From henning@tle.ENET.dec.com Mon Jan 11 18:41:28 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id SAA95606 for ; Mon, 11 Jan 1999 18:41:28 -0800 (PST) Received: from fddi-sgi.engr.sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id SAA83921 for ; Mon, 11 Jan 1999 18:41:27 -0800 (PST) mail_from (henning@tle.ENET.dec.com) Received: from mail11.digital.com (mail11.digital.com [192.208.46.10]) by fddi-sgi.engr.sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id SAA09869 for ; Mon, 11 Jan 1999 18:41:26 -0800 (PST) mail_from (henning@tle.ENET.dec.com) Received: from us9rmc.zko.dec.com (us9rmc.zko.dec.com [16.33.16.29]) by mail11.digital.com (8.9.1a/8.9.1/WV2.0c) with SMTP id VAA07326 for ; Mon, 11 Jan 1999 21:42:30 -0500 (EST) Received: from tle.enet by us9rmc.zko.dec.com (5.65/rmc-17Jan97) id AA13310; Mon, 11 Jan 99 21:38:00 -0500 Message-Id: <9901120238.AA13310@us9rmc.zko.dec.com> Received: from tle.enet; by us9rmc.enet; Mon, 11 Jan 99 21:38:00 EST Date: Mon, 11 Jan 99 21:38:00 EST From: John Henning 11-Jan-1999 2138 -0500 To: mccalpin Cc: me@tle.ENET.dec.com Apparently-To: mccalpin Subject: Compaq AlphaServer DS20 Status: RO Hi John, Here's the AlphaServer DS20 - John Script started on Thu Jan 7 19:28:17 1999 % head -1 /etc/motd Digital UNIX V4.0E (Rev. 1091); Thu Dec 17 14:29:49 EST 1998 % diff stream_d.f_as_at_ftp_site_22may97 tryit.f 53c53,54 < PARAMETER (n=2000000,offset=0,ndim=n+offset,ntimes=10) --- > PARAMETER (n = 1000200 ,offset=0) > PARAMETER (ndim=n+offset,ntimes=10) 334a336,359 > END > > DOUBLE PRECISION FUNCTION SECOND > * > * Assume we will be called often enough so that rpcc counter > * doesn't overflow! > * > INTEGER*8 TICK, PREV_TICK > DOUBLE PRECISION RPCC_DELTA, PREV_SEC > INTEGER*8 FIXED_RPCC > EXTERNAL FIXED_RPCC, RPCC_DELTA > > DATA PREV_TICK /0/ > > TICK = FIXED_RPCC() > IF (PREV_TICK .EQ. 0) THEN > SECOND = 0D0 > ELSE > SECOND = PREV_SEC + RPCC_DELTA (%VAL(PREV_TICK), %VAL(TICK)) > END IF > > PREV_SEC = SECOND > PREV_TICK = TICK > RETURN % cat rpcc.c /* rpcc.c j.henning hacked from example by d.grunwald*/ #include #include #include #include #include #include static double scale = -1; static pid_t mypid = -1; /* fix_rpcc adds the two halves of the register (bias & offset), since some tools return it w/o this fixup */ unsigned long fix_rpcc(unsigned long rpcc) { unsigned long fixed; unsigned long offset = (rpcc >> 32) & 0xffffffff; unsigned long cycles = (rpcc & 0xffffffff); fixed = (unsigned long) offset + (unsigned long) cycles; /* printf ("%8lx %8lx %8lx ", offset, cycles, fixed); */ return fixed; } /* Compute time in *seconds* (note that Grunwald's original version did microseconds) given two fixedup RPCCs */ double rpcc_delta(unsigned long ul_start, unsigned long ul_stop) { double retval; if ( scale == -1 ) { int err; double freq; struct cpu_info cpubuf ; err = getsysinfo(GSI_CPU_INFO, &cpubuf, sizeof(cpubuf)); /*printf ("err= %d freq = %d current_cpu = %d\n", err, cpubuf.mhz, cpubuf.current_cpu);*/ if (err < 0) { freq = 300; } else { freq = cpubuf.mhz; } freq *= 1000000; /* * Scale is the 1 over the clock frequency */ scale = 1/(double) freq; } if (ul_stop < ul_start) { retval = ((0xffffffff - ul_start) + ul_stop) * scale; } else { retval = (ul_stop - ul_start) * scale; } /* printf ("%f\n", retval); */ return retval; } unsigned long fixed_rpcc(void) { unsigned long i; i = asm("rpcc %v0"); return fix_rpcc(i); } % f77 -non_shared -fast -O5 -unroll 8 -arch ev6 -assume nounderscore -V rpcc.o tryit.f % grep COMPILER: tryit.l COMPILER: DIGITAL Fortran 77 V5.2-171-428BH % unlimit % ./a.out ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 1000200 Offset = 0 The total memory requirement is 22 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 11034 microseconds (= 11034 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 1076.8628 0.0150 0.0149 0.0156 Scale: 1243.9420 0.0131 0.0129 0.0138 Add: 1179.0963 0.0204 0.0204 0.0211 Triad: 1323.2813 0.0182 0.0181 0.0187 Sum of a is = 1.153531441436304E+018 Sum of b is = 2.307062882849082E+017 Sum of c is = 3.076083843811272E+017 % ls -l tryit.f a.out -rwxr-xr-x 1 john users 917504 Jan 7 19:29 a.out -rw-r--r-- 1 john users 11106 Jan 7 19:27 tryit.f % exit % script done on Thu Jan 7 19:31:44 1999 From dsiebert@engineering.uiowa.edu Fri Jan 15 08:12:14 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id IAA02013 for ; Fri, 15 Jan 1999 08:12:13 -0800 (PST) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id IAA57645 for ; Fri, 15 Jan 1999 08:12:12 -0800 (PST) mail_from (dsiebert@engineering.uiowa.edu) Received: from server07.icaen.uiowa.edu (server07.icaen.uiowa.edu [128.255.17.47]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id IAA09024 for ; Fri, 15 Jan 1999 08:12:11 -0800 (PST) mail_from (dsiebert@engineering.uiowa.edu) From: dsiebert@engineering.uiowa.edu Received: from server01.icaen.uiowa.edu (root@server01.icaen.uiowa.edu [128.255.17.41]) by server07.icaen.uiowa.edu (8.8.7/8.7.1) with ESMTP id KAA10575 for ; Fri, 15 Jan 1999 10:12:10 -0600 (CST) Received: from l-ecn007.icaen.uiowa.edu (dsiebert@l-ecn007.icaen.uiowa.edu [128.255.17.107]) by server01.icaen.uiowa.edu (8.8.7/8.7.1) with ESMTP id KAA07278; Fri, 15 Jan 1999 10:12:09 -0600 (CST) Received: (from dsiebert@localhost) by l-ecn007.icaen.uiowa.edu (8.8.6 (PHNE_14041)/client-6.6) id KAA29111; Fri, 15 Jan 1999 10:12:08 -0600 (CST) Message-Id: <199901151612.KAA29111@l-ecn007.icaen.uiowa.edu> Organization: Iowa Computer Aided Engineering Network, University of Iowa Subject: Re: PA-8500 SPEC results mystifying To: mccalpin (John McCalpin) Date: Fri, 15 Jan 1999 10:12:08 -0600 (CST) In-Reply-To: <9901060924.ZM85519@frakir.engr.sgi.com> from "John McCalpin" at Jan 6, 99 09:24:33 am Reply-To: douglas-siebert@uiowa.edu X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Status: RO Hi John, Since this PA-8500 / C360 thing has me interested, I did a few STREAM runs on some boxes I have here. Unfortunately, I do not have the f77 compilers installed -- only the f90 which are still not up to par with the f77 in optimization capability. With my minimal knowledge of Fortran I got it running but it was actually slower in the Fortran version, which probably shouldn't happen. Anyway, I am just submitting C results. Compiler is about six months old, but that is probably not an issue for this sort of thing. I compiled on the C360 and used the same binary on a C180, C200, and C240 as well. Systems were lightly used (users logged into the console doing not much of anything other than Netscape -- only a couple % CPU usage) The C360, BTW, was a C200 upgraded to C360. I don't know if the C360s you buy as C360s are different at all (probably not, HP is supposed to be coming out with some new stuff soon) /usr/bin/cc: LINT A.10.32.16 CXREF A.10.32.16 HP92453-01 A.10.32.18 HP C Compiler /usr/lib/libc: $Revision: 76.3 $ /opt/ansic/lbin/ccom: LINT A.10.32.16 CXREF A.10.32.16 HP92453-01 A.10.32.18 HP C Compiler HP-UX SLLIC/OPTIMIZER UX.10.20.559 (DAVIS): 05/26/98 Ucode Code Generator - UX10.20.61 (PACG_UX10.MULTI_BL33) REV: HP SESD Code High Level Optimizer - UX.10.21.971124 (UX10.MULTI) [-DHLO_RELEASE +O3] - 05-Mar-98.13:41 /usr/lib/libc: $Revision: 76.3 $ Compiler options (just guessing at the ones to use) +Oall -Wl,-a,archive Its interesting that the C180 results are really not much different than the C360 results. HP has a lot of room for improvement in this area -- the PA-8500 would look a whole lot better if they put a good memory system in the box. As it is, they are working with something that is about on par with the best PC motherboards, which isn't much to brag about in the workstation world, especially given the DS20 results and the likely better results for the POWER3 machines. C180 (PA-8000@180MHz 512K + 512K offchip L1, IIRC) ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 7000000, Offset = 0 Total memory required = 160.2 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 9999 microseconds. Each test below will take on the order of 260000 microseconds. (= 26 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 302.7027 0.3740 0.3700 0.3800 Scale: 266.6667 0.4260 0.4200 0.4300 Add: 190.9091 0.8850 0.8800 0.9000 Triad: 204.8780 0.8290 0.8200 0.8400 C200 (PA-8200@200MHz 512Ki + 1MBd offchip L1) ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 7000000, Offset = 0 Total memory required = 160.2 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 9999 microseconds. Each test below will take on the order of 270000 microseconds. (= 27 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 287.1795 0.3910 0.3900 0.4000 Scale: 273.1707 0.4190 0.4100 0.4200 Add: 202.4096 0.8310 0.8300 0.8400 Triad: 224.0000 0.7540 0.7500 0.7600 C240 (PA-8200@236MHz 2MB + 2MB off chip L1) ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 7000000, Offset = 0 Total memory required = 160.2 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 9999 microseconds. Each test below will take on the order of 240000 microseconds. (= 24 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 339.3939 0.3320 0.3300 0.3400 Scale: 320.0000 0.3530 0.3500 0.3600 Add: 215.3846 0.7850 0.7800 0.7900 Triad: 218.1818 0.7700 0.7700 0.7700 C360 (PA-8500@367MHz 512Ki + 1MBd on chip L1) ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 7000000, Offset = 0 Total memory required = 160.2 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 9999 microseconds. Each test below will take on the order of 209999 microseconds. (= 21 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 311.1111 0.3670 0.3600 0.3700 Scale: 273.1707 0.4110 0.4100 0.4200 Add: 210.0000 0.8070 0.8000 0.8100 Triad: 207.4074 0.8100 0.8100 0.8100 -- Douglas Siebert Director of Computing Facilities douglas-siebert@uiowa.edu Division of Mathematical Sciences, U of Iowa I plan to live forever, or die trying. From pulsar@springnet1.com Wed Jan 13 10:37:48 PST 1999 Article: 28644 of comp.benchmarks Path: news.corp.sgi.com!news.sgi.com!howland.erols.net!vixen.cso.uiuc.edu!pulsar From: pulsar@springnet1.com (Pulsar) Newsgroups: comp.sys.mac.advocacy,comp.benchmarks Subject: Re: Memory bandwidth of New G3 boxes? Date: Wed, 13 Jan 1999 03:58:23 -0600 Organization: University of Illinois at Urbana-Champaign Lines: 31 Message-ID: References: <77crd9$9er$1@hecate.umd.edu> <77fa8b$l4o$1@hecate.umd.edu> NNTP-Posting-Host: spider.cso.uiuc.edu Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Newsreader: Yet Another NewsWatcher 2.4.0 X-Proxy-Client: e-cowan@uiuc.edu from cloudy.users.shout.net Xref: news.corp.sgi.com comp.sys.mac.advocacy:498928 comp.benchmarks:28644 Status: RO In article <77fa8b$l4o$1@hecate.umd.edu>, davewang@cslab.kecl.ntt.co.jp.DELETE.delete.DELETE@Glue.umd.edu (David T. Wang) wrote: > Pulsar (pulsar@springnet1.com) wrote: > : In article <77crd9$9er$1@hecate.umd.edu>, > : davewang@cslab.kecl.ntt.co.jp.DELETE.delete.DELETE@Glue.umd.edu (David T. > : Wang) wrote: > > : > I'm curious about what kind of bandwidth the new G3 boxes with 100 MHz > : > busses provide. If you've been fortunate enough to acquire one of those > : > boxes, please see if you can run Stream and collect some numbers. > : > > : > http://www.cs.virginia.edu/stream/ > > : I'll try it on an iMac and a first series PowerMac G3 tomorrow and report > : the results. > > Thank you. They look quite respectable. Please tell us what you get > with the iMac. http://www.cloudmaster.ml.org/pulsar/stream/ I have various scores up on a web page. There are 233MHz iMac and new 266MHz iMac scores listed. There is also a 266MHz PowerMac G3 and a 266MHz PowerBook G3 listed. Ryan Tokarek From jr@cfm.brown.edu Wed Jan 13 11:14:32 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id LAA98425 for ; Wed, 13 Jan 1999 11:14:32 -0800 (PST) Received: from deliverator.sgi.com (deliverator.sgi.com [150.166.91.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id LAA85974 for ; Wed, 13 Jan 1999 11:14:31 -0800 (PST) mail_from (jr@cfm.brown.edu) Received: from mail.cfm.brown.edu (e151.cfm.brown.edu [128.148.160.151]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id LAA20120 for ; Wed, 13 Jan 1999 11:05:23 -0800 (PST) mail_from (jr@cfm.brown.edu) Received: from cough (cough [128.148.160.138]) by mail.cfm.brown.edu (8.9.1/8.9.1) with SMTP id OAA08932 for ; Wed, 13 Jan 1999 14:07:57 -0500 (EST) Date: Wed, 13 Jan 1999 14:08:14 -0500 (EST) From: Jon Reiter To: mccalpin@cthulhu Subject: stream on power 3 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Status: RO We have a 43P-260 here. I got Assignment: 586.6667 Scaling : 640.0000 Summing : 812.3077 SAXPYing : 812.3077 On one processor and Assignment: 845.8352 Scaling: 839.8298 Summing: 990.7404 SAXPYing: 994.5894 on two processors. We are still trying to track down the problems (all of this seems too slow). It looks like we aren't using the same compiler they are (no power-3 specific optimization) but we _are_ using what IBM claims is the latest compiler. So it is... -jon From davewang@cslab.kecl.ntt.co.jp.DELETE.delete.DELETE@Glue.umd.edu Tue Jan 12 09:14:03 PST 1999 Article: 28634 of comp.benchmarks From: davewang@cslab.kecl.ntt.co.jp.DELETE.delete.DELETE@Glue.umd.edu (David T. Wang) Newsgroups: comp.sys.mac.advocacy,comp.benchmarks Subject: Re: Memory bandwidth of New G3 boxes? Date: 12 Jan 1999 11:04:43 GMT Organization: University of Maryland, College Park Lines: 82 Message-ID: <77fa8b$l4o$1@hecate.umd.edu> References: <77crd9$9er$1@hecate.umd.edu> NNTP-Posting-Host: vdal6.eng.umd.edu X-Newsreader: TIN [version 1.2 PL2] Path: news.corp.sgi.com!news.sgi.com!csulb.edu!logbridge.uoregon.edu!newspeer.monmouth.com!news.lightlink.com!news4.his.com!news.cs.jhu.edu!haven.umd.edu!hecate.umd.edu!vdal6.eng.umd.edu!davewang@cslab.kecl.ntt.co.jp.DELETE.delete.DELETE Xref: news.corp.sgi.com comp.sys.mac.advocacy:498368 comp.benchmarks:28634 Status: RO Pulsar (pulsar@springnet1.com) wrote: : In article <77crd9$9er$1@hecate.umd.edu>, : davewang@cslab.kecl.ntt.co.jp.DELETE.delete.DELETE@Glue.umd.edu (David T. : Wang) wrote: : > I'm curious about what kind of bandwidth the new G3 boxes with 100 MHz : > busses provide. If you've been fortunate enough to acquire one of those : > boxes, please see if you can run Stream and collect some numbers. : > : > http://www.cs.virginia.edu/stream/ : I'll try it on an iMac and a first series PowerMac G3 tomorrow and report : the results. : PowerMac G3 : 100Mhz bus : 400Mhz processor : 200MHz 1 meg L2 cache : 64 megs of RAM. : ------------------------------------------------------------- : This system uses 8 bytes per DOUBLE PRECISION word. : ------------------------------------------------------------- : Array size = 400000, Offset = 0 : Total memory required = 9.2 MB. : Each test is run 10 times, but only : the *best* time for each is used. : ------------------------------------------------------------- : Your clock granularity/precision appears to be 17 microseconds. : Each test below will take on the order of 20713 microseconds. : (= 1218 clock ticks) : Increase the size of the arrays if this shows that : you are not getting at least 20 clock ticks per test. : ------------------------------------------------------------- : WARNING -- The above is only a rough guideline. : For best results, please be sure you know the : precision of your system timer. : STREAM PPC604 OPT : ------------------------------------------------------------- : Function Rate (MB/s) RMS time Min time Max time : Copy: 233.4914 0.0276 0.0274 0.0278 : Scale: 222.5236 0.0291 0.0288 0.0300 : Add: 230.6473 0.0417 0.0416 0.0419 : Triad: 212.9736 0.0452 0.0451 0.0454 : STREAM PPC601 OPT : ------------------------------------------------------------- : Function Rate (MB/s) RMS time Min time Max time : Copy: 236.8265 0.0272 0.0270 0.0277 : Scale: 220.6060 0.0292 0.0290 0.0294 : Add: 215.3625 0.0448 0.0446 0.0451 : Triad: 212.9878 0.0457 0.0451 0.0508 : STREAM PPC UNROLL : ------------------------------------------------------------- : Function Rate (MB/s) RMS time Min time Max time : Copy: 184.2999 0.0348 0.0347 0.0350 : Scale: 180.5767 0.0356 0.0354 0.0358 : Add: 185.3640 0.0519 0.0518 0.0520 : Triad: 184.6402 0.0522 0.0520 0.0531 : STREAM PPC BASE : ------------------------------------------------------------- : Function Rate (MB/s) RMS time Min time Max time : Copy: 178.3674 0.0360 0.0359 0.0361 : Scale: 173.3290 0.0370 0.0369 0.0371 : Add: 182.5269 0.0527 0.0526 0.0528 : Triad: 182.6693 0.0528 0.0526 0.0535 : Ryan Tokarek : : Thank you. They look quite respectable. Please tell us what you get with the iMac. -- davewang@cslab.kecl.ntt.co.jp.I.like.green.eggs.and.ham,not.spam All statements are personal opinions Not speaking for NTT or University of Maryland Kyoto, Japan. From alanc@west.sun.com Tue Jan 12 13:57:37 PST 1999 Article: 28637 of comp.benchmarks Path: news.corp.sgi.com!news.sgi.com!howland.erols.net!feed1.news.rcn.net!rcn!chippy.visi.com!news-out.visi.com!nntp.teleport.com!news1.teleport.com!alanc From: alanc@west.sun.com (Alan Charlesworth) Newsgroups: comp.sys.mac.advocacy,comp.benchmarks Subject: Re: Memory bandwidth of New G3 boxes? Message-ID: References: <77crd9$9er$1@hecate.umd.edu> <77fa8b$l4o$1@hecate.umd.edu> X-Newsreader: MT-NewsWatcher 2.4.4 Lines: 80 Date: Tue, 12 Jan 1999 10:20:41 -0800 NNTP-Posting-Host: 216.26.6.13 X-Complaints-To: news@teleport.com X-Trace: news1.teleport.com 916165192 216.26.6.13 (Tue, 12 Jan 1999 10:19:52 PDT) NNTP-Posting-Date: Tue, 12 Jan 1999 10:19:52 PDT Xref: news.corp.sgi.com comp.sys.mac.advocacy:498512 comp.benchmarks:28637 Status: RO Below are Stream rates for the original iMac, and a PMac 7300/200 (604e). Apple has increased its memory bandwidth by 2.7x since the Feb 97 intro of the 7300/200: Stream average (PPC604 opt) PMac 7300 200 MHz 604e 50 MHz bus 82 MBps iMac 233 MHz G3 67 MHz Bus 133 MBps Blue G3 400 MHz G3 100 MHz Bus 225 MBps (from previous posting) The fastest PC in the Streeam web site (http://www.cs.virginia.edu/stream/) listing is a "Generic_440BX_400", with an average bandwidth of 323 MBps. iMac, 67 MHz Bus, 233 MHz G3, 116 MHz, 0.5 MB cache, 160 MB memory STREAM PPC604 OPT ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 139.7502 0.0462 0.0458 0.0469 Scale: 135.9331 0.0473 0.0471 0.0475 Add: 129.8877 0.0742 0.0739 0.0744 Triad: 126.7544 0.0760 0.0757 0.0762 STREAM PPC601 OPT ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 134.9841 0.0478 0.0474 0.0484 Scale: 126.9136 0.0507 0.0504 0.0509 Add: 126.7946 0.0760 0.0757 0.0762 Triad: 126.1979 0.0762 0.0761 0.0765 STREAM PPC UNROLL ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 115.7826 0.0556 0.0553 0.0563 Scale: 114.4738 0.0560 0.0559 0.0563 Add: 120.7805 0.0797 0.0795 0.0798 Triad: 119.3050 0.0806 0.0805 0.0809 STREAM PPC BASE ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 115.6696 0.0562 0.0553 0.0581 Scale: 114.5639 0.0566 0.0559 0.0585 Add: 120.9205 0.0802 0.0794 0.0831 Triad: 120.4638 0.0807 0.0797 0.0835 PMac 7300, 50 MHz Bus, 200 MHz 604e, 1 MB cache, 288 MB memory STREAM PPC604 OPT ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 84.7222 0.0766 0.0755 0.0844 Scale: 72.0859 0.0891 0.0888 0.0894 Add: 91.3772 0.1053 0.1051 0.1056 Triad: 81.5779 0.1180 0.1177 0.1184 STREAM PPC601 OPT ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 83.6787 0.0777 0.0765 0.0854 Scale: 71.1981 0.0901 0.0899 0.0904 Add: 90.4517 0.1063 0.1061 0.1065 Triad: 79.9480 0.1202 0.1201 0.1206 STREAM PPC UNROLL ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 62.9803 0.1027 0.1016 0.1110 Scale: 54.2387 0.1183 0.1180 0.1185 Add: 74.1301 0.1297 0.1295 0.1299 Triad: 64.7384 0.1484 0.1483 0.1486 STREAM PPC BASE ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 64.8062 0.1006 0.0988 0.1143 Scale: 54.2630 0.1188 0.1179 0.1232 Add: 74.9216 0.1291 0.1281 0.1323 Triad: 65.3315 0.1479 0.1469 0.1547 From tomdean@ix.netcom.com Mon Feb 15 04:03:24 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id EAA20221 for ; Mon, 15 Feb 1999 04:03:23 -0800 (PST) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id EAA53029 for ; Mon, 15 Feb 1999 04:03:22 -0800 (PST) mail_from (tomdean@ix.netcom.com) Received: from ares.cs.Virginia.EDU (ares.cs.Virginia.EDU [128.143.136.41] (may be forged)) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id EAA04857 for ; Mon, 15 Feb 1999 04:03:21 -0800 (PST) mail_from (tomdean@ix.netcom.com) Received: from ix.netcom.com (sil-wa15-29.ix.netcom.com [207.93.148.29]) by ares.cs.Virginia.EDU (8.8.5/8.8.5) with ESMTP id HAA15268 for ; Mon, 15 Feb 1999 07:03:17 -0500 (EST) Received: (from tomdean@localhost) by ix.netcom.com (8.9.3/8.8.8) id EAA00464 for mccalpin@cs.virginia.edu; Mon, 15 Feb 1999 04:03:15 -0800 (PST) (envelope-from tomdean) Date: Mon, 15 Feb 1999 04:03:15 -0800 (PST) From: Thomas Dean Message-Id: <199902151203.EAA00464@ix.netcom.com> To: mccalpin@cs.virginia.edu Subject: Diff to cast int to double w/gcc Status: RO *** stream_d.c~ Sat Aug 24 12:18:22 1996 --- stream_d.c Mon Feb 15 04:00:21 1999 *************** *** 94,100 **** printf(HLINE); printf("Array size = %d, Offset = %d\n" , N, OFFSET); printf("Total memory required = %.1f MB.\n", ! (3 * N * BytesPerWord) / 1048576.0); printf("Each test is run %d times, but only\n", NTIMES); printf("the *best* time for each is used.\n"); --- 94,100 ---- printf(HLINE); printf("Array size = %d, Offset = %d\n" , N, OFFSET); printf("Total memory required = %.1f MB.\n", ! (double)(3 * N * BytesPerWord) / 1048576.0); printf("Each test is run %d times, but only\n", NTIMES); printf("the *best* time for each is used.\n"); From tomdean@ix.netcom.com Mon Feb 15 04:42:59 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id EAA20300 for ; Mon, 15 Feb 1999 04:42:59 -0800 (PST) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id EAA33544 for ; Mon, 15 Feb 1999 04:42:58 -0800 (PST) mail_from (tomdean@ix.netcom.com) Received: from ares.cs.Virginia.EDU (ftp.cs.Virginia.EDU [128.143.137.19]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id EAA00617 for ; Mon, 15 Feb 1999 04:42:57 -0800 (PST) mail_from (tomdean@ix.netcom.com) Received: from ix.netcom.com (sil-wa15-29.ix.netcom.com [207.93.148.29]) by ares.cs.Virginia.EDU (8.8.5/8.8.5) with ESMTP id HAA15781 for ; Mon, 15 Feb 1999 07:42:52 -0500 (EST) Received: (from tomdean@localhost) by ix.netcom.com (8.9.3/8.8.8) id EAA00743; Mon, 15 Feb 1999 04:42:51 -0800 (PST) (envelope-from tomdean) Date: Mon, 15 Feb 1999 04:42:51 -0800 (PST) Message-Id: <199902151242.EAA00743@ix.netcom.com> From: Thomas Dean To: mccalpin@cs.virginia.edu Subject: Output of stream_d Status: RO My system is a DEC CELEBRIS 5133DP, 2ea, 133Mhz Pentium Processors, running FreeBSD: Copyright (c) 1982, 1986, 1989, 1991, 1993 The Regents of the University of California. All rights reserved. FreeBSD 4.0-CURRENT #4: Mon Feb 15 03:34:29 PST 1999 tomdean@celebris:/usr/src/sys/compile/CELEBRIS-SMP Timecounter-tdd "i8254" frequency 1193032 Hz cost 2547 ns CPU: Pentium/P54C (586-class CPU) Origin = "GenuineIntel" Id = 0x525 Stepping=5 Features=0x3bf real memory = 100663296 (98304K bytes) avail memory = 95055872 (92828K bytes) Programming 16 pins in IOAPIC #0 FreeBSD/SMP: Multiprocessor motherboard cpu0 (BSP): apic id: 0, version: 0x00030010, at 0xfee00000 cpu1 (AP): apic id: 1, version: 0x00030010, at 0xfee00000 io0 (APIC): apic id: 2, version: 0x000f0011, at 0xfec00000 Preloaded elf kernel "kernel" at 0xf02b0000. Probing for devices on PCI bus 0: chip0: rev 0x11 on pci0.0. 0 ncr0: rev 0x02 int a irq 11 on pci0.1.0 chip1: rev 0x88 on pci0.2.0 ============================= Here is the result of running stream_d on my system. # time stream_d ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity appears to be less than one microsecond. Each test below will take on the order of 273437 microseconds. Floating exception (core dumped) 1.524u 2.264s 0:10.00 37.8% 6+30720k 0+1697io 0pf+0w celebris: {71} !cc cc -O2 -m486 stream_d.c second_cpu.c -o stream_d -lm celebris: {72} time stream_d ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 7812 microseconds. Each test below will take on the order of 273437 microseconds. (= 35 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 33.5738 0.4899 0.4766 0.5000 Scale: 47.6279 0.3407 0.3359 0.3516 Add: 52.0678 0.4696 0.4609 0.4766 Triad: 51.2000 0.4836 0.4688 0.4922 19.191u 0.788s 0:19.97 100.0% 5+30244k 0+0io 0pf+0w From ashizawa@nsg.sgi.com Mon Mar 1 22:18:01 1999 Received: from tristan.nsg.sgi.com (tristan.nsg.sgi.com [134.14.128.9]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via SMTP id WAA46329 for ; Mon, 1 Mar 1999 22:17:57 -0800 (PST) Received: from nsg.sgi.com by tristan.nsg.sgi.com via ESMTP (951211.SGI.8.6.12.PATCH1502/951211.SGI) id PAA08557; Tue, 2 Mar 1999 15:18:00 +0900 Sender: ashizawa@nsg.sgi.com Message-ID: <36DB828E.56BABC05@nsg.sgi.com> Date: Tue, 02 Mar 1999 15:17:50 +0900 From: Yoshio Ashizawa Organization: Nihon SiliconGraphics Cray K.K. X-Mailer: Mozilla 4.04 [ja_euc] (X11; I; IRIX 6.3 IP32) MIME-Version: 1.0 To: John McCalpin CC: ashizawa@nsg.sgi.com Subject: STREAM performance on 256 pe O2k. Content-Type: text/plain; charset=iso-2022-jp Content-Transfer-Encoding: 7bit Status: RO Hi McCalpin-san, Thank you for your reply and suggestion that I had better check the number. And I checked source again I found that I made a mistake in timing fuction. Now I get the reasonable expected numbers, 42 GB/s. We can make a reasonable criteria for the procurement. And I submitted a bug #676604 that I had expericenced with STREAM. Thank you again, ashizawa typhoon 37% ./go + export OMP_NUM_THREADS=256 + export OMP_DYNAMIC=FALSE + export MPC_GANG=OFF + export _DSM_PPM=2 + export _DSM_MUSTRUN=1 + ./stream ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 160000000 Offset = 0 The total memory requirement is 3662 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 4 microseconds The tests below will each take a time on the order of 34356 microseconds (= 8589 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 42825.1375 0.0618 0.0598 0.0721 Scale: 45299.3503 0.0622 0.0565 0.0739 Add: 51802.3491 0.0982 0.0741 0.2121 Triad: 52977.9809 0.0738 0.0725 0.0765 Sum of a is = 1.8452812500173852E+20 Sum of b is = 3.6905625000360264E+19 Sum of c is = 4.9207500000800375E+19 Large Case: typhoon 21% ./go + export OMP_NUM_THREADS=256 + export OMP_DYNAMIC=FALSE + export MPC_GANG=OFF + export _DSM_PPM=2 + export _DSM_MUSTRUN=1 + ./stream ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 400000000 Offset = 0 The total memory requirement is 9155 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 4 microseconds The tests below will each take a time on the order of 90979 microseconds (= 22745 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 42824.2493 0.1505 0.1494 0.1525 Scale: 43213.4826 0.1638 0.1481 0.2423 Add: 48285.8374 0.2003 0.1988 0.2025 Triad: 49275.5196 0.1982 0.1948 0.2171 Sum of a is = 4.6132031251661586E+20 Sum of b is = 9.2264062499237265E+19 Sum of c is = 1.2301875000272059E+20 -- ------------------------------------------------------ Yoshio Ashizawa Scalable Systems Technology Center Silicon Graphics Inc. Japan E-mail: ashizawa@nsg.sgi.com PHONE : +81-3-5488-1838 FAX : +81-3-5420-2397 From ce107@cfm.brown.edu Fri Mar 19 13:13:08 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id NAA80156 for ; Fri, 19 Mar 1999 13:13:08 -0800 (PST) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id NAA77066 for ; Fri, 19 Mar 1999 13:13:07 -0800 (PST) mail_from (ce107@cfm.brown.edu) Received: from e151.cfm.brown.edu (e151.cfm.brown.edu [128.148.160.151]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id QAA08792 for ; Fri, 19 Mar 1999 16:13:06 -0500 (EST) mail_from (ce107@cfm.brown.edu) Received: from urania.cfm.brown.edu (urania [128.148.160.125]) by e151.cfm.brown.edu (8.9.1/8.9.1) with ESMTP id QAA16432 for ; Fri, 19 Mar 1999 16:11:55 -0500 (EST) From: "C. Evangelinos" Received: (from ce107@localhost) by urania.cfm.brown.edu (8.9.1/8.9.1) id QAA29892 for mccalpin@frakir.engr.sgi.com; Fri, 19 Mar 1999 16:13:05 -0500 (EST) Message-Id: <199903192113.QAA29892@urania.cfm.brown.edu> Subject: Re: Hardware Prefetching and other CRI T3E goodies (was Re: MIPS R10k/R12k references/URL?) To: mccalpin (John McCalpin) Date: Fri, 19 Mar 1999 16:13:05 -0500 (EST) In-Reply-To: <9903191139.ZM79784@frakir.engr.sgi.com> from "John McCalpin" at Mar 19, 99 11:39:34 am MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Status: RO > I don't know the details. I know that there is a barrier tree in > the router network and that the design came from the T3E folks > in Chippewa Falls. I don't have time to recheck but this sounds more like the faster T3D barrier tree. > >PS> Have you had any IBM formal submission of STREAM numbers for Model > > 260? I wouldn't want to submit mine. > > No word from IBM. I have gotten numbers from a number of customers, > and they are all in the 550-800 MB/s range. I may put them in the > standard table soon, just to try to get IBM's attention.... Well in that case I'm not breaking any embarrassing secret: These are the best results I managed to get compiling with flags for Power3 and a high res PowerPC timer: xlf 6.1.0.0 and also 5.1.1.0 -O3 -qarch=pwr3 -qtune=pwr3 1 548.9 545.3 766.0 767.6 Using the standard timer may actually give slightly higher numbers but the number of timer ticks used is very small. Following the suggestions in the RS/6000 Scientific and Technical Computing: POWER3 Introduction and Tuning Guide (that describe how to tune BLAS/ESSL dcopy()) I transformed the STREAM source code for COPY and managed to get approximately 862.3MB/s for COPY user-visible bandwidth(*). This is still below IBM claims (and anyway the standard STREAM results are supposed to be attainable without hand-tuning *and* the technique used is so simple the compiler should have it in as a special case given that copying vectors is very frequently done in codes). (*) I had to compile with -O2 instead of -O3 as -O3 negated the effects of the hand-tuning. Constantinos Evangelinos Center for Fluid Mechanics Brown University/Division of Applied Mathematics From stefand@ferrari.lcam.u-psud.fr Fri Feb 5 09:25:17 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id JAA02792 for ; Fri, 5 Feb 1999 09:25:17 -0800 (PST) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id JAA71157 for ; Fri, 5 Feb 1999 09:25:07 -0800 (PST) mail_from (stefand@ferrari.lcam.u-psud.fr) Received: from ares.cs.Virginia.EDU (mail.cs.Virginia.EDU [128.143.137.19]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id JAB07609 for ; Fri, 5 Feb 1999 09:24:54 -0800 (PST) mail_from (stefand@ferrari.lcam.u-psud.fr) X-UIDL: 919201530.057 From: stefand@ferrari.lcam.u-psud.fr Received: from ferrari.lcam.u-psud.fr (ferrari.lcam.u-psud.fr [194.57.28.131]) by ares.cs.Virginia.EDU (8.8.5/8.8.5) with ESMTP id MAA22224 for ; Fri, 5 Feb 1999 12:23:25 -0500 (EST) Received: (from stefand@localhost) by ferrari.lcam.u-psud.fr (8.7.1/8.7.1) id SAA20642; Fri, 5 Feb 1999 18:28:22 GMT Message-Id: <199902051828.SAA20642@ferrari.lcam.u-psud.fr> Subject: Stream results for IBM RS/6k 595 To: mccalpin@cs.virginia.edu Date: Fri, 5 Feb 1999 18:28:22 +0000 (WET) X-Loop: stefand@ferrari.lcam.u-psud.fr Reply-To: sad@utk.edu X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Zm-Priority: Low Status: RO Hi John, I just tried to get stream data for our IBM RS/6k 43-260 Power3 machine but that didn't work very well: So far only the xlf 4.x compiler is installed, and it doesn't know yet about the Power3 machines. Hence, the numbers are not representative (in fact, smaller than the ones you list in the experimental results section). I am planning to have the compiler upgraded and then I will make another attempt. However, once on it I decided to benchmark our IBM RS/6k 595 (AIX 4.1, 2 GB RAM) and found the so far fastest results for this machine with these compiler options: xlf -O3 -qnozerosize -qarch=pwr2 -qtune=pwr2 stream_d.f second_cpu.f \ -bmaxdata:500000000 -qhot -qhot=arraypad -qalign=4k -o strem_595 to be in the following price range Function Rate (MB/s) RMS time Min time Max time Copy: 847.0588 .4002 .3400 .4500 Scale: 900.0000 .3965 .3200 .4400 Add: 881.6327 .5402 .4900 .6600 Triad: 960.0000 .5273 .4500 .6000 (The Copy result is from a re-run of the same binary, well within the 5% threshold you also mentioned). The complete log is appended for your records. This is faster than the currently listed 595 entry (probably still with xlf 3.x?) -- maybe you could also list the dates of the entries so one can deduce such things? -- and also faster than I can get the Power3 machine with this compiler. You will notice that I had to increase n quite a bit since the box was too fast. For fun I'll build an OS/2 version tonight for pentium and p6 class machines. Cheers, Stefan Anyway, here is the complete log: Concord: (IBM 595 2 GB RAM) xlf 4.x xlf -O3 -qnozerosize -qarch=pwr2 -qtune=pwr2 stream_d.f second_cpu.f -bmaxdata:500000000 -qhot -o strem_595 timex ./strem_595 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 18000000 Offset = 0 The total memory requirement is 411 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 10000 microseconds The tests below will each take a time on the order of 220000 microseconds (= 22 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 738.4615 .4071 .3900 .4200 Scale: 738.4615 .3981 .3900 .4100 Add: 830.7692 .5381 .5200 .5600 Triad: 847.0588 .5333 .5100 .5600 Sum of a is = 0.207594140609274962E+20 Sum of b is = 0.415188281269478605E+19 Sum of c is = 0.553584374975187046E+19 real 46.64 user 20.25 sys 2.65 xlf -O3 -qnozerosize -qarch=pwr2 -qtune=pwr2 stream_d.f second_cpu.f \ -bmaxdata:500000000 -qhot -qhot=arraypad -o strem_595 concord pts/4 [169] /lscrtch/sad/stream> timex ./strem_595 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 18000000 Offset = 0 The total memory requirement is 411 MB You are running each test 10 times The *best* time for each test is used -------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 757.8947 .3991 .3800 .4100 Scale: 757.8947 .4002 .3800 .4300 Add: 864.0000 .5382 .5000 .5600 Triad: 864.0000 .5212 .5000 .5400 Sum of a is = 0.207594140609274962E+20 Sum of b is = 0.415188281269478605E+19 Sum of c is = 0.553584374975187046E+19 real 46.46 user 19.93 sys 2.89 xlf -O3 -qnozerosize -qarch=pwr2 -qtune=pwr2 stream_d.f second_cpu.f \ -bmaxdata:500000000 -qhot -qhot=arraypad -qalign=4k -o strem_595 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 18000000 Offset = 0 The total memory requirement is 411 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 10000 microseconds The tests below will each take a time on the order of 220000 microseconds (= 22 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 738.4615 .4092 .3900 .4300 Scale: 757.8947 .3991 .3800 .4100 Add: 864.0000 .5262 .5000 .5500 Triad: 881.6327 .5253 .4900 .5500 Sum of a is = 0.207594140609274962E+20 Sum of b is = 0.415188281269478605E+19 Sum of c is = 0.553584374975187046E+19 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 18000000 Offset = 0 The total memory requirement is 411 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 10000 microseconds The tests below will each take a time on the order of 220000 microseconds (= 22 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 757.8947 .3971 .3800 .4100 Scale: 757.8947 .3911 .3800 .4000 Add: 847.0588 .5372 .5100 .5500 Triad: 847.0588 .5241 .5100 .5500 Sum of a is = 0.207594140609274962E+20 Sum of b is = 0.415188281269478605E+19 Sum of c is = 0.553584374975187046E+19 real 46.24 user 19.90 sys 2.74 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 18000000 Offset = 0 The total memory requirement is 411 MB You are running each test 80 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 10000 microseconds The tests below will each take a time on the order of 260000 microseconds (= 26 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 800.0000 .4022 .3600 .4500 Scale: 900.0000 .3965 .3200 .4400 Add: 881.6327 .5402 .4900 .6600 Triad: 960.0000 .5273 .4500 .6000 Sum of a is = 0.440152552974613180E+102 Sum of b is = 0.880305106010380985E+101 Sum of c is = 0.117374014130075336E+102 real 187.16 user 150.50 sys 2.99 * Another run with the same binary resulted in: Function Rate (MB/s) RMS time Min time Max time Copy: 847.0588 .4002 .3400 .4500 Scale: 778.3784 .3996 .3700 .5200 Add: 919.1489 .5340 .4700 .5800 Triad: 900.0000 .5306 .4800 .5700 Sum of a is = 0.440152552974613180E+102 Sum of b is = 0.880305106010380985E+101 Sum of c is = 0.117374014130075336E+102 real 167.85 user 150.43 sys 3.25 -- ========================================================================= Dr. Stefan A. Deutscher | (+33-(0)1) voice fax Laboratoire des Collisions Atomiques et | LCAM : 6915-7699 6915-7671 Mol\'{e}culaires (LCAM), B\^{a}timent 351 | home : 5624-0992 call first Universit\'{e} de Paris-Sud | email: sad@utk.edu 91405 Orsay Cedex, France (Europe) | (forwarded to France) ========================================================================= Do you know what they call a quarter-pounder with cheese in Paris? From gratz@ite.inf.tu-dresden.de Tue Apr 6 08:09:36 PDT 1999 Article: 88087 of comp.arch Path: news.corp.sgi.com!enews.sgi.com!torn!sunqbc.risq.qc.ca!feed2.news.luth.se!luth.se!fu-berlin.de!zrz.TU-Berlin.DE!news.tu-chemnitz.de!news.inf.tu-dresden.de!not-for-mail From: Achim Gratz Newsgroups: comp.arch Subject: Re: STREAM benchmark: Compiler dependent ? Date: 6 Apr 1999 15:25:07 +0200 Organization: Institute of Computer Engineering, CS department, TU Dresden, Germany Lines: 129 Sender: gratz@ite127.inf.tu-dresden.de Message-ID: References: <37028AD1.EDAA4D00@moene.indiv.nluug.nl> <7duutu$hcd@news1.newsguy.com> NNTP-Posting-Host: ite127.inf.tu-dresden.de Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: kastor.inf.tu-dresden.de 923405108 24639 141.76.85.27 (6 Apr 1999 13:25:08 GMT) X-Complaints-To: usenet@irz.inf.tu-dresden.de NNTP-Posting-Date: 6 Apr 1999 13:25:08 GMT X-Newsreader: Gnus v5.6.11/Emacs 19.34 Xref: news.corp.sgi.com comp.arch:88087 Status: RO ce107@cfm.brown.edu (C. Evangelinos) writes: > Which in turn poses the question: On systems where non-cacheable > load/stores (like VIS) can be used or on others where certain stores > can avoid write-allocate traffic (eg. the new PIII SIMD extensions > unless I'm mistaken) are there compilers that will actually generate > the code (preferably in the absence of non-portable directives) or > does one always need to go and use assembler or special tools like > Intel's Vtune? STREAM experimental numbers contain very impressive > results for Suns using VIS but at least the 4.2 compilers didn't > appear capable to generate such code even for a simple copy loop with > a hard-coded loop length... No idea if the 5.0 compilers are any > better. The SunPro 5.0 compilers can handle limited cases (I think they have an inline template for copy loops and probably some vector loops). If you look at the latest SPEC submissions you should find the extra flags to force these optimizations. One of my students is working on a code generator for SUIF that uses VIS instructions and some less well-known special features of UltraSPARC to produce optimal code for STREAM and slightly more complicated loops that arise e.g. in sparse matrix solvers. I have one isolated loop where I have a parametrizable m4 assembly code generator for. Below are results for C++ compiled and C++ with an assembly inset. The results are in MB/s, the rows are results for different sizes of the respective arrays and the columns are the size of a single array (there are six of it), minimum bandwith obtained for any repetition, root mean squared (RMS over all repetitions), two arithmetic averages (one with loop overhead and one without) and at last maximum bandwith obtained for any repetition. Machine tested is an U60/2360 (only one processor used). Even though I've used the high-res timer, the first three rows have to be takenn with a grain of salt due to the timing overhead. The code can't use any VIS, but I've used read and write prefetch in the assembly version to make the in-cache results more stable. The fun thing is that while the assembly was originally intended for out-of-cache sizes it turns out that it does much better than the compiled code for the in-cache case and actually approaches the theoretical limit, whereas the compiled code barely makes half of it. ~/solver (5) vh5 Size of one vector in MB: 13.57 Number of repetitions: 1024 Min RMS Avg1 Avg2 Max 16 388.8 894.9 622.7 895.6 896.8 32 628.3 1108. 855.6 1108. 1129. 64 744.3 1332. 1144. 1333. 1342. 128 766.0 1410. 1297. 1414. 1423. 256 830.7 1433. 1366. 1439. 1456. 512 773.9 979.4 961.2 980.5 989.1 1024 744.4 771.5 765.4 771.5 773.4 2048 600.1 730.4 730.0 733.0 762.0 4096 717.7 753.5 752.1 753.7 759.1 8192 677.3 724.1 720.6 724.8 755.3 16384 601.5 656.5 656.7 657.1 678.5 32768 546.7 561.8 562.0 562.2 615.8 65536 511.5 526.5 526.6 526.8 578.6 131072 377.3 386.2 386.9 386.9 460.6 262144 299.1 308.1 308.6 308.6 363.8 524288 294.5 302.8 302.9 302.9 331.2 1048576 298.3 301.8 301.8 301.8 309.5 ~/solver (6) vh5.RWA Size of one vector in MB: 13.57 Number of repetitions: 1024 Min RMS Avg1 Avg2 Max 16 158.8 1356. 901.5 1396. 1429. 32 764.8 1881. 1406. 1885. 1893. 64 1291. 2171. 1811. 2173. 2180. 128 1742. 2374. 2141. 2375. 2384. 256 1992. 2403. 2242. 2404. 2415. 512 2174. 2448. 2360. 2448. 2458. 1024 2056. 2315. 2274. 2317. 2337. 2048 2120. 2425. 2402. 2428. 2480. 4096 1936. 2254. 2246. 2258. 2315. 8192 1893. 2233. 2232. 2238. 2304. 16384 1697. 1716. 1714. 1716. 1720. 32768 1118. 1171. 1172. 1173. 1378. 65536 953.4 990.1 992.3 992.7 1222. 131072 598.2 630.9 633.2 633.4 825.5 262144 446.6 459.2 460.4 460.4 572.2 524288 440.2 447.8 448.1 448.1 496.2 1048576 431.0 439.9 440.0 440.0 454.2 STREAM results for the same machine (these results could probably be improved by more sophisticated placements of the arrays): 1 CPU base result (C 4.2 compiled) Assignment: 355.8263 0.1428 0.1415 0.1474 Scaling : 344.8744 0.1463 0.1459 0.1480 Summing : 364.6393 0.2083 0.2070 0.2126 SAXPYing : 350.3134 0.2159 0.2155 0.2174 1 CPU result (C 4.2 compiled with compiler generated prefetch) Function Rate (MB/s) RMS time Min time Max time Assignment: 348.7358 0.1391 0.1376 0.1533 Scaling : 354.9112 0.1362 0.1352 0.1401 Summing : 407.6570 0.1779 0.1766 0.1792 SAXPYing : 411.1044 0.1764 0.1751 0.1770 2 CPU result (C 4.2 compiled with compiler generated prefetch) Assignment: 480.4420 0.1014 0.0999 0.1244 Scaling : 484.4671 0.1007 0.0991 0.1242 Summing : 525.6779 0.1388 0.1370 0.1711 SAXPYing : 561.3117 0.1288 0.1283 0.1298 1 CPU (C 4.2 based, experimental with VIS/assembly, no prefetch) Function Rate (MB/s) RMS time Min time Max time Assignment: 617.6496 0.0780 0.0777 0.0792 Scaling : 458.0195 0.1060 0.1048 0.1123 Summing : 511.3926 0.1414 0.1408 0.1433 SAXPYing : 496.0045 0.1464 0.1452 0.1489 2 CPU (C 4.2 based, experimental with VIS/assembly, no prefetch) Function Rate (MB/s) RMS time Min time Max time Assignment: 741.3924 0.0651 0.0647 0.0665 Scaling : 685.2633 0.0703 0.0700 0.0710 Summing : 734.5970 0.0982 0.0980 0.0985 SAXPYing : 726.9866 0.0992 0.0990 0.0995 While the prefetch isn't very effective at improving the numbers per se, it helps tremendously with badly placed data. Achim Gratz. --+<[ It's the small pleasures that make life so miserable. ]>+-- WWW: http://www.inf.tu-dresden.de/~ag7/{english/} E-Mail: gratz@ite.inf.tu-dresden.de Phone: +49 351 463 - 8325 From lindahl@cs.virginia.edu Tue Mar 30 16:40:01 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id QAA04323 for ; Tue, 30 Mar 1999 16:40:01 -0800 (PST) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id QAA51766 for ; Tue, 30 Mar 1999 16:40:00 -0800 (PST) mail_from (lindahl@cs.virginia.edu) Received: from ares.cs.Virginia.EDU (ares.cs.Virginia.EDU [128.143.136.41] (may be forged)) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id QAB00780 for ; Tue, 30 Mar 1999 16:39:58 -0800 (PST) mail_from (lindahl@cs.virginia.edu) Received: from adder.cs.Virginia.EDU (adder.cs.Virginia.EDU [128.143.137.18]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id TAA13722 for ; Tue, 30 Mar 1999 19:39:54 -0500 (EST) From: Greg Lindahl Received: (from lindahl@localhost) by adder.cs.Virginia.EDU (8.8.5/8.8.5) id TAA03968 for mccalpin@cs.virginia.edu; Tue, 30 Mar 1999 19:39:54 -0500 (EST) Message-Id: <199903310039.TAA03968@adder.cs.Virginia.EDU> Subject: stream results, Compaq XP1000 To: mccalpin@cs.virginia.edu Date: Tue, 30 Mar 1999 19:39:53 -0500 (EST) X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Status: RO This is the new single-cpu motherboard Compaq workstation: Compiled with a modern Digital Fortran compiler (not sure what version), f77 -O5 -tune ev6 -non_shared stream_d.f, then run under Linux on a 500 mhz XP1000. Almost but not quite as good as the DS20. gcc majorly sucks in this situation, even the latest egcs. I'm not sure how it could fuck up something so simple. Sure, it can't issue the 'write hint' instruction, but for the loops themselves, almost all of your time is waiting for main memory, so you'd think even stupid code would be sufficient. Ah well. ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 2000000 Offset = 0 The total memory requirement is 45 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 24636 microseconds (= 24636 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 896.7368 0.0358 0.0357 0.0364 Scale: 879.5568 0.0364 0.0364 0.0365 Add: 892.0275 0.0538 0.0538 0.0539 Triad: 888.3963 0.0541 0.0540 0.0541 Sum of a is = 2.306601562591874E+018 Sum of b is = 4.613203124856438E+017 Sum of c is = 6.150937500141256E+017 From ashoks@Eng.Sun.COM Fri Apr 2 17:53:56 PST 1999 Article: 87992 of comp.arch Path: news.corp.sgi.com!news.sgi.com!news.maxwell.syr.edu!newsfeed.direct.ca!su-news-hub1.bbnplanet.com!denver-news-feed1.bbnplanet.com!news.gtei.net!namche.sun.com!ebaynews1.Ebay.Sun.COM!engnews2.Eng.Sun.COM!not-for-mail From: ashoks@Eng.Sun.COM (Ashok Singhal) Newsgroups: comp.arch Subject: Re: STREAM benchmark: Compiler dependent ? Date: 2 Apr 1999 22:25:42 GMT Organization: Sun Microsystems Inc. Lines: 38 Message-ID: <7e3g56$57c$1@engnews2.Eng.Sun.COM> References: Reply-To: ashoks@Eng.Sun.COM NNTP-Posting-Host: dovetail.eng.sun.com Xref: news.corp.sgi.com comp.arch:87992 Status: RO In article fsf@cfm.brown.edu, ce107@cfm.brown.edu (C. Evangelinos) writes: > > Which in turn poses the question: On systems where non-cacheable > load/stores (like VIS) can be used or on others where certain stores > can avoid write-allocate traffic (eg. the new PIII SIMD extensions > unless I'm mistaken) are there compilers that will actually generate > the code (preferably in the absence of non-portable directives) or > does one always need to go and use assembler or special tools like > Intel's Vtune? STREAM experimental numbers contain very impressive > results for Suns using VIS but at least the 4.2 compilers didn't > appear capable to generate such code even for a simple copy loop with > a hard-coded loop length... No idea if the 5.0 compilers are any > better. > The 5.0 compilers can handle the copy loop, but apparently they don't do the others. The compilers can now also generate prefetch instructions. [In addition, the processors are now faster (400 MHz) and the system speed has also improved to 100 MHz] On a 8-way 400 MHz E3500, here is what we get (compiled with 5.0 compilers, no assembly): CPUs COPY SCALE SUM TRIAD 1 577.7 369.1 375.9 431.7 2 1160.9 721.9 690.5 847.4 4 1884.1 984.3 1043.9 1189.9 8 3021.7 1538.5 1512.1 1768.8 (With only 4 boards, this box runs out of memory bank bandwidth before it runs out of backplane bandwidth for the 8CPU case.) Ashok Singhal Sun Microsystems From rfridman@ucalgary.ca Mon Apr 12 12:52:41 PDT 1999 Article: 29080 of comp.benchmarks Path: news.corp.sgi.com!news.sgi.com!newshub.sdsu.edu!newsfeed.berkeley.edu!cyclone.bc.net!rover.ucs.ualberta.ca!news.ucalgary.ca!news From: Robert Fridman Newsgroups: comp.benchmarks Subject: Re: What about Apple PowerPC ? Date: Mon, 12 Apr 1999 13:36:57 -0600 Organization: WurcNet Inc. Lines: 181 Message-ID: <7eti5s$806@ds2.acs.ucalgary.ca> References: <370b77da.117241834@news.cea.fr> NNTP-Posting-Host: @moonbase.hpc.maci.ucalgary.ca Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="------------05B5052B0072B81992C224ED" X-Mailer: Mozilla 4.5 [en] (X11; I; OSF1 V4.0 alpha) X-Accept-Language: en Xref: news.corp.sgi.com comp.benchmarks:29080 Status: RO This is a multi-part message in MIME format. --------------05B5052B0072B81992C224ED Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Denis COURTIER wrote: > > Hi ! > > Is anybody knows something about benchmarks on Apple PowerPC ? > I would like some information about perfs on I-Mac or Macintosh G3, > but I can't find any on www.specbench.org , Wintune, and so on... > > My goal is to compare I-Mac and PC with Celeron or Pentium... I have an Apple laptop (running Linux, 233 MHz PPC750 with .5 MB L2 cache 66MHz bus). I find that benchmarks don't reflect the intended use of the computer. For example, the STREAM bechmark reported a modeds 97MB/sec of memory bandwidth. I then tried STREAM with smaller data sizes and was amazed that if the data could fit in the L2 cache, the bandwidth trippled. So this architecture rewards good, tight programming. I included my STREAM runs at the end of this email. Another 'benchmark' I tried was running the benchmark gforth (GNU Forth) programs. My laptop performed well against a 450MHz celeron. The results show that the laptop is a lot slower on one test (.5 the speed), 18% faster on another and 50% and 25% slower on the last 2. The architecture of my laptop (due to a slow system bus) rewards using the cache. This seems to suite may applications. Benchmarks don't seem to reflect this. If anyone has any PII or PIII laptops runing Linux, I'd be interesed in their performance numbers for STREAM and gforth. -- Robert. ---------------------------------------------------------------------- Robert Fridman rfridman@ucalgary.ca WurcNet Inc. University of Calgary Calgary, Alberta phone (403) 220-6779 Canada fax (403) 284-4707 --------------05B5052B0072B81992C224ED Content-Type: text/plain; charset=us-ascii; name="plotdata" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="plotdata" Array size = 10000, Offset = 0 Total memory required = 0.2 MB. Function Rate (MB/s) RMS time Min time Max time Copy: 338.9337 0.0005 0.0005 0.0005 Scale: 333.9580 0.0034 0.0005 0.0105 Add: 332.4415 0.0035 0.0007 0.0108 Triad: 338.5347 0.0007 0.0007 0.0007 Array size = 15000, Offset = 0 Total memory required = 0.3 MB. Function Rate (MB/s) RMS time Min time Max time Copy: 238.8216 0.0036 0.0010 0.0110 Scale: 219.7889 0.0011 0.0011 0.0012 Add: 204.4340 0.0067 0.0018 0.0122 Triad: 212.6390 0.0055 0.0017 0.0117 Array size = 20000, Offset = 0 Total memory required = 0.5 MB. Function Rate (MB/s) RMS time Min time Max time Copy: 182.1260 0.0071 0.0018 0.0219 Scale: 194.2932 0.0055 0.0016 0.0118 Add: 183.7676 0.0061 0.0026 0.0127 Triad: 176.2776 0.0067 0.0027 0.0129 Array size = 40000, Offset = 0 Total memory required = 0.9 MB. Function Rate (MB/s) RMS time Min time Max time Copy: 121.2802 0.0098 0.0053 0.0164 Scale: 113.4362 0.0057 0.0056 0.0059 Add: 117.1304 0.0120 0.0082 0.0185 Triad: 117.8348 0.0110 0.0081 0.0183 Array size = 80000, Offset = 0 Total memory required = 1.8 MB. Function Rate (MB/s) RMS time Min time Max time Copy: 102.7780 0.0175 0.0125 0.0335 Scale: 99.4795 0.0160 0.0129 0.0233 Add: 104.2630 0.0247 0.0184 0.0386 Triad: 105.4943 0.0222 0.0182 0.0384 Array size = 100000, Offset = 0 Total memory required = 2.3 MB. Function Rate (MB/s) RMS time Min time Max time Copy: 100.9976 0.0216 0.0158 0.0367 Scale: 99.6945 0.0178 0.0160 0.0260 Add: 104.8172 0.0275 0.0229 0.0531 Triad: 106.2185 0.0253 0.0226 0.0425 Array size = 250000, Offset = 0 Total memory required = 5.7 MB. Function Rate (MB/s) RMS time Min time Max time Copy: 98.7996 0.0463 0.0405 0.0815 Scale: 97.5706 0.0466 0.0410 0.0806 Add: 104.0945 0.0597 0.0576 0.0745 Triad: 104.3550 0.0575 0.0575 0.0576 Array size = 500000, Offset = 0 Total memory required = 11.4 MB. Function Rate (MB/s) RMS time Min time Max time Copy: 98.2149 0.0946 0.0815 0.1623 Scale: 97.6838 0.0860 0.0819 0.0992 Add: 104.3832 0.1187 0.1150 0.1313 Triad: 104.5187 0.1190 0.1148 0.1313 Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Function Rate (MB/s) RMS time Min time Max time Copy: 97.9756 0.1768 0.1633 0.2679 Scale: 97.7368 0.1641 0.1637 0.1655 Add: 104.4727 0.2302 0.2297 0.2320 Triad: 104.5920 0.2298 0.2295 0.2304 --------------05B5052B0072B81992C224ED Content-Type: text/plain; charset=us-ascii; name="Benchres" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="Benchres" Some benchmark results for various combinations of hardware, Gforth version, and Gforth configuration. Unless specified otherwise, the default configurations were used. You can measure your combination with `make bench'. You can find a table comparing Gforth with six interpretive Forth systems in the manual (Section Performance), and a comparison with more systems in http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz. All times are given in seconds of user time. siev bubble matrix fib machine and configuration 2.39 3.38 4.13 4.18 Celeron 450 (Mendocino); gcc-2.7.2.3 -DFORCE-REG -DDIRECT-THREADED; gforth-0.3.0; ELF 4.63 5.06 3.36 5.25 Apple PowerBook G3 (233MHz); egcs-1.0.2 prerelease; gforth-0.4.0 6.88 7.43 5.80 8.15 AMD K6-2 300MHz 1M PB cache 100MHz; gcc-2.7.2.3 -DFORCE-REG; gforth-0.4.0; ELF 7.36 8.16 7.73 9.04 Pentium-MMX 200MHz 512K PB cache; egcs-1.1b -DFORCE-REG; gforth-0.4.0; ELF 10.91 11.94 11.24 13.13 Pentium 133MHz 256K PB cache; gcc-2.6.3 -DFORCE_REG; gforth-0.1beta; a.out 11.16 11.86 10.64 12.53 Pentium 133MHz 512K PB cache; gcc-2.7.2p -DFORCE_REG, gforth-0.3.0; ELF 12.62 13.56 11.04 14.97 AMD K6 166MHz 512K PB cache; gcc-2.7.2p -DFORCE_REG, gforth-0.3.0; ELF 11.81 14.39 13.61 15.07 IBM/Cyrix-6x86 133MHz (P166+) 512K PB cache; gcc-2.7.2.1 -DFORCE_REG -DDIRECT_THREADED; gforth-0.3.0; ELF 12.08 11.90 11.06 12.09 Cyrix-6x86MX 166MHz (PR200) 512K PB Cache; gcc-2.7.2.1 -DFORCE_REG -DDIRECT_THREADED; gforth-0.3.0; ELF 29.89 35.42 26.96 34.59 i486 66MHz 256K cache; gcc-2.6.3 -DFORCE_REG -DDIRECT_THREADED; gforth-0.1beta; a.out 39.50 45.91 36.73 44.90 i486 50MHz 256K cache; gcc-2.7.0 -DFORCE_REG -DDIRECT_THREADED; gforth-0.1beta 42.82 46.74 38.69 48.30 i486 50MHz 256K cache; gcc-2.7.0 -DFORCE_REG; gforth-0.1beta 3.09 3.24 2.39 3.42 21164A (Alpha,164LX) 600MHz 2M cache; gcc-2.7.2.1+gas; gforth-0.3.0 3.7 3.8 2.8 4.1 21164A (Alpha,PC164) 500MHz 2M cache; gcc-2.7.2.1+as (Digital Unix); gforth-0.3.0 7.0 7.6 6.2 7.7 21064A (Alpha,Cabriolet) 300MHz 2M cache; gcc-2.7.2; gforth-0.2.0 7.49 7.85 6.21 8.07 R4400 250 Mhz 2Mb cache; gcc-2.7.2.2 8.17 9.01 6.24 9.35 R10000 (SGI PowerChallenge XL) 195MHz 2M cache; gcc-2.7.2 -DFORCE_REG; gforth-0.3.0 17.3 19.0 14.1 18.3 R4000 (DecStation 5000/150) 100MHz 1M cache; gcc-2.4.5; gforth-0.1beta 50.9 56.8 42.4 52.0 R3000 (DecStation 5000/200) 25MHz 64K+64K cache; gcc-2.5.8 -DFORCE_REG; gforth-0.1beta 7.8 8.6 7.0 10.3 UltraSparc-II 248MHz; Solaris.5.5.1; gcc-2.7.1; gforth-0.3.0 28.5 31.1 26.3 33.3 SuperSparc (Sparcstation 10) 40MHz; Solaris.5.5.1; gcc-2.7.1; gforth-0.3.0 59.5 65.8 69.5 61.9 FJMB86903 (SPARC ELC) 33MHz; gcc-2.5.8; gforth-0.1beta 84.34 91.49 76.16 88.83 L64801 25MHz (SPARC IPC) 64K WT cache; gcc-2.4.5; gforth-0.1beta 11.6 12.1 10.8 15.6 PA8000 (HP C160) 160MHz 64M RAM; gcc-2.7.2; gforth-0.3.0 30.0 34.1 20.5 33.0 PA-RISC 1.1 (HP 720) 50MHz 64K cache; gcc-2.6.3 -DDIRECT_THREADED; gforth-0.1beta 6.81 7.53 5.10 8.12 PPC604e (PowerMac) 200MHz; Linux; gcc-2.7.2.1; gforth-0.4.0 8.25 10.09 6.45 10.34 PPC604e (PowerMac) 200MHz; Linux; gcc-2.7.2.1; gforth-0.3.0 (indirect threaded) 14.05 16.96 11.14 17.51 PPC604 (PowerMac) 132MHz 256K L2 cache; MkLinux 2.1; gcc-2.7.2.1; gforth-0.3.0 --------------05B5052B0072B81992C224ED-- From kristian@snabel.com Mon Apr 19 03:33:25 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id DAA46601 for ; Mon, 19 Apr 1999 03:33:02 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id DAA61925 for ; Mon, 19 Apr 1999 03:32:55 -0700 (PDT) mail_from (kristian@snabel.com) Received: from ares.cs.Virginia.EDU (ares.cs.Virginia.EDU [128.143.136.41] (may be forged)) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id DAA09001 for ; Mon, 19 Apr 1999 03:32:52 -0700 (PDT) mail_from (kristian@snabel.com) Received: from snabelco_pdc.snabel.com (snabelco.dk [195.78.66.94]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id GAA25110 for ; Mon, 19 Apr 1999 06:32:49 -0400 (EDT) Received: from snabel.com (195.97.174.138 [195.97.174.138]) by snabelco_pdc.snabel.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.1960.3) id 28QTCGYZ; Mon, 19 Apr 1999 12:33:06 +0100 Sender: kristian@cs.virginia.edu Message-ID: <371B063E.3E165A80@snabel.com> Date: Mon, 19 Apr 1999 12:32:30 +0200 From: "Kristian Elof Sørensen" Organization: Snabel & co X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.36 i686) MIME-Version: 1.0 To: mccalpin@cs.virginia.edu Subject: Stream results for a Celeron Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Status: RO Hi I have just got my new Linux box up and running at work for use as a Java development machine and a CVS server, and have ran some benchmarks on it. You don't seem to have the numbers for this particular PC variant in the chart at http://www.cs.virginia.edu/stream/standard/Bandwidth.html It's a PC with a 400 MHz Celeron Mendocino (a PII with 128 KB L2 cache running at the CPU's clockspeed built into the CPU housing) on an ASUS P2B motherboard with 128 MB 100 MHz RAM Function Rate (MB/s) RMS time Min time Max time Copy: 197.5237 0.0866 0.0810 0.1053 Scale: 202.2192 0.0813 0.0791 0.0913 Add: 238.8392 0.1029 0.1005 0.1139 Triad: 218.6668 0.1121 0.1098 0.1166 It sure beats the K6-2 350 MHz I have at home for running Java programs under the Linux JVM. I have found that the performance of this particular application maps rather lineary to the CPU memory bandwidth as measured by Stream. Of course with some very effective jit's or some other such tricks this observation doesn't neccesarrily hold, though it makes perfect sence that Java applications must be limited nearly exclusively by the memory bandwith to and from the CPU. From alexander@vvv.srcc.msu.su Tue Apr 20 09:45:14 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id JAA49491 for ; Tue, 20 Apr 1999 09:42:22 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id JAA05144 for ; Tue, 20 Apr 1999 09:42:21 -0700 (PDT) mail_from (alexander@vvv.srcc.msu.su) Received: from ares.cs.Virginia.EDU (ares.cs.Virginia.EDU [128.143.137.19]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id JAA05822 for ; Tue, 20 Apr 1999 09:42:15 -0700 (PDT) mail_from (alexander@vvv.srcc.msu.su) Received: from vvv.srcc.msu.su (root@vvv.srcc.msu.su [212.192.244.50]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id MAA10742 for ; Tue, 20 Apr 1999 12:41:46 -0400 (EDT) Received: from greenland (parallel.srcc.msu.su [212.192.244.167]) by vvv.srcc.msu.su (8.9.1/8.9.1) with SMTP id UAA03487 for ; Tue, 20 Apr 1999 20:41:29 +0400 Message-ID: <009a01be8b4c$accddae0$a7f4c0d4@srcc.msu.su> From: "Alexander N. Andreyev" To: Subject: STREAM results for Siemens RM600 Date: Tue, 20 Apr 1999 20:41:37 +0400 MIME-Version: 1.0 Content-Type: text/plain; charset="koi8-r" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2014.211 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2014.211 Status: RO Hello, John I send you results for the STREAM benchmark on Siemens' RM600E system. The OS is Reliant Unix 5.44, the compiler flags to "cc" were "-O2 -lm". As you may see, Add rate is larger than Copy rate! Is this at all possible and how this may happen? ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 16 microseconds. Each test below will take on the order of 189254 microseconds. (= 11828 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 75.2676 0.2141 0.2126 0.2171 Scale: 74.9931 0.2156 0.2134 0.2180 Add: 97.4220 0.2518 0.2464 0.2656 Triad: 74.6078 0.3239 0.3217 0.3271 Best wishes, Alexander N. Andreyev Laboratory of Parallel Information Technologies, SRCC, MSU alexander@vvv.srcc.msu.su, http://alex.motor.ru, ICQ: #3523091 Parallel Computing: http://parallel.srcc.msu.su [russian language] From john@mccalpin.com Mon Apr 26 07:04:36 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id HAA11431 for ; Mon, 26 Apr 1999 07:04:36 -0700 (PDT) Received: from conn.engr.sgi.com (conn.engr.sgi.com [130.62.53.66]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id HAA00805 for ; Mon, 26 Apr 1999 07:04:36 -0700 (PDT) mail_from (guest@conn.engr.sgi.com) Received: (from guest@localhost) by conn.engr.sgi.com (8.8.7/8.8.7) id GAA12477 for mccalpin@engr.sgi.com; Mon, 26 Apr 1999 06:11:24 -0700 Date: Mon, 26 Apr 1999 06:11:24 -0700 From: John McCalpin Message-Id: <199904261311.GAA12477@conn.engr.sgi.com> To: mccalpin@cthulhu Subject: stream on 320 Status: RO Compiled with -O2 -strength-reduce -funroll-loops The machine is a 450 MHz uniprocessor Silicon Graphics Model 320.... The performance stinks.... [guest@conn]$ repeat 10 ./stream_wall ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74271 microseconds. (= 74271 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.6738 0.0935 0.0932 0.0956 Scale: 180.6012 0.0887 0.0886 0.0889 Add: 227.8294 0.1065 0.1053 0.1155 Triad: 227.9440 0.1054 0.1053 0.1058 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74317 microseconds. (= 74317 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.6904 0.0935 0.0932 0.0954 Scale: 180.5216 0.0894 0.0886 0.0954 Add: 227.9873 0.1054 0.1053 0.1060 Triad: 227.9829 0.1054 0.1053 0.1055 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74095 microseconds. (= 74095 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7862 0.0934 0.0931 0.0954 Scale: 180.6950 0.0886 0.0885 0.0886 Add: 227.9765 0.1053 0.1053 0.1055 Triad: 228.0154 0.1053 0.1053 0.1054 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74160 microseconds. (= 74160 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7807 0.0945 0.0931 0.1056 Scale: 180.6685 0.0886 0.0886 0.0888 Add: 228.0891 0.1053 0.1052 0.1056 Triad: 227.9853 0.1053 0.1053 0.1056 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74065 microseconds. (= 74065 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7513 0.0934 0.0932 0.0953 Scale: 180.6175 0.0887 0.0886 0.0888 Add: 228.0870 0.1064 0.1052 0.1153 Triad: 227.9981 0.1054 0.1053 0.1057 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74225 microseconds. (= 74225 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7346 0.0945 0.0932 0.1032 Scale: 180.6563 0.0886 0.0886 0.0888 Add: 228.1390 0.1053 0.1052 0.1055 Triad: 228.0653 0.1053 0.1052 0.1055 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74145 microseconds. (= 74145 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7548 0.0935 0.0932 0.0955 Scale: 180.7072 0.0886 0.0885 0.0886 Add: 227.8575 0.1064 0.1053 0.1154 Triad: 227.9700 0.1054 0.1053 0.1055 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74180 microseconds. (= 74180 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.5966 0.0946 0.0932 0.1033 Scale: 180.5809 0.0887 0.0886 0.0889 Add: 227.9853 0.1053 0.1053 0.1055 Triad: 227.9440 0.1054 0.1053 0.1059 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74172 microseconds. (= 74172 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.8048 0.0934 0.0931 0.0953 Scale: 180.7114 0.0886 0.0885 0.0888 Add: 227.8704 0.1072 0.1053 0.1154 Triad: 228.0871 0.1054 0.1052 0.1059 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74191 microseconds. (= 74191 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.8194 0.0945 0.0931 0.1031 Scale: 180.7214 0.0886 0.0885 0.0887 Add: 228.1674 0.1052 0.1052 0.1053 Triad: 228.3041 0.1052 0.1051 0.1054 [guest@conn]$ cat >stream_wall.LOG Compiled with -O2 -strength-reduce -funroll-loops -rw-r--r-- 1 guest 998 8688 Apr 26 05:51 stream_d.o -rwxr-xr-x 1 guest 998 9751 Apr 26 05:52 stream_wall -rw-r--r-- 1 guest 998 381 Apr 26 05:53 times.txt [guest@conn]$ repeat 10 ./stream_wall ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74271 microseconds. (= 74271 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.6738 0.0935 0.0932 0.0956 Scale: 180.6012 0.0887 0.0886 0.0889 Add: 227.8294 0.1065 0.1053 0.1155 Triad: 227.9440 0.1054 0.1053 0.1058 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74317 microseconds. (= 74317 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.6904 0.0935 0.0932 0.0954 Scale: 180.5216 0.0894 0.0886 0.0954 Add: 227.9873 0.1054 0.1053 0.1060 Triad: 227.9829 0.1054 0.1053 0.1055 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74095 microseconds. (= 74095 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7862 0.0934 0.0931 0.0954 Scale: 180.6950 0.0886 0.0885 0.0886 Add: 227.9765 0.1053 0.1053 0.1055 Triad: 228.0154 0.1053 0.1053 0.1054 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74160 microseconds. (= 74160 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7807 0.0945 0.0931 0.1056 Scale: 180.6685 0.0886 0.0886 0.0888 Add: 228.0891 0.1053 0.1052 0.1056 Triad: 227.9853 0.1053 0.1053 0.1056 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74065 microseconds. (= 74065 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7513 0.0934 0.0932 0.0953 Scale: 180.6175 0.0887 0.0886 0.0888 Add: 228.0870 0.1064 0.1052 0.1153 Triad: 227.9981 0.1054 0.1053 0.1057 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74225 microseconds. (= 74225 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7346 0.0945 0.0932 0.1032 Scale: 180.6563 0.0886 0.0886 0.0888 Add: 228.1390 0.1053 0.1052 0.1055 Triad: 228.0653 0.1053 0.1052 0.1055 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74145 microseconds. (= 74145 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.7548 0.0935 0.0932 0.0955 Scale: 180.7072 0.0886 0.0885 0.0886 Add: 227.8575 0.1064 0.1053 0.1154 Triad: 227.9700 0.1054 0.1053 0.1055 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74180 microseconds. (= 74180 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.5966 0.0946 0.0932 0.1033 Scale: 180.5809 0.0887 0.0886 0.0889 Add: 227.9853 0.1053 0.1053 0.1055 Triad: 227.9440 0.1054 0.1053 0.1059 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74172 microseconds. (= 74172 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.8048 0.0934 0.0931 0.0953 Scale: 180.7114 0.0886 0.0885 0.0888 Add: 227.8704 0.1072 0.1053 0.1154 Triad: 228.0871 0.1054 0.1052 0.1059 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 74191 microseconds. (= 74191 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 171.8194 0.0945 0.0931 0.1031 Scale: 180.7214 0.0886 0.0885 0.0887 Add: 228.1674 0.1052 0.1052 0.1053 Triad: 228.3041 0.1052 0.1051 0.1054 From john@mccalpin.com Mon Apr 26 11:02:25 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id LAA11944 for ; Mon, 26 Apr 1999 11:02:25 -0700 (PDT) Received: from conn.engr.sgi.com (conn.engr.sgi.com [130.62.53.66]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id LAA12915 for ; Mon, 26 Apr 1999 11:02:24 -0700 (PDT) mail_from (guest@conn.engr.sgi.com) Received: (from guest@localhost) by conn.engr.sgi.com (8.8.7/8.8.7) id KAA00969 for mccalpin@engr.sgi.com; Mon, 26 Apr 1999 10:01:33 -0700 Date: Mon, 26 Apr 1999 10:01:33 -0700 From: John McCalpin Message-Id: <199904261701.KAA00969@conn.engr.sgi.com> To: mccalpin@cthulhu Subject: offset tests on 320 Status: RO 0 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78325 microseconds. (= 78325 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.6180 0.0963 0.0960 0.0984 Scale: 173.5508 0.0923 0.0922 0.0923 Add: 220.7607 0.1098 0.1087 0.1187 Triad: 219.6574 0.1103 0.1093 0.1183 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78278 microseconds. (= 78278 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.7396 0.0963 0.0960 0.0984 Scale: 173.5677 0.0922 0.0922 0.0924 Add: 220.6856 0.1169 0.1088 0.1670 Triad: 219.5712 0.1094 0.1093 0.1101 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78315 microseconds. (= 78315 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.5782 0.0964 0.0961 0.0984 Scale: 173.4811 0.0923 0.0922 0.0923 Add: 220.8602 0.1098 0.1087 0.1188 Triad: 219.6233 0.1093 0.1093 0.1095 1 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 1 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78322 microseconds. (= 78322 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.8292 0.0958 0.0953 0.0987 Scale: 170.9055 0.0938 0.0936 0.0946 Add: 217.1513 0.1108 0.1105 0.1117 Triad: 218.7645 0.1111 0.1097 0.1206 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 1 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78322 microseconds. (= 78322 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.8468 0.0956 0.0953 0.0979 Scale: 170.9402 0.0936 0.0936 0.0937 Add: 217.2378 0.1108 0.1105 0.1124 Triad: 218.8923 0.1107 0.1096 0.1197 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 1 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 2 microseconds. Each test below will take on the order of 78304 microseconds. (= 39152 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.7993 0.0967 0.0954 0.1054 Scale: 170.9000 0.0937 0.0936 0.0937 Add: 217.1159 0.1106 0.1105 0.1107 Triad: 218.8822 0.1110 0.1096 0.1214 2 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 2 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78290 microseconds. (= 78290 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.8457 0.1034 0.0959 0.1231 Scale: 171.5911 0.0933 0.0932 0.0934 Add: 218.6011 0.1133 0.1098 0.1214 Triad: 219.1060 0.1109 0.1095 0.1214 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 2 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78277 microseconds. (= 78277 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.8404 0.0974 0.0959 0.1059 Scale: 171.6149 0.0933 0.0932 0.0936 Add: 218.6231 0.1099 0.1098 0.1101 Triad: 218.9041 0.1105 0.1096 0.1173 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 2 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 2 microseconds. Each test below will take on the order of 78263 microseconds. (= 39131 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.8684 0.0989 0.0959 0.1076 Scale: 172.5849 0.0947 0.0927 0.1002 Add: 218.6032 0.1150 0.1098 0.1357 Triad: 218.3983 0.1162 0.1099 0.1234 3 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 3 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78260 microseconds. (= 78260 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.8961 0.0957 0.0953 0.0979 Scale: 171.9432 0.0932 0.0931 0.0934 Add: 219.4145 0.1094 0.1094 0.1097 Triad: 219.1701 0.1106 0.1095 0.1196 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 3 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78297 microseconds. (= 78297 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.8540 0.0967 0.0953 0.1054 Scale: 171.9913 0.0931 0.0930 0.0931 Add: 219.4988 0.1097 0.1093 0.1116 Triad: 219.2643 0.1096 0.1095 0.1099 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 3 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78301 microseconds. (= 78301 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.8750 0.0991 0.0953 0.1167 Scale: 171.9358 0.0947 0.0931 0.1068 Add: 219.2042 0.1108 0.1095 0.1195 Triad: 219.2882 0.1109 0.1094 0.1216 4 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 4 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78244 microseconds. (= 78244 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.9727 0.0962 0.0958 0.0982 Scale: 173.7035 0.0932 0.0921 0.1021 Add: 221.0717 0.1086 0.1086 0.1087 Triad: 220.0543 0.1092 0.1091 0.1094 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 4 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 2 microseconds. Each test below will take on the order of 78303 microseconds. (= 39151 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.8892 0.0962 0.0959 0.0984 Scale: 173.6922 0.0922 0.0921 0.0924 Add: 220.9964 0.1099 0.1086 0.1187 Triad: 219.8487 0.1094 0.1092 0.1106 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 4 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78293 microseconds. (= 78293 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.9814 0.0961 0.0958 0.0983 Scale: 173.6960 0.0922 0.0921 0.0922 Add: 221.1390 0.1097 0.1085 0.1186 Triad: 220.0160 0.1100 0.1091 0.1176 5 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 5 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78266 microseconds. (= 78266 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.9861 0.0956 0.0952 0.0978 Scale: 170.8525 0.0937 0.0936 0.0937 Add: 217.2830 0.1105 0.1105 0.1107 Triad: 219.0201 0.1107 0.1096 0.1199 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 5 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78204 microseconds. (= 78204 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.9509 0.0966 0.0953 0.1053 Scale: 170.8654 0.0937 0.0936 0.0939 Add: 217.2694 0.1107 0.1105 0.1121 Triad: 219.0622 0.1096 0.1096 0.1098 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 5 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78308 microseconds. (= 78308 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.8803 0.0967 0.0953 0.1054 Scale: 170.9128 0.0936 0.0936 0.0937 Add: 217.2122 0.1106 0.1105 0.1107 Triad: 219.0680 0.1096 0.1096 0.1100 6 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 6 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78306 microseconds. (= 78306 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.0199 0.1148 0.0958 0.1585 Scale: 172.0707 0.1064 0.0930 0.1615 Add: 218.7564 0.1319 0.1097 0.1525 Triad: 219.1601 0.1359 0.1095 0.2018 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 6 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 106654 microseconds. (= 106654 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.9570 0.1048 0.0958 0.1321 Scale: 171.4862 0.1007 0.0933 0.1305 Add: 219.4086 0.1098 0.1094 0.1101 Triad: 219.2643 0.1138 0.1095 0.1436 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 6 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78237 microseconds. (= 78237 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.0669 0.0960 0.0958 0.0982 Scale: 171.7088 0.0943 0.0932 0.1033 Add: 219.0202 0.1097 0.1096 0.1103 Triad: 219.6355 0.1094 0.1093 0.1097 7 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 7 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78284 microseconds. (= 78284 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.0636 0.0955 0.0952 0.0977 Scale: 171.9561 0.1000 0.0930 0.1482 Add: 219.4005 0.1105 0.1094 0.1193 Triad: 219.1281 0.1096 0.1095 0.1096 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 7 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78308 microseconds. (= 78308 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.0990 0.0955 0.0952 0.0977 Scale: 172.0227 0.0932 0.0930 0.0943 Add: 219.5873 0.1105 0.1093 0.1193 Triad: 219.3404 0.1095 0.1094 0.1096 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 7 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78349 microseconds. (= 78349 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.0690 0.0955 0.0952 0.0976 Scale: 171.9764 0.0932 0.0930 0.0934 Add: 219.5270 0.1104 0.1093 0.1194 Triad: 219.1921 0.1097 0.1095 0.1102 8 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 8 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 2 microseconds. Each test below will take on the order of 78245 microseconds. (= 39122 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.0442 0.0961 0.0958 0.0983 Scale: 173.6657 0.0922 0.0921 0.0923 Add: 221.0679 0.1086 0.1086 0.1087 Triad: 220.1511 0.1102 0.1090 0.1191 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 8 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78306 microseconds. (= 78306 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 166.9222 0.0972 0.0959 0.1059 Scale: 173.5320 0.0922 0.0922 0.0924 Add: 220.9376 0.1087 0.1086 0.1089 Triad: 219.9857 0.1093 0.1091 0.1105 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 8 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 79766 microseconds. (= 79766 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.0791 0.0961 0.0958 0.0985 Scale: 173.7505 0.0932 0.0921 0.1022 Add: 221.0637 0.1087 0.1086 0.1088 Triad: 220.2581 0.1091 0.1090 0.1092 9 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 9 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78331 microseconds. (= 78331 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.0144 0.0955 0.0952 0.0976 Scale: 170.7833 0.0944 0.0937 0.1003 Add: 217.1415 0.1118 0.1105 0.1211 Triad: 219.0100 0.1097 0.1096 0.1098 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 9 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78234 microseconds. (= 78234 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.1361 0.0955 0.0952 0.0977 Scale: 171.0151 0.0938 0.0936 0.0954 Add: 217.2772 0.1116 0.1105 0.1204 Triad: 219.2541 0.1096 0.1095 0.1098 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 9 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78324 microseconds. (= 78324 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.0549 0.0955 0.0952 0.0977 Scale: 170.8052 0.0937 0.0937 0.0938 Add: 217.0707 0.1117 0.1106 0.1207 Triad: 218.8382 0.1098 0.1097 0.1101 10 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 10 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 88211 microseconds. (= 88211 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.1646 0.0962 0.0957 0.0982 Scale: 171.6241 0.0934 0.0932 0.0943 Add: 218.7526 0.1101 0.1097 0.1117 Triad: 219.6033 0.1101 0.1093 0.1149 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 10 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78329 microseconds. (= 78329 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.0860 0.0982 0.0958 0.1082 Scale: 171.1908 0.1021 0.0935 0.1140 Add: 218.6050 0.1123 0.1098 0.1178 Triad: 219.0221 0.1184 0.1096 0.1387 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 10 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 94522 microseconds. (= 94522 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.1385 0.0961 0.0957 0.0981 Scale: 171.7089 0.0933 0.0932 0.0943 Add: 218.7924 0.1098 0.1097 0.1100 Triad: 219.5449 0.1094 0.1093 0.1099 11 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 11 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78339 microseconds. (= 78339 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.0530 0.0956 0.0952 0.0977 Scale: 171.8767 0.0932 0.0931 0.0933 Add: 219.4465 0.1105 0.1094 0.1195 Triad: 219.3644 0.1095 0.1094 0.1097 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 11 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78215 microseconds. (= 78215 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.3396 0.0960 0.0950 0.1019 Scale: 172.0503 0.0931 0.0930 0.0931 Add: 219.7340 0.1103 0.1092 0.1193 Triad: 219.6574 0.1095 0.1093 0.1113 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 11 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78300 microseconds. (= 78300 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.1396 0.0955 0.0952 0.0977 Scale: 172.0079 0.0932 0.0930 0.0935 Add: 219.4988 0.1094 0.1093 0.1095 Triad: 219.2722 0.1106 0.1095 0.1194 12 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 12 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78170 microseconds. (= 78170 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.1944 0.0971 0.0957 0.1057 Scale: 173.7223 0.0922 0.0921 0.0924 Add: 221.1919 0.1086 0.1085 0.1089 Triad: 220.1754 0.1090 0.1090 0.1091 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 12 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78320 microseconds. (= 78320 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.0321 0.0961 0.0958 0.0982 Scale: 173.4812 0.0933 0.0922 0.1023 Add: 220.7667 0.1088 0.1087 0.1090 Triad: 220.0139 0.1094 0.1091 0.1111 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 12 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78158 microseconds. (= 78158 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.2031 0.1034 0.0957 0.1314 Scale: 173.7035 0.0952 0.0921 0.1193 Add: 221.0963 0.1163 0.1086 0.1452 Triad: 220.2885 0.1376 0.1089 0.2048 13 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 13 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 114877 microseconds. (= 114877 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.1432 0.1091 0.0952 0.1323 Scale: 170.8872 0.0995 0.0936 0.1324 Add: 217.0432 0.1201 0.1106 0.1486 Triad: 219.1621 0.1214 0.1095 0.1467 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 13 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78174 microseconds. (= 78174 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.4193 0.0953 0.0950 0.0975 Scale: 170.9951 0.0936 0.0936 0.0938 Add: 217.3696 0.1117 0.1104 0.1205 Triad: 219.4786 0.1095 0.1094 0.1099 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 13 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78277 microseconds. (= 78277 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.2068 0.0955 0.0951 0.0976 Scale: 170.7941 0.0940 0.0937 0.0947 Add: 217.2613 0.1127 0.1105 0.1283 Triad: 219.0661 0.1100 0.1096 0.1107 14 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 14 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78366 microseconds. (= 78366 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.3326 0.0968 0.0956 0.1041 Scale: 171.6996 0.0955 0.0932 0.1137 Add: 218.8283 0.1130 0.1097 0.1229 Triad: 219.5529 0.1113 0.1093 0.1183 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 14 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 2 microseconds. Each test below will take on the order of 78248 microseconds. (= 39124 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.2224 0.0997 0.0957 0.1231 Scale: 171.4532 0.0935 0.0933 0.0950 Add: 218.6311 0.1118 0.1098 0.1198 Triad: 219.3865 0.1122 0.1094 0.1260 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 14 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78245 microseconds. (= 78245 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 167.3377 0.0960 0.0956 0.0982 Scale: 171.6683 0.0942 0.0932 0.1022 Add: 218.8822 0.1108 0.1096 0.1197 Triad: 219.6494 0.1095 0.1093 0.1098 15 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 15 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 2 microseconds. Each test below will take on the order of 78638 microseconds. (= 39319 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.0231 0.1702 0.0952 0.4075 Scale: 171.6923 0.1381 0.0932 0.2181 Add: 218.6170 0.1576 0.1098 0.2558 Triad: 219.4064 0.1603 0.1094 0.2669 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 15 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 77591 microseconds. (= 77591 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.9903 0.0952 0.0947 0.0972 Scale: 172.7339 0.1263 0.0926 0.2862 Add: 219.9795 0.1096 0.1091 0.1116 Triad: 219.9132 0.1265 0.1091 0.2288 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 15 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 78082 microseconds. (= 78082 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 168.6661 0.0952 0.0949 0.0972 Scale: 172.3228 0.0940 0.0928 0.1032 Add: 219.5630 0.1099 0.1093 0.1108 Triad: 219.6917 0.1097 0.1092 0.1105 From john@mccalpin.com Wed Apr 28 18:38:48 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id SAA17041 for ; Wed, 28 Apr 1999 18:38:48 -0700 (PDT) Received: from mccloud.engr.sgi.com (mccloud.engr.sgi.com [192.48.196.40]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id SAA83130 for ; Wed, 28 Apr 1999 18:38:48 -0700 (PDT) mail_from (guest@mccloud.engr.sgi.com) Received: (from guest@localhost) by mccloud.engr.sgi.com (8.8.7/8.8.7) id SAA31516 for mccalpin@engr.sgi.com; Wed, 28 Apr 1999 18:37:04 -0700 Date: Wed, 28 Apr 1999 18:37:04 -0700 From: John McCalpin Message-Id: <199904290137.SAA31516@mccloud.engr.sgi.com> To: mccalpin@cthulhu Subject: STREAM on cloudburst Status: RO ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39100 microseconds. (= 39100 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 322.2560 0.0499 0.0496 0.0515 Scale: 323.2520 0.0495 0.0495 0.0497 Add: 366.9443 0.0655 0.0654 0.0657 Triad: 365.4525 0.0657 0.0657 0.0658 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39051 microseconds. (= 39051 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 321.8540 0.0499 0.0497 0.0515 Scale: 323.4346 0.0495 0.0495 0.0495 Add: 366.7367 0.0655 0.0654 0.0655 Triad: 365.4358 0.0658 0.0657 0.0658 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39177 microseconds. (= 39177 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 322.1455 0.0499 0.0497 0.0514 Scale: 323.2844 0.0495 0.0495 0.0496 Add: 366.8883 0.0654 0.0654 0.0655 Triad: 365.2578 0.0658 0.0657 0.0660 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39055 microseconds. (= 39055 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 321.7440 0.0499 0.0497 0.0515 Scale: 323.2783 0.0495 0.0495 0.0496 Add: 366.7092 0.0655 0.0654 0.0655 Triad: 364.9525 0.0658 0.0658 0.0659 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39099 microseconds. (= 39099 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 322.2621 0.0500 0.0496 0.0515 Scale: 323.2062 0.0495 0.0495 0.0496 Add: 366.8659 0.0655 0.0654 0.0655 Triad: 365.2911 0.0657 0.0657 0.0658 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39068 microseconds. (= 39068 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 321.9448 0.0499 0.0497 0.0515 Scale: 323.2588 0.0495 0.0495 0.0496 Add: 366.9218 0.0654 0.0654 0.0655 Triad: 365.1909 0.0658 0.0657 0.0658 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39059 microseconds. (= 39059 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 322.2043 0.0499 0.0497 0.0515 Scale: 323.3173 0.0495 0.0495 0.0498 Add: 366.8437 0.0655 0.0654 0.0655 Triad: 365.2747 0.0658 0.0657 0.0658 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39053 microseconds. (= 39053 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 321.8989 0.0499 0.0497 0.0515 Scale: 323.1997 0.0495 0.0495 0.0496 Add: 366.7988 0.0655 0.0654 0.0655 Triad: 365.3916 0.0657 0.0657 0.0658 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39044 microseconds. (= 39044 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 322.3466 0.0499 0.0496 0.0515 Scale: 323.4610 0.0495 0.0495 0.0496 Add: 366.8153 0.0655 0.0654 0.0657 Triad: 365.4470 0.0657 0.0657 0.0658 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 0 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 39054 microseconds. (= 39054 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 322.0612 0.0499 0.0497 0.0515 Scale: 323.1672 0.0495 0.0495 0.0496 Add: 366.8101 0.0655 0.0654 0.0655 Triad: 365.2246 0.0658 0.0657 0.0658 From kcollins@rsn.hp.com Wed May 12 11:41:15 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id LAA43002 for ; Wed, 12 May 1999 11:40:54 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id LAA31869 for ; Wed, 12 May 1999 11:40:53 -0700 (PDT) mail_from (kcollins@rsn.hp.com) Received: from ares.cs.Virginia.EDU (mail.cs.Virginia.EDU [128.143.137.19]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id LAA01120 for ; Wed, 12 May 1999 11:40:47 -0700 (PDT) mail_from (kcollins@rsn.hp.com) Received: from atlrel2.hp.com (atlrel2.hp.com [156.153.255.202]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id OAA09505 for ; Wed, 12 May 1999 14:40:45 -0400 (EDT) Received: from cmail2.rsn.hp.COM (cmail2.rsn.hp.com [15.99.101.1]) by atlrel2.hp.com (8.8.6 (PHNE_17135)/8.8.5tis) with ESMTP id OAA18203 for ; Wed, 12 May 1999 14:40:39 -0400 (EDT) Received: from kirbyob.rsn.hp.com (kcollinspc.rsn.hp.com [15.99.100.197]) by cmail2.rsn.hp.COM with SMTP (8.7.1/8.7.1) id NAA29469; Wed, 12 May 1999 13:40:37 -0500 (CDT) Message-Id: <3.0.3.32.19990512134040.00883380@cmail2.rsn.hp.com> X-Sender: kcollins@cmail2.rsn.hp.com X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32) Date: Wed, 12 May 1999 13:40:40 -0500 To: mccalpin@cs.virginia.edu From: Kirby Collins Subject: stream results for HP N4000, 1 cpu Cc: isom@cmail2.rsn.hp.COM, kcollins@cmail2.rsn.hp.COM, koester@cmail2.rsn.hp.COM Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Status: RO run on a prerelease of HP-UX 11.0 March 1999 Extension Pack compiled with HP Fortran 90 Compiler B.11.01.06 f90 -o stream_d.mp +DA2.0 +DS2.0 +O3 +Odataprefetch +FPD -Wl,+pd,64M +Oparallel stream_d.f second_wall.o ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 20000000 Offset = 0 The total memory requirement is 457 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 3 microseconds The tests below will each take a time on the order of 271725 microseconds (= 90575 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 759.9506 0.4212 0.4211 0.4216 Scale: 758.6967 0.4220 0.4218 0.4226 Add: 834.5851 0.5756 0.5751 0.5770 Triad: 836.1392 0.5743 0.5741 0.5749 Sum of a is = 2.306601562980265E+19 Sum of b is = 4.613203126278327E+18 Sum of c is = 6.150937498077968E+18 From kcollins@rsn.hp.com Wed May 12 13:36:33 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id NAA43453 for ; Wed, 12 May 1999 13:36:21 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id NAA26371 for ; Wed, 12 May 1999 13:36:19 -0700 (PDT) mail_from (kcollins@rsn.hp.com) Received: from ares.cs.Virginia.EDU (ares.cs.Virginia.EDU [128.143.136.41] (may be forged)) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id NAA07594 for ; Wed, 12 May 1999 13:36:19 -0700 (PDT) mail_from (kcollins@rsn.hp.com) Received: from atlrel2.hp.com (atlrel2.hp.com [156.153.255.202]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id QAA12993 for ; Wed, 12 May 1999 16:36:16 -0400 (EDT) Received: from cmail2.rsn.hp.COM (cmail2.rsn.hp.com [15.99.101.1]) by atlrel2.hp.com (8.8.6 (PHNE_17135)/8.8.5tis) with ESMTP id QAA01089 for ; Wed, 12 May 1999 16:36:14 -0400 (EDT) Received: from kirbyob.rsn.hp.com (kcollinspc.rsn.hp.com [15.99.100.197]) by cmail2.rsn.hp.COM with SMTP (8.7.1/8.7.1) id PAA04206; Wed, 12 May 1999 15:36:11 -0500 (CDT) Message-Id: <3.0.3.32.19990512153614.007f9dd0@cmail2.rsn.hp.com> X-Sender: kcollins@cmail2.rsn.hp.com X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32) Date: Wed, 12 May 1999 15:36:14 -0500 To: mccalpin@cs.virginia.edu From: Kirby Collins Subject: HP V2500 stream results, 1 node, 1 cpu Cc: kcollins@cmail2.rsn.hp.COM, isom@cmail2.rsn.hp.COM, koester@cmail2.rsn.hp.COM Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Status: RO ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 20000000 Offset = 0 The total memory requirement is 457 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 3 microseconds The tests below will each take a time on the order of 597567 microseconds (= 199189 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 371.4137 0.9370 0.8616 0.9459 Scale: 366.9035 0.9138 0.8722 0.9194 Add: 346.9886 1.3841 1.3833 1.3849 Triad: 355.6894 1.3502 1.3495 1.3511 Sum of a is = 2.306601562980265E+19 Sum of b is = 4.613203126278327E+18 Sum of c is = 6.150937498077968E+18 From kcollins@rsn.hp.com Wed May 12 13:36:57 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id NAA43523 for ; Wed, 12 May 1999 13:36:44 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id NAA33508 for ; Wed, 12 May 1999 13:36:43 -0700 (PDT) mail_from (kcollins@rsn.hp.com) Received: from ares.cs.Virginia.EDU (ares.cs.Virginia.EDU [128.143.136.41] (may be forged)) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id NAA05903 for ; Wed, 12 May 1999 13:36:42 -0700 (PDT) mail_from (kcollins@rsn.hp.com) Received: from atlrel2.hp.com (atlrel2.hp.com [156.153.255.202]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id QAA13001 for ; Wed, 12 May 1999 16:36:41 -0400 (EDT) Received: from cmail2.rsn.hp.COM (cmail2.rsn.hp.com [15.99.101.1]) by atlrel2.hp.com (8.8.6 (PHNE_17135)/8.8.5tis) with ESMTP id QAA01264 for ; Wed, 12 May 1999 16:36:38 -0400 (EDT) Received: from kirbyob.rsn.hp.com (kcollinspc.rsn.hp.com [15.99.100.197]) by cmail2.rsn.hp.COM with SMTP (8.7.1/8.7.1) id PAA04222; Wed, 12 May 1999 15:36:37 -0500 (CDT) Message-Id: <3.0.3.32.19990512153641.00851c70@cmail2.rsn.hp.com> X-Sender: kcollins@cmail2.rsn.hp.com X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32) Date: Wed, 12 May 1999 15:36:41 -0500 To: mccalpin@cs.virginia.edu From: Kirby Collins Subject: HP V2500 Stream results, 1 node, 32 cpu Cc: kcollins@cmail2.rsn.hp.COM, isom@cmail2.rsn.hp.COM, koester@cmail2.rsn.hp.COM Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Status: RO ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 20000000 Offset = 0 The total memory requirement is 457 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 3 microseconds The tests below will each take a time on the order of 58322 microseconds (= 19441 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 3934.3020 0.0996 0.0813 0.1022 Scale: 4193.2557 0.0964 0.0763 0.0989 Add: 3926.3165 0.1321 0.1223 0.1947 Triad: 4027.4864 0.1251 0.1192 0.1626 Sum of a is = 2.306601562980265E+19 Sum of b is = 4.613203126278327E+18 Sum of c is = 6.150937498077968E+18 From henning@perfom.zko.dec.com Wed May 19 08:35:17 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id IAA55757 for ; Wed, 19 May 1999 08:35:05 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id IAA47025 for ; Wed, 19 May 1999 08:35:04 -0700 (PDT) mail_from (henning@perfom.zko.dec.com) Received: from mail11.digital.com (mail11.digital.com [192.208.46.10]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id IAA08206 for ; Wed, 19 May 1999 08:35:00 -0700 (PDT) mail_from (henning@perfom.zko.dec.com) Received: from perfom (perfom.zko.dec.com [16.31.64.90]) by mail11.digital.com (8.9.2/8.9.2/WV2.0g) with SMTP id LAA10085 for ; Wed, 19 May 1999 11:37:59 -0400 (EDT) Date: Wed, 19 May 1999 11:34:24 -0400 Message-Id: <99051911342465@perfom.zko.dec.com> From: henning@perfom.zko.dec.com (John Henning) To: MCCALPIN Subject: AlphaServer ES40 X-VMS-To: MCCALPIN@FRAKIR.ENGR.SGI.COM X-VMS-Cc: ME Status: RO Hi John, Here's some more data for your chart. This is the AlphaServer ES40 (1 cpu) If you could post it soon, I'd be grateful! Best regards - John Script started on Thu May 13 10:35:06 1999 $ ls rpcc.c tryit.sh stream_d.f_as_at_ftp_site_22may97 typescript tryit.f $ diff stream_d.f_as_at_ftp_site_22may97 tryit.f^R diff stream_d.f_as_at_ftp_site_22may97 tryit.f 53c53,54 < PARAMETER (n=2000000,offset=0,ndim=n+offset,ntimes=10) --- > PARAMETER (n = 1000000 ,offset=0) > PARAMETER (ndim=n+offset,ntimes=10) 334a336,359 > END > > DOUBLE PRECISION FUNCTION SECOND > * > * Assume we will be called often enough so that rpcc counter > * doesn't overflow! > * > INTEGER*8 TICK, PREV_TICK > DOUBLE PRECISION RPCC_DELTA, PREV_SEC > INTEGER*8 FIXED_RPCC > EXTERNAL FIXED_RPCC, RPCC_DELTA > > DATA PREV_TICK /0/ > > TICK = FIXED_RPCC() > IF (PREV_TICK .EQ. 0) THEN > SECOND = 0D0 > ELSE > SECOND = PREV_SEC + RPCC_DELTA (%VAL(PREV_TICK), %VAL(TICK)) > END IF > > PREV_SEC = SECOND > PREV_TICK = TICK > RETURN $ $ cat rpcc.c /* rpcc.c j.henning hacked from example by d.grunwald*/ #include #include #include #include #include #include static double scale = -1; static pid_t mypid = -1; /* fix_rpcc adds the two halves of the register (bias & offset), since some tools return it w/o this fixup */ unsigned long fix_rpcc(unsigned long rpcc) { unsigned long fixed; unsigned long offset = (rpcc >> 32) & 0xffffffff; unsigned long cycles = (rpcc & 0xffffffff); fixed = (unsigned long) offset + (unsigned long) cycles; /* printf ("%8lx %8lx %8lx ", offset, cycles, fixed); */ return fixed; } /* Compute time in *seconds* (note that Grunwald's original version did microseconds) given two fixedup RPCCs */ double rpcc_delta(unsigned long ul_start, unsigned long ul_stop) { double retval; if ( scale == -1 ) { int err; double freq; struct cpu_info cpubuf ; err = getsysinfo(GSI_CPU_INFO, &cpubuf, sizeof(cpubuf)); /*printf ("err= %d freq = %d current_cpu = %d\n", err, cpubuf.mhz, cpubuf.current_cpu);*/ if (err < 0) { freq = 300; } else { freq = cpubuf.mhz; } freq *= 1000000; /* * Scale is the 1 over the clock frequency */ scale = 1/(double) freq; } if (ul_stop < ul_start) { retval = ((0xffffffff - ul_start) + ul_stop) * scale; } else { retval = (ul_stop - ul_start) * scale; } /* printf ("%f\n", retval); */ return retval; } unsigned long fixed_rpcc(void) { unsigned long i; i = asm("rpcc %v0"); return fix_rpcc(i); } $ $ cat tryit.sh # Print compiler version echo "" echo "" echo -n " " echo " end" >tmp$$.f f77 -V -c tmp$$.f grep "COMPILER:" tmp$$.l | cut -b 10-99 rm tmp$$.l tmp$$.o tmp$$.f echo "" echo "" cc -c rpcc.c f77 -non_shared -fast -O5 -unroll 8 -arch ev6 -assume nounderscore rpcc.o tryit.f ./a.out $ $ ./tryit.sh -n DIGITAL Fortran 77 V5.2-171-428BH ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 1000000 Offset = 0 The total memory requirement is 22 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 11945 microseconds (= 11945 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 1004.5576 0.0160 0.0159 0.0163 Scale: 1193.8713 0.0134 0.0134 0.0135 Add: 1058.6282 0.0227 0.0227 0.0228 Triad: 1237.0523 0.0194 0.0194 0.0195 Sum of a is = 1.153300781280042E+018 Sum of b is = 2.306601562536570E+017 Sum of c is = 3.075468750061256E+017 $ exit script done on Thu May 13 10:35:47 1999 From kcollins@rsn.hp.com Wed May 19 11:45:02 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id LAA55970 for ; Wed, 19 May 1999 11:44:49 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id LAA78399 for ; Wed, 19 May 1999 11:44:48 -0700 (PDT) mail_from (kcollins@rsn.hp.com) Received: from ares.cs.Virginia.EDU (ftp.cs.Virginia.EDU [128.143.137.19]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id LAA05359 for ; Wed, 19 May 1999 11:44:44 -0700 (PDT) mail_from (kcollins@rsn.hp.com) Received: from atlrel2.hp.com (atlrel2.hp.com [156.153.255.202]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id OAA22546 for ; Wed, 19 May 1999 14:44:42 -0400 (EDT) Received: from cmail2.rsn.hp.COM (cmail2.rsn.hp.com [15.99.101.1]) by atlrel2.hp.com (8.8.6 (PHNE_17135)/8.8.5tis) with ESMTP id OAA24354 for ; Wed, 19 May 1999 14:44:39 -0400 (EDT) Received: from kirbyob.rsn.hp.com (kcollinspc.rsn.hp.com [15.99.100.197]) by cmail2.rsn.hp.COM with SMTP (8.7.1/8.7.1) id NAA13437 for ; Wed, 19 May 1999 13:44:38 -0500 (CDT) Message-Id: <3.0.3.32.19990519134439.00894940@cmail2.rsn.hp.com> X-Sender: kcollins@cmail2.rsn.hp.com X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32) Date: Wed, 19 May 1999 13:44:39 -0500 To: mccalpin@cs.virginia.edu From: Kirby Collins Subject: stream results for HP N4000, 8 way parallel Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Status: RO run on a prerelease of HP-UX 11.0 March 1999 Extension Pack compiled with HP Fortran 90 Compiler B.11.01.06 f90 -o stream_d.mp +DA2.0 +DS2.0 +O3 +Odataprefetch +FPD -Wl,+pd,64M +Oparallel stream_d.f second_wall.o ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 20000000 Offset = 0 The total memory requirement is 457 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 3 microseconds The tests below will each take a time on the order of 113757 microseconds (= 37919 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 1759.0928 0.1852 0.1819 0.2080 Scale: 1755.9066 0.1828 0.1822 0.1837 Add: 1917.2391 0.2508 0.2504 0.2520 Triad: 1919.6778 0.2503 0.2500 0.2507 Sum of a is = 2.306601562980265E+19 Sum of b is = 4.613203126278327E+18 Sum of c is = 6.150937498077968E+18 From henning@perfom.zko.dec.com Mon May 24 06:36:13 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id GAA64827 for ; Mon, 24 May 1999 06:36:02 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id GAA33200 for ; Mon, 24 May 1999 06:36:01 -0700 (PDT) mail_from (henning@perfom.zko.dec.com) Received: from mail11.digital.com (mail11.digital.com [192.208.46.10]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id GAA06793 for ; Mon, 24 May 1999 06:36:00 -0700 (PDT) mail_from (henning@perfom.zko.dec.com) Received: from perfom (perfom.zko.dec.com [16.31.64.90]) by mail11.digital.com (8.9.2/8.9.2/WV2.0g) with SMTP id JAA06345 for ; Mon, 24 May 1999 09:39:08 -0400 (EDT) Date: Mon, 24 May 1999 09:34:46 -0400 Message-Id: <99052409344660@perfom.zko.dec.com> From: henning@perfom.zko.dec.com (John Henning) To: MCCALPIN Subject: XP1000 Streams results X-VMS-To: MCCALPIN@FRAKIR.ENGR.SGI.COM X-VMS-Cc: ME Status: RO Hi John, Here's the results for the XP1000. The submission was long delayed, as you can see from the date; that was my fault. Thanks! - john Script started on Tue Mar 2 08:16:45 1999 % ls *ftp* stream_d.f_as_at_ftp_site_22may97 % diff *ftp* tryit.f 53c53,54 < PARAMETER (n=2000000,offset=0,ndim=n+offset,ntimes=10) --- > PARAMETER (n = 1000020,offset=0) > PARAMETER (ndim=n+offset,ntimes=10) 334a336,359 > END > > DOUBLE PRECISION FUNCTION SECOND > * > * Assume we will be called often enough so that rpcc counter > * doesn't overflow! > * > INTEGER*8 TICK, PREV_TICK > DOUBLE PRECISION RPCC_DELTA, PREV_SEC > INTEGER*8 FIXED_RPCC > EXTERNAL FIXED_RPCC, RPCC_DELTA > > DATA PREV_TICK /0/ > > TICK = FIXED_RPCC() > IF (PREV_TICK .EQ. 0) THEN > SECOND = 0D0 > ELSE > SECOND = PREV_SEC + RPCC_DELTA (%VAL(PREV_TICK), %VAL(TICK)) > END IF > > PREV_SEC = SECOND > PREV_TICK = TICK > RETURN % cat rpcc.c /* rpcc.c j.henning hacked from example by d.grunwald*/ #include #include #include #include #include #include static double scale = -1; static pid_t mypid = -1; /* fix_rpcc adds the two halves of the register (bias & offset), since some tools return it w/o this fixup */ unsigned long fix_rpcc(unsigned long rpcc) { unsigned long fixed; unsigned long offset = (rpcc >> 32) & 0xffffffff; unsigned long cycles = (rpcc & 0xffffffff); fixed = (unsigned long) offset + (unsigned long) cycles; /* printf ("%8lx %8lx %8lx ", offset, cycles, fixed); */ return fixed; } /* Compute time in *seconds* (note that Grunwald's original version did microseconds) given two fixedup RPCCs */ double rpcc_delta(unsigned long ul_start, unsigned long ul_stop) { double retval; if ( scale == -1 ) { int err; double freq; struct cpu_info cpubuf ; err = getsysinfo(GSI_CPU_INFO, &cpubuf, sizeof(cpubuf)); /*printf ("err= %d freq = %d current_cpu = %d\n", err, cpubuf.mhz, cpubuf.current_cpu);*/ if (err < 0) { freq = 300; } else { freq = cpubuf.mhz; } freq *= 1000000; /* * Scale is the 1 over the clock frequency */ scale = 1/(double) freq; } if (ul_stop < ul_start) { retval = ((0xffffffff - ul_start) + ul_stop) * scale; } else { retval = (ul_stop - ul_start) * scale; } /* printf ("%f\n", retval); */ return retval; } unsigned long fixed_rpcc(void) { unsigned long i; i = asm("rpcc %v0"); return fix_rpcc(i); } % cc -c rpcc.c % f77 -non_shared -fast -O5 -unroll 8 -arch ev6 -assume nounderscore -V rpcc.o tryit.f % grep COMPILER: tryit.l COMPILER: DIGITAL Fortran 77 V5.2-171-428BH % ./a.out ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 1000020 Offset = 0 The total memory requirement is 22 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 12706 microseconds (= 12706 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 900.2357 0.0179 0.0178 0.0192 Scale: 971.1877 0.0166 0.0165 0.0172 Add: 917.2598 0.0262 0.0262 0.0265 Triad: 967.7494 0.0248 0.0248 0.0250 Sum of a is = 1.153323847295668E+018 Sum of b is = 2.306647694567821E+017 Sum of c is = 3.075530259436257E+017 % exit % script done on Tue Mar 2 08:17:29 1999 From henning@perfom.zko.dec.com Wed May 26 07:15:36 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/960327.SGI.AUTOCF) via ESMTP id HAA69179 for ; Wed, 26 May 1999 07:15:25 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id HAA95703 for ; Wed, 26 May 1999 07:15:24 -0700 (PDT) mail_from (henning@perfom.zko.dec.com) Received: from mail13.digital.com (mail13.digital.com [192.208.46.30]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id HAA01703 for ; Wed, 26 May 1999 07:15:21 -0700 (PDT) mail_from (henning@perfom.zko.dec.com) Received: from perfom (perfom.zko.dec.com [16.31.64.90]) by mail13.digital.com (8.9.2/8.9.2/WV2.0g) with SMTP id KAA00248 for ; Wed, 26 May 1999 10:15:20 -0400 (EDT) Date: Wed, 26 May 1999 10:15:04 -0400 Message-Id: <99052610150494@perfom.zko.dec.com> From: henning@perfom.zko.dec.com (John Henning) To: MCCALPIN Subject: AlphaServer DS10 X-VMS-To: MCCALPIN@FRAKIR.ENGR.SGI.COM X-VMS-Cc: ME Status: RO Hi John, Here's the AlphaServer DS10. Please add this (and the XP1000, sent earlier in the week) to your page? Thanks - john PS. I had encouraged HP to send you results for the N4000. Glad to see they did. Who was the submitter? Script started on Fri May 21 12:23:22 1999 % cat tryit.csh #!/bin/csh set verbose unlimit cat rpcc.c cc -c rpcc.c sleep 2 diff stream_d.f_as_at_ftp_site_22may97 tryit.f sleep 2 f90 -non_shared -fast -O5 -unroll 8 -arch ev6 \ -assume nounderscore rpcc.o tryit.f sleep 2 ./a.out % ./tryit.csh unlimit cat rpcc.c /* rpcc.c j.henning hacked from example by d.grunwald*/ #include #include #include #include #include #include static double scale = -1; static pid_t mypid = -1; /* fix_rpcc adds the two halves of the register (bias & offset), since some tools return it w/o this fixup */ unsigned long fix_rpcc(unsigned long rpcc) { unsigned long fixed; unsigned long offset = (rpcc >> 32) & 0xffffffff; unsigned long cycles = (rpcc & 0xffffffff); fixed = (unsigned long) offset + (unsigned long) cycles; /* printf ("%8lx %8lx %8lx ", offset, cycles, fixed); */ return fixed; } /* Compute time in *seconds* (note that Grunwald's original version did microseconds) given two fixedup RPCCs */ double rpcc_delta(unsigned long ul_start, unsigned long ul_stop) { double retval; if ( scale == -1 ) { int err; double freq; struct cpu_info cpubuf ; err = getsysinfo(GSI_CPU_INFO, &cpubuf, sizeof(cpubuf)); /*printf ("err= %d freq = %d current_cpu = %d\n", err, cpubuf.mhz, cpubuf.current_cpu);*/ if (err < 0) { freq = 300; } else { freq = cpubuf.mhz; } freq *= 1000000; /* * Scale is the 1 over the clock frequency */ scale = 1/(double) freq; } if (ul_stop < ul_start) { retval = ((0xffffffff - ul_start) + ul_stop) * scale; } else { retval = (ul_stop - ul_start) * scale; } /* printf ("%f\n", retval); */ return retval; } unsigned long fixed_rpcc(void) { unsigned long i; i = asm("rpcc %v0"); return fix_rpcc(i); } cc -c rpcc.c sleep 2 diff stream_d.f_as_at_ftp_site_22may97 tryit.f 53c53,54 < PARAMETER (n=2000000,offset=0,ndim=n+offset,ntimes=10) --- > PARAMETER (n = 1000063 ,offset=0) > PARAMETER (ndim=n+offset,ntimes=10) 334a336,359 > END > > DOUBLE PRECISION FUNCTION SECOND > * > * Assume we will be called often enough so that rpcc counter > * doesn't overflow! > * > INTEGER*8 TICK, PREV_TICK > DOUBLE PRECISION RPCC_DELTA, PREV_SEC > INTEGER*8 FIXED_RPCC > EXTERNAL FIXED_RPCC, RPCC_DELTA > > DATA PREV_TICK /0/ > > TICK = FIXED_RPCC() > IF (PREV_TICK .EQ. 0) THEN > SECOND = 0D0 > ELSE > SECOND = PREV_SEC + RPCC_DELTA (%VAL(PREV_TICK), %VAL(TICK)) > END IF > > PREV_SEC = SECOND > PREV_TICK = TICK > RETURN sleep 2 f90 -non_shared -fast -O5 -unroll 8 -arch ev6 -assume nounderscore rpcc.o tryit.f sleep 2 ./a.out ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 1000063 Offset = 0 The total memory requirement is 22 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 16817 microseconds (= 16817 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 881.1317 0.0185 0.0182 0.0197 Scale: 929.8878 0.0174 0.0172 0.0174 Add: 839.8763 0.0289 0.0286 0.0290 Triad: 900.9164 0.0268 0.0266 0.0279 Sum of a is = 1.153373439229264E+018 Sum of b is = 2.306746878435012E+017 Sum of c is = 3.075662504592511E+017 % exit % script done on Fri May 21 12:23:43 1999 From Brian.Whitney@West.Sun.COM Fri Jun 11 15:50:36 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via ESMTP id PAA07389 for ; Fri, 11 Jun 1999 15:50:35 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id PAA42576 for ; Fri, 11 Jun 1999 15:50:34 -0700 (PDT) mail_from (Brian.Whitney@West.Sun.COM) Received: from mercury.Sun.COM (mercury.Sun.COM [192.9.25.1]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id PAA05191 for ; Fri, 11 Jun 1999 15:50:25 -0700 (PDT) mail_from (Brian.Whitney@West.Sun.COM) Received: from westmail2.West.Sun.COM ([129.153.100.30]) by mercury.Sun.COM (8.9.3+Sun/8.9.3) with ESMTP id PAA07966 for ; Fri, 11 Jun 1999 15:50:20 -0700 (PDT) Received: from oregon.West.Sun.COM (oregon.West.Sun.COM [129.153.37.10]) by westmail2.West.Sun.COM (8.9.1b+Sun/8.9.1/ENSMAIL,v1.6) with SMTP id PAA11567 for ; Fri, 11 Jun 1999 15:50:18 -0700 (PDT) Received: from sailfish by oregon.West.Sun.COM (SMI-8.6/SMI-SVR4) id PAA09872; Fri, 11 Jun 1999 15:50:18 -0700 Message-Id: <199906112250.PAA09872@oregon.West.Sun.COM> Date: Fri, 11 Jun 1999 15:50:18 -0700 (PDT) From: Brian Whitney Reply-To: Brian Whitney Subject: Stream Results - Sun E10000 To: mccalpin Cc: Brian.Whitney@West.Sun.COM MIME-Version: 1.0 Content-Type: MULTIPART/mixed; BOUNDARY=Family_of_Sardines_330_000 X-Mailer: dtmail 1.3.0 CDE Version 1.3 SunOS 5.7 sun4u sparc Status: RO --Family_of_Sardines_330_000 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: cVQFHksMZKIb3sV8GWsjfQ== Dear Dr. McCalpin, Please find attached STREAM runs on the Sun Enterprise 10000. I modified the code to flush the cache between kernels. Without this modification, the results were faster than the interconnect supports, because of caching. The copy results are faster than the other three because of a compiler flag which allows us to bypass the cache, allowing for faster bcopy performance. I have made a family of runs, a set on a 16 board system, a set on a 12 board system and a set on an 8 board system. Results are reported for 1, 16, 32, 48, and 64 CPUs (as appropriate). File names are mod.stream.NNonB where NN is the number of cpus used B is the number of boards in the system. Each board has 4 GBs of memory on it, 4 400 MHz, 8MB l2$ UltraSPARC II cpus. Thanks, Brian Brian Whitney Sun Microsystems, Inc. Strategic Applications Engineering 8300 SW Creekside Pl. office: (503) 520-7616 Beaverton, OR 97008 fax: (503) 520-7724 e-mail: Brian.Whitney@west.sun.com --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.1on16"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.1on16 Content-MD5: s4yG+2+VgjjyhJxjXdblIA== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 31602 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 3453153 microseconds (= 3453153 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 364.1918 2.2020 2.1966 2.2072 Scale: 215.3819 3.7150 3.7143 3.7160 Add: 287.4328 4.1755 4.1749 4.1763 Triad: 296.0569 4.0539 4.0533 4.0557 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.16on16"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.16on16 Content-MD5: 4Jv/AkDTV0BXRjKteWVYuA== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 24002 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 234538 microseconds (= 234538 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 4925.4124 0.1629 0.1624 0.1648 Scale: 3030.9344 0.2643 0.2639 0.2646 Add: 3703.3682 0.3244 0.3240 0.3247 Triad: 3661.7034 0.3279 0.3277 0.3283 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.32on16"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.32on16 Content-MD5: QwOOLOlES7WsaKpy2UAE8A== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 16102 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 124946 microseconds (= 124946 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 9053.1679 0.0892 0.0884 0.0931 Scale: 5176.1229 0.1548 0.1546 0.1549 Add: 6260.2872 0.1919 0.1917 0.1922 Triad: 6301.5355 0.1905 0.1904 0.1906 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.48on16"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.48on16 Content-MD5: js05ssCIHGsPylQjXc/7HQ== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 8002 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 70413 microseconds (= 70413 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 12056.9571 0.0672 0.0664 0.0731 Scale: 6356.9620 0.1262 0.1258 0.1266 Add: 7658.5054 0.1570 0.1567 0.1576 Triad: 7572.8558 0.1587 0.1585 0.1590 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.64on16"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.64on16 Content-MD5: BWHH2E4LuDuEZyaO1JsT5A== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 102 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 231511 microseconds (= 231511 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 12430.3516 0.0721 0.0644 0.1157 Scale: 7222.7912 0.1223 0.1108 0.1557 Add: 8202.1488 0.1474 0.1463 0.1481 Triad: 7607.8306 0.1660 0.1577 0.2003 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.1on12"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.1on12 Content-MD5: a5vTHstwR/YCOggLxn7iKw== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 23501 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 3448896 microseconds (= 3448896 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 362.3409 2.2154 2.2079 2.2196 Scale: 215.8045 3.7080 3.7071 3.7092 Add: 288.2899 4.1637 4.1625 4.1651 Triad: 296.1738 4.0525 4.0517 4.0532 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.16on12"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.16on12 Content-MD5: uDpPl6aZzEBhXkE2/3BHaw== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 16001 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 235383 microseconds (= 235383 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 5181.8820 0.1547 0.1544 0.1563 Scale: 2976.4744 0.2690 0.2688 0.2693 Add: 3655.5044 0.3286 0.3283 0.3292 Triad: 3696.3705 0.3249 0.3246 0.3251 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.32on12"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.32on12 Content-MD5: MvyqpW/8XUpsnLaVq+Sxnw== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 8001 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 128476 microseconds (= 128476 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 8474.0441 0.0947 0.0944 0.0966 Scale: 4746.2450 0.1781 0.1686 0.2459 Add: 5726.7007 0.2097 0.2095 0.2100 Triad: 5770.7797 0.2080 0.2079 0.2082 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.48on12"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.48on12 Content-MD5: Xd3R1sEVAPvpm785WqJuYw== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 1 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 89427 microseconds (= 89427 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 10536.3358 0.0770 0.0759 0.0831 Scale: 4950.2602 0.1647 0.1616 0.1847 Add: 6336.5245 0.1895 0.1894 0.1897 Triad: 5940.5107 0.2023 0.2020 0.2026 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.1on8"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.1on8 Content-MD5: L9I+VdLHslOUOKq8IkqxbA== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 15501 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 3445439 microseconds (= 3445439 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 362.2980 2.2103 2.2081 2.2120 Scale: 215.9640 3.7059 3.7043 3.7067 Add: 288.7717 4.1566 4.1555 4.1575 Triad: 296.4783 4.0490 4.0475 4.0507 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.16on8"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.16on8 Content-MD5: 1KfD3wfrOFMqlt99PrLg2Q== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 8001 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 245438 microseconds (= 245438 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 4892.7287 0.1639 0.1635 0.1652 Scale: 2885.0024 0.2775 0.2773 0.2777 Add: 3465.8932 0.3464 0.3462 0.3466 Triad: 3445.9909 0.3484 0.3482 0.3486 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000 Content-Type: TEXT/plain; name="mod.stream.32on8"; charset=us-ascii; x-unix-mode=0644 Content-Description: mod.stream.32on8 Content-MD5: ah+dLje+IUk1mLrRaaqdVA== shmget status: Error 0 shmat status: Error 0 shmctl to destroy the block: Error 0 Value of shared_memory_key: 1 Got memory at 9e800000 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 50000000 Offset = 0 The total memory requirement is 1144 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 166871 microseconds (= 166871 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 6932.9629 0.1190 0.1154 0.1423 Scale: 3556.5623 0.2327 0.2249 0.2468 Add: 4312.7981 0.2915 0.2782 0.3188 Triad: 4031.9559 0.3077 0.2976 0.3217 Sum of a is = 5.7665039122923D+19 Sum of b is = 1.1533007808505D+19 Sum of c is = 1.5377343764968D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Family_of_Sardines_330_000-- From John@McCalpin.com Mon Jun 21 08:49:45 1999 Received: from mccloud.engr.sgi.com (mccloud.engr.sgi.com [192.48.196.40]) by frakir.engr.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via ESMTP id IAA06598 for ; Mon, 21 Jun 1999 08:49:44 -0700 (PDT) Received: (from guest@localhost) by mccloud.engr.sgi.com (8.8.7/8.8.7) id JAA28228 for mccalpin@frakir.engr.sgi.com; Mon, 21 Jun 1999 09:50:37 -0700 Date: Mon, 21 Jun 1999 09:50:37 -0700 From: John McCalpin Message-Id: <199906211650.JAA28228@mccloud.engr.sgi.com> To: mccalpin Subject: STREAM on cloudburst Status: RO [guest@mccloud ~/STREAM]$ more LOG.0020 a 0x804a980 b 0x87ebc20 c 0x8f8cec0 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 20 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 3 microseconds. Each test below will take on the order of 38929 microseconds. (= 12976 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 339.4937 0.0475 0.0471 0.0503 Scale: 325.9317 0.0491 0.0491 0.0492 Add: 368.8960 0.0651 0.0651 0.0652 Triad: 374.4734 0.0642 0.0641 0.0642 From Brian.Whitney@West.Sun.COM Fri Jun 18 09:05:43 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via ESMTP id JAA12200 for ; Fri, 18 Jun 1999 09:05:43 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id JAA53967 for ; Fri, 18 Jun 1999 09:05:42 -0700 (PDT) mail_from (Brian.Whitney@West.Sun.COM) Received: from mercury.Sun.COM (mercury.Sun.COM [192.9.25.1]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id JAA05139 for ; Fri, 18 Jun 1999 09:05:40 -0700 (PDT) mail_from (Brian.Whitney@West.Sun.COM) Received: from westmail2.West.Sun.COM ([129.153.100.30]) by mercury.Sun.COM (8.9.3+Sun/8.9.3) with ESMTP id JAA18585 for ; Fri, 18 Jun 1999 09:05:39 -0700 (PDT) Received: from oregon.West.Sun.COM (oregon.West.Sun.COM [129.153.37.10]) by westmail2.West.Sun.COM (8.9.1b+Sun/8.9.1/ENSMAIL,v1.6) with SMTP id JAA19145 for ; Fri, 18 Jun 1999 09:05:37 -0700 (PDT) Received: from sailfish by oregon.West.Sun.COM (SMI-8.6/SMI-SVR4) id JAA28559; Fri, 18 Jun 1999 09:05:38 -0700 Message-Id: <199906181605.JAA28559@oregon.West.Sun.COM> Date: Fri, 18 Jun 1999 09:05:38 -0700 (PDT) From: Brian Whitney Reply-To: Brian Whitney Subject: Re: Stream Results - Sun E10000 To: mccalpin Cc: Brian.Whitney@West.Sun.COM MIME-Version: 1.0 Content-Type: MULTIPART/mixed; BOUNDARY=Yoke_of_Oxen_967_000 X-Mailer: dtmail 1.3.0 CDE Version 1.3 SunOS 5.7 sun4u sparc Status: RO --Yoke_of_Oxen_967_000 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: otrfKUWcCjKb2NH9DXNOiA== Dr McCalpin... I have rerun things in 64-bit mode for 64, 48, 32, and 16 CPUs. I did 5 runs per CPU config. Attached are the logs. The array size is 300,000,000. This is 4.6M array elements per processor, or 36.8MB array per processor. This should be large enough (correct?). All runs are on a full system, just less CPUs. Results track pretty well from what I submitted before, but these are a bit higher in general. As far as the presentation of you table, it would be good to have translations of the machine name to a bit more detail. I just need a way to translate the listed name to what a particular company offerred. The names of the attached files should be pretty self explanatory, but pstx2-64.logx16.400x8.64 64 processor results pstx2-64.logx16.400x8.48 48 processor results pstx2-64.logx16.400x8.32 32 processor results pstx2-64.logx16.400x8.16 16 processor results keeper@sae3:/home/keeper% 201) sys sae3 06/18/99 09:00:05 SunOS 5.7 Generic_106541-02 sun4u 400/100 MHz 64 CPUs 8 MB E$ 64.00 GB memory SUNW,Ultra-Enterprise-10000 keeper@sae3:/home/keeper% 202) Brian --Yoke_of_Oxen_967_000 Content-Type: TEXT/plain; name="pstx2-64.logx16.400x8.64"; charset=us-ascii; x-unix-mode=0644 Content-Description: pstx2-64.logx16.400x8.64 Content-MD5: aPAwm7WadGOy1ApH1VthvQ== ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 596697 microseconds (= 596697 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 12141.5503 0.4165 0.3953 0.4814 Scale: 7970.4096 0.6312 0.6022 0.6602 Add: 8599.0450 0.8534 0.8373 0.8617 Triad: 8600.4443 0.8726 0.8372 0.8786 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 595651 microseconds (= 595651 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 12130.0444 0.4414 0.3957 0.6169 Scale: 7958.3941 0.6284 0.6031 0.6484 Add: 8599.2201 0.8521 0.8373 0.8739 Triad: 8630.3129 0.8726 0.8343 0.8958 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 579513 microseconds (= 579513 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 12167.0028 0.4196 0.3945 0.5000 Scale: 7970.0226 0.6201 0.6023 0.6277 Add: 8605.7956 0.8408 0.8366 0.8469 Triad: 8616.0746 0.8604 0.8356 0.8677 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 597482 microseconds (= 597482 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 12132.9098 0.4200 0.3956 0.4814 Scale: 7966.3353 0.6192 0.6025 0.6306 Add: 8472.3587 0.8530 0.8498 0.8620 Triad: 8613.7797 0.8651 0.8359 0.8741 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 581655 microseconds (= 581655 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 12159.9570 0.4243 0.3947 0.5513 Scale: 7977.7957 0.6290 0.6017 0.6637 Add: 8610.8937 0.8489 0.8362 0.8539 Triad: 8251.6698 0.8734 0.8726 0.8741 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Yoke_of_Oxen_967_000 Content-Type: TEXT/plain; name="pstx2-64.logx16.400x8.48"; charset=us-ascii; x-unix-mode=0644 Content-Description: pstx2-64.logx16.400x8.48 Content-MD5: KjLg9ZbkIFGqALhj10YXxQ== ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 599304 microseconds (= 599304 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 11577.5532 0.4209 0.4146 0.4680 Scale: 7353.7291 0.6549 0.6527 0.6694 Add: 7789.0422 0.9253 0.9244 0.9274 Triad: 7201.2731 1.0018 0.9998 1.0027 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 600044 microseconds (= 600044 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 11729.9561 0.4144 0.4092 0.4155 Scale: 7395.0202 0.6515 0.6491 0.6678 Add: 7823.1004 0.9209 0.9204 0.9219 Triad: 7229.6332 0.9965 0.9959 0.9975 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 599581 microseconds (= 599581 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 11595.9554 0.4250 0.4139 0.5069 Scale: 7415.4925 0.6500 0.6473 0.6657 Add: 7840.7754 0.9189 0.9183 0.9195 Triad: 7238.2820 0.9956 0.9947 0.9964 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 600269 microseconds (= 600269 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 11610.4149 0.4328 0.4134 0.5741 Scale: 7479.3262 0.6480 0.6418 0.6954 Add: 7862.3724 0.9164 0.9158 0.9169 Triad: 7274.7876 0.9904 0.9897 0.9909 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 599749 microseconds (= 599749 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 11599.4181 0.4229 0.4138 0.4933 Scale: 7477.0253 0.6449 0.6420 0.6606 Add: 7840.6027 0.9189 0.9183 0.9203 Triad: 7306.8919 0.9860 0.9854 0.9868 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Yoke_of_Oxen_967_000 Content-Type: TEXT/plain; name="pstx2-64.logx16.400x8.32"; charset=us-ascii; x-unix-mode=0644 Content-Description: pstx2-64.logx16.400x8.32 Content-MD5: 0OWFR5aKoLArP6XxYzCpRw== ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 780404 microseconds (= 780404 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 9367.4950 0.5307 0.5124 0.6678 Scale: 5874.4214 0.8187 0.8171 0.8261 Add: 6068.6759 1.1876 1.1864 1.1885 Triad: 5176.2323 1.3918 1.3910 1.3928 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 781327 microseconds (= 781327 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 9362.0170 0.5134 0.5127 0.5140 Scale: 5872.6900 0.8187 0.8173 0.8272 Add: 6066.3548 1.1878 1.1869 1.1885 Triad: 5174.4295 1.3922 1.3915 1.3934 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 780310 microseconds (= 780310 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 9382.2079 0.5224 0.5116 0.6046 Scale: 5878.0314 0.8178 0.8166 0.8247 Add: 6073.9815 1.1861 1.1854 1.1870 Triad: 5175.0300 1.3920 1.3913 1.3931 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 780756 microseconds (= 780756 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 9447.7536 0.5272 0.5081 0.6665 Scale: 5902.1595 0.8144 0.8133 0.8209 Add: 6096.9868 1.1818 1.1809 1.1825 Triad: 5179.7939 1.3907 1.3900 1.3912 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 780661 microseconds (= 780661 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 9457.4364 0.5110 0.5075 0.5116 Scale: 5898.7208 0.8154 0.8137 0.8229 Add: 6088.9977 1.1833 1.1825 1.1840 Triad: 5175.8864 1.3916 1.3911 1.3923 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Yoke_of_Oxen_967_000 Content-Type: TEXT/plain; name="pstx2-64.logx16.400x8.16"; charset=us-ascii; x-unix-mode=0644 Content-Description: pstx2-64.logx16.400x8.16 Content-MD5: vhxZkncx390T9JiP2FILLQ== ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 1423088 microseconds (= 1423088 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 5415.7614 0.8872 0.8863 0.8914 Scale: 3489.8197 1.3760 1.3754 1.3776 Add: 3581.2702 2.0126 2.0105 2.0216 Triad: 2788.4158 2.5837 2.5821 2.5846 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 1420296 microseconds (= 1420296 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 5427.8851 0.8868 0.8843 0.8899 Scale: 3479.8422 1.3805 1.3794 1.3834 Add: 3560.7318 2.0231 2.0221 2.0242 Triad: 2786.9876 2.5845 2.5834 2.5853 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 1422022 microseconds (= 1422022 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 5417.5998 0.8906 0.8860 0.9264 Scale: 3497.1673 1.3732 1.3725 1.3743 Add: 3593.3089 2.0056 2.0037 2.0117 Triad: 2790.1061 2.5825 2.5805 2.5836 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 1420779 microseconds (= 1420779 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 5385.5530 0.8929 0.8913 0.9047 Scale: 3468.2695 1.3848 1.3840 1.3862 Add: 3552.4322 2.0277 2.0268 2.0285 Triad: 2781.6282 2.5892 2.5884 2.5902 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 1422451 microseconds (= 1422451 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 5426.0808 0.8864 0.8846 0.8881 Scale: 3492.3740 1.3755 1.3744 1.3764 Add: 3579.7885 2.0130 2.0113 2.0185 Triad: 2788.6280 2.5829 2.5819 2.5837 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) --Yoke_of_Oxen_967_000-- From Brian.Whitney@West.Sun.COM Wed Jun 16 15:54:51 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via ESMTP id PAA08479 for ; Wed, 16 Jun 1999 15:54:51 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id PAA30980 for ; Wed, 16 Jun 1999 15:54:50 -0700 (PDT) mail_from (Brian.Whitney@West.Sun.COM) Received: from mercury.Sun.COM (mercury.Sun.COM [192.9.25.1]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id PAA09185 for ; Wed, 16 Jun 1999 15:54:49 -0700 (PDT) mail_from (Brian.Whitney@West.Sun.COM) Received: from westmail2.West.Sun.COM ([129.153.100.30]) by mercury.Sun.COM (8.9.3+Sun/8.9.3) with ESMTP id PAA16680 for ; Wed, 16 Jun 1999 15:54:48 -0700 (PDT) Received: from oregon.West.Sun.COM (oregon.West.Sun.COM [129.153.37.10]) by westmail2.West.Sun.COM (8.9.1b+Sun/8.9.1/ENSMAIL,v1.6) with SMTP id PAA28896 for ; Wed, 16 Jun 1999 15:54:47 -0700 (PDT) Received: from sailfish by oregon.West.Sun.COM (SMI-8.6/SMI-SVR4) id PAA11108; Wed, 16 Jun 1999 15:54:47 -0700 Message-Id: <199906162254.PAA11108@oregon.West.Sun.COM> Date: Wed, 16 Jun 1999 15:54:47 -0700 (PDT) From: Brian Whitney Reply-To: Brian Whitney Subject: Re: Stream Results - Sun E10000 To: mccalpin MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: lop+r9Zy2gBr5jVxCjJDyA== X-Mailer: dtmail 1.3.0 CDE Version 1.3 SunOS 5.7 sun4u sparc X-Zm-Priority: Medium Status: RO Ran a slightly larger data set size (yes, 300,000,000)... I think I like these results, they are making more sense. Now to see if smaller size make sense... Brian keeper@sae3:/home/keeper/stream% 208) setenv PARALLEL 64 keeper@sae3:/home/keeper/stream% 209) pstx2-64 ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 300000000 Offset = 0 The total memory requirement is 6866 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity appears to be less than one microsecond Your clock granularity/precision appears to be 1 microseconds The tests below will each take a time on the order of 605053 microseconds (= 605053 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 12071.9246 0.4192 0.3976 0.4464 Scale: 7967.1486 0.6185 0.6025 0.6578 Add: 8497.6855 0.8515 0.8473 0.8592 Triad: 8590.0299 0.8652 0.8382 0.8755 Sum of a is = 3.4599023433963D+20 Sum of b is = 6.9198046533463D+19 Sum of c is = 9.2264062741574D+19 Note: Nonstandard floating-point mode enabled See the Numerical Computation Guide, ieee_sun(3M) From Daniel.S.Katz@jpl.nasa.gov Fri Jun 25 13:59:25 1999 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by frakir.engr.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via ESMTP id NAA15867 for ; Fri, 25 Jun 1999 13:59:23 -0700 (PDT) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id NAA97871 for ; Fri, 25 Jun 1999 13:59:22 -0700 (PDT) mail_from (Daniel.S.Katz@jpl.nasa.gov) Received: from ares.cs.Virginia.EDU (ares.cs.Virginia.EDU [128.143.136.41] (may be forged)) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id NAA09060 for ; Fri, 25 Jun 1999 13:59:20 -0700 (PDT) mail_from (Daniel.S.Katz@jpl.nasa.gov) Received: from eis-msg-014.jpl.nasa.gov (eis-msg-014.jpl.nasa.gov [137.78.160.189]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id QAA21880 for ; Fri, 25 Jun 1999 16:59:18 -0400 (EDT) Received: from jpl.nasa.gov (kangaroo.jpl.nasa.gov [137.78.73.151]) by eis-msg-014.jpl.nasa.gov (8.8.5/8.8.5) with ESMTP id NAA25417 for ; Fri, 25 Jun 1999 13:59:46 -0700 (PDT) Message-ID: <3773ED66.5C9394D9@jpl.nasa.gov> Date: Fri, 25 Jun 1999 13:58:15 -0700 From: "Daniel S. Katz" Reply-To: Daniel.S.Katz@jpl.nasa.gov Organization: Jet Propulsion Laboratory X-Mailer: Mozilla 4.61 (Macintosh; U; PPC) X-Accept-Language: en MIME-Version: 1.0 To: mccalpin@cs.virginia.edu Subject: STREAM results - Mac G3-400 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Status: RO Hi, I ran the benchmarks on my Mac G3-400 laptop: PPC BASE.out: ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 400000, Offset = 0 Total memory required = 9.2 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 17 microseconds. Each test below will take on the order of 28143 microseconds. (= 1655 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 125.6577 0.0513 0.0509 0.0528 Scale: 123.0485 0.0522 0.0520 0.0525 Add: 129.5651 0.0749 0.0741 0.0758 Triad: 129.4987 0.0745 0.0741 0.0748 PPC UNROLL.out: ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 400000, Offset = 0 Total memory required = 9.2 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 17 microseconds. Each test below will take on the order of 28381 microseconds. (= 1669 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 125.5714 0.0516 0.0510 0.0537 Scale: 125.0464 0.0519 0.0512 0.0549 Add: 130.6958 0.0747 0.0735 0.0777 Triad: 129.5634 0.0753 0.0741 0.0779 PPC601 OPT.out: ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 400000, Offset = 0 Total memory required = 9.2 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 16 microseconds. Each test below will take on the order of 28254 microseconds. (= 1765 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 151.5439 0.0427 0.0422 0.0436 Scale: 140.7305 0.0461 0.0455 0.0471 Add: 141.1848 0.0683 0.0680 0.0691 Triad: 138.2887 0.0697 0.0694 0.0706 PPC604 OPT.out: ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 400000, Offset = 0 Total memory required = 9.2 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 17 microseconds. Each test below will take on the order of 28644 microseconds. (= 1684 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 148.9758 0.0434 0.0430 0.0447 Scale: 145.2795 0.0447 0.0441 0.0455 Add: 148.4666 0.0650 0.0647 0.0659 Triad: 144.6480 0.0666 0.0664 0.0668 From fjohn@us.ibm.com Mon Sep 27 11:16:17 1999 X-Lotus-FromDomain: IBMUS To: mccalpin@austin.ibm.com Message-ID: <85256823.00724389.00@d54mta05.raleigh.ibm.com> From: Frank Johnston Date: Mon, 27 Sep 1999 11:16:01 -0500 Subject: STREAM on Power3smp Mime-Version: 1.0 Content-type: text/plain; charset=us-ascii Content-Disposition: inline X-Mozilla-Status: 8001 X-Mozilla-Status2: 00000000 Status: RO To whom it may concern, I would like to submit STREAM benchmark results for an IBM RS/6000 SP 222 MHz POWER3smp. Please add the following summary to the web page of standard STREAM results: Machine ID npus COPY SCALE ADD TRIAD IBM-SP_222MHz-POWER3smp 1 790.4 794.1 951.0 953.3 IBM-SP_222MHz-POWER3smp 2 1549.2 1524.4 1843.5 1844.5 IBM-SP_222MHz-POWER3smp 4 2677.0 2630.9 3091.3 3101.2 IBM-SP_222MHz-POWER3smp 8 4332.9 4211.4 4843.3 4937.8 Each CPU on this machine has a private 4 MB L2 cache. The array size at each CPU count exceeds the aggregate available L2. However the arrays were kept as small as possible in order to give system tables mapping virtual to real memory a chance to reside in the L2. When data in these tables comes from main memory, there is additional memory traffic which STREAM does not count. In addition to changing the array size and offset, the following changes were made to the Fortran STREAM code: 1) Increase "ntimes" to 1000 in order to reduce run-to-run fluctuations. 2) Change the "scalar" constant to sqrt(2) - 1 to prevent overflows (INFs). 3) Add compiler directives to allow the code to run in parallel on multiple CPUs and to allow certain arrays to be "prefetched" by the hardware. Here is an example (the COPY loop): !smp$ parallel do private(j) , schedule(static) DO 30 j = 1,n !ibm* prefetch_by_load( c(j) ) c(j) = a(j) 30 CONTINUE If you have any questions, please let me know. Frank Johnston (fjohn@us.ibm.com, xxx-xxx-xxxx) P.S. Here are the actual outputs: output for 1 CPU: ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 524300 Offset = 330 The total memory requirement is 12 MB You are running each test 1000 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 5 microseconds The tests below will each take a time on the order of 8259 microseconds (= 1652 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 790.4472 .0107 .0106 .0116 Scale: 794.1492 .0106 .0106 .0109 Add: 951.0364 .0133 .0132 .0144 Triad: 953.3298 .0133 .0132 .0137 Sum of a is = 1048600.00000001118 Sum of b is = 434344.341503158561 Sum of c is = 1482944.34150735638 output for 2 CPUs: ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 1048600 Offset = 404 The total memory requirement is 24 MB You are running each test 1000 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 5 microseconds The tests below will each take a time on the order of 15856 microseconds (= 3171 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 1549.2400 .0109 .0108 .0201 Scale: 1524.3722 .0111 .0110 .0120 Add: 1843.4750 .0137 .0137 .0147 Triad: 1844.5380 .0137 .0136 .0143 Sum of a is = 2097200.00000001118 Sum of b is = 868688.683004021645 Sum of c is = 2965888.68302347837 output for 4 CPUs: ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 2097200 Offset = 260 The total memory requirement is 48 MB You are running each test 1000 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 5 microseconds The tests below will each take a time on the order of 32019 microseconds (= 6404 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 2676.9512 .0133 .0125 .0329 Scale: 2630.9133 .0133 .0128 .0229 Add: 3091.3001 .0169 .0163 .0181 Triad: 3101.2452 .0168 .0162 .0180 Sum of a is = 4194400.00000001118 Sum of b is = 1737377.36602994637 Sum of c is = 5931777.36605572235 output for 8 CPUs: ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 4194400 Offset = 468 The total memory requirement is 96 MB You are running each test 1000 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 5 microseconds The tests below will each take a time on the order of 64196 microseconds (= 12839 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 4332.9165 .0166 .0155 .0576 Scale: 4211.3665 .0170 .0159 .0588 Add: 4843.2744 .0217 .0208 .0607 Triad: 4937.7796 .0213 .0204 .0486 Sum of a is = 8388800.00000001118 Sum of b is = 3474754.73209443502 Sum of c is = 11863554.7321202103 From fjohn@us.ibm.com Tue Oct 12 16:49:16 1999 X-Lotus-FromDomain: IBMUS To: mccalpin@austin.ibm.com From: Frank Johnston Message-ID: <85256823.0072A381.00@d54mta05.raleigh.ibm.com> Date: Tue, 12 Oct 1999 16:49:16 -0500 Subject: STREAM (base) on POWER3smp Status: RO From: Frank Johnston on 10/12/99 04:49 PM To: John D Mccalpin/Austin/IBM@IBMUS John, Here are STREAM results without the compiler directive for a Nighthawk1 (IBM RS/6000 SP 222 MHz POWER3smp). Would it be possible to publish both these and the results obtained with the directive on the table of "Standard Results", using slightly different machine IDs? I still believe that use of a simple Fortran compiler directive is much closer in spirit to the "Standard Results", than "Experimental Results" obtained with hand-coded assembler. Machine ID npus COPY SCALE ADD TRIAD IBM-SP_222MHz-POWER3smp 1 413.0 421.2 587.4 614.2 IBM-SP_222MHz-POWER3smp 2 820.7 800.4 1182.1 1165.4 IBM-SP_222MHz-POWER3smp 4 1603.5 1535.9 2218.3 2187.5 IBM-SP_222MHz-POWER3smp 8 2954.1 2821.0 3889.0 3872.2 Frank Johnston output for 1 CPU: ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 524300 Offset = 260 The total memory requirement is 12 MB You are running each test 1000 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 5 microseconds The tests below will each take a time on the order of 8319 microseconds (= 1664 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 412.9788 .0204 .0203 .0213 Scale: 421.1776 .0200 .0199 .0210 Add: 587.4248 .0216 .0214 .0226 Triad: 614.2449 .0205 .0205 .0216 Sum of a is = 1048600.00000001118 Sum of b is = 434344.341503158561 Sum of c is = 1482944.34150735638 output for 2 CPUs: ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 1048600 Offset = 360 The total memory requirement is 24 MB You are running each test 750 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 5 microseconds The tests below will each take a time on the order of 8704 microseconds (= 1741 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 820.6599 .0205 .0204 .0217 Scale: 800.4499 .0210 .0210 .0222 Add: 1182.0659 .0214 .0213 .0226 Triad: 1165.4323 .0217 .0216 .0229 Sum of a is = 2097200.00000000373 Sum of b is = 868688.683004021295 Sum of c is = 2965888.68302347790 output for 4 CPUs: ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 2097200 Offset = 272 The total memory requirement is 48 MB You are running each test 1000 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 5 microseconds The tests below will each take a time on the order of 11072 microseconds (= 2214 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 1603.5355 .0219 .0209 .0232 Scale: 1535.8922 .0225 .0218 .0235 Add: 2218.2522 .0235 .0227 .0249 Triad: 2187.5434 .0236 .0230 .0247 Sum of a is = 4194400.00000001118 Sum of b is = 1737377.36602994637 Sum of c is = 5931777.36605572235 output for 8 CPUs: ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 4194400 Offset = 224 The total memory requirement is 96 MB You are running each test 1000 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 5 microseconds The tests below will each take a time on the order of 13120 microseconds (= 2624 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 2954.0690 .0240 .0227 .0347 Scale: 2820.9498 .0247 .0238 .0724 Add: 3889.0287 .0271 .0259 .0456 Triad: 3872.2120 .0270 .0260 .0612 From jeppesen@mrl.ucsb.edu Thu Oct 14 21:10:42 1999 Message-ID: <0.711fc4b7.25388127@aol.com> Date: Thu, 14 Oct 1999 21:10:42 -0700 From: Claus Jeppesen Subject: athlon To: mccalpin@austin.ibm.com MIME-Version: 1.0 Status: RO Hi John, A result for your streams table. Thanx, Claus Jeppesen Computer: Polywell 800K7 CPU: 600Mhz AMD K7 (Athlon) Memory: 256MB PC100 RAM Compiler: Portland Group - pgf77 -fast streams_d.f I used the version of second_cpu.f which implies etime(...) - results from time and streams_d.f : ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 8000000 Offset = 0 The total memory requirement is 183 MB You are running each test 10 times The *best* time for each test is used ---------------------------------------------------- Your clock granularity/precision appears to be 10000 microseconds The tests below will each take a time on the order of 220000 microseconds (= 22 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ---------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ---------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 387.8788 0.3390 0.3300 0.3400 Scale: 474.0741 0.2700 0.2700 0.2700 Add: 480.0000 0.4061 0.4000 0.4200 Triad: 468.2927 0.4150 0.4100 0.4200 Sum of a is = 9.2264062494077470E+018 Sum of b is = 1.8452812500593927E+018 Sum of c is = 2.4603749999959350E+018 15.690u 0.870s 0:17.11 96.7% 0+0k 0+0io 103pf+0w From mason.cabot@intel.com Wed Oct 27 16:03:57 1999 Received: from netmail.austin.ibm.com (netmail.austin.ibm.com [9.53.250.98]) by gmp246.austin.ibm.com (AIX4.3/UCB 8.8.8/8.7-client1.01) with ESMTP id QAA24800 for ; Wed, 27 Oct 1999 16:03:55 -0500 Received: from ausmail.austin.ibm.com (am00.austin.ibm.com [9.3.199.19]) by netmail.austin.ibm.com (8.8.5/8.8.5) with ESMTP id QAA20356 for ; Wed, 27 Oct 1999 16:03:52 -0500 Received: from thalia.fm.intel.com (thalia.fm.intel.com [132.233.247.11]) by ausmail.austin.ibm.com (8.9.1/8.8.5) with SMTP id QAA16920 for ; Wed, 27 Oct 1999 16:05:10 -0500 Received: from fmsmsx27.FM.INTEL.COM (fmsmsx27.fm.intel.com [132.233.42.27]) by thalia.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.10 1999/10/20 18:19:05 spurcell Exp $) with ESMTP id RAA26752 for ; Wed, 27 Oct 1999 17:04:22 GMT Received: by fmsmsx27.fm.intel.com with Internet Mail Service (5.5.2448.0) id ; Wed, 27 Oct 1999 14:03:53 -0700 Message-ID: <438BA6CBABC5D111AC4100A0C96B75B302957EF4@fmsmsx41.fm.intel.com> From: "Cabot, Mason" To: "'mccalpin@austin.ibm.com'" Subject: RE: funny BW on 820-based system Date: Wed, 27 Oct 1999 14:03:50 -0700 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2448.0) Status: RO John, I don't think that there are any issues with the board they are using. There is a newer BIOS (P04) available for the CC820 board that fixes some things, but I don't think that these fixes are performance related. I think the main problem here is that they are using the DOS version of STREAM rather than recompiling to fit the hardware/OS. For the systems that are reviewed in this article and the version of STREAM for NT posted on the web site, I see: *****i820 CuMine 733/133***** Copy: 441.9930 Scale: 438.3188 Add: 553.0855 Triad: 544.3054 *****i820 Katmai 600/133***** Copy: 379.1829 Scale: 386.3530 Add: 448.4531 Triad: 448.3660 *****BX Katmai 600/100***** Copy: 342.2239 Scale: 340.2257 Add: 411.9562 Triad: 409.2117 I've attached the full output from these tests. I haven't sent you any of my Athlon tests because I need to recompile to optimize for their FPU. Do you know if the latest versions of GCC optimize for K7? thanks, Mason -----Original Message----- From: John D. McCalpin, Ph.D. [mailto:mccalpin@austin.ibm.com] Sent: Wednesday, October 27, 1999 11:46 AM To: mason.cabot@intel.com Subject: funny BW on 820-based system Hi Mason A hardware review at: http://www.aceshardware.com/Spades/read.php?article_id=84 shows funny results for the 733EB+820 system with the Intel motherboard. Specifically, the 733EB+820 runs 30% slower than the PIII-600/440BX and more than 30% slower than the 600B+820 on a different vendor's motherboard. Are there any config issues that the authors need to be aware of? -- John D. McCalpin, Ph.D. mccalpin@austin.ibm.com Senior Scientist IBM Future Systems "I am willing to make mistakes as long as someone else is willing to learn from them." ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 999936 Total memory required = 24.9 MB. Each test is run 50 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 37238 microseconds. (= 37238 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 342.2239 0.0470 0.0468 0.0519 Scale: 340.2257 0.0472 0.0470 0.0495 Add: 411.9562 0.0585 0.0583 0.0588 Triad: 409.2117 0.0589 0.0586 0.0591 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 999936 Total memory required = 24.9 MB. Each test is run 50 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 31814 microseconds. (= 31814 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 379.1829 0.0424 0.0422 0.0475 Scale: 386.3530 0.0416 0.0414 0.0443 Add: 448.4531 0.0536 0.0535 0.0540 Triad: 448.3660 0.0536 0.0535 0.0540 ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 999936 Total memory required = 24.9 MB. Each test is run 50 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 23162 microseconds. (= 23162 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 441.9930 0.0364 0.0362 0.0394 Scale: 438.3188 0.0366 0.0365 0.0382 Add: 553.0855 0.0435 0.0434 0.0437 Triad: 544.3054 0.0442 0.0441 0.0445 From john@mccalpin.com Fri Oct 29 06:39:17 1999 Received: from netmail2.austin.ibm.com (netmail2.austin.ibm.com [9.53.250.97]) by gmp246.austin.ibm.com (AIX4.3/UCB 8.8.8/8.7-client1.01) with ESMTP id GAA45158 for ; Fri, 29 Oct 1999 06:39:17 -0500 Received: from ausmail.austin.ibm.com (am00.austin.ibm.com [9.3.199.19]) by netmail2.austin.ibm.com (8.8.5/8.8.5) with ESMTP id GAA22198 for ; Fri, 29 Oct 1999 06:39:19 -0500 Received: from Mail.austin.rr.com (sm2.texas.rr.com [24.93.35.55]) by ausmail.austin.ibm.com (8.9.1/8.8.5) with SMTP id GAA33076 for ; Fri, 29 Oct 1999 06:40:33 -0500 Received: from austin.rr.com ([24.93.44.235]) by Mail.austin.rr.com with Microsoft SMTPSVC(5.5.1875.185.18); Fri, 29 Oct 1999 06:32:24 -0500 Message-ID: <38198752.143D94F6@austin.rr.com> Date: Fri, 29 Oct 1999 06:38:58 -0500 From: John McCalpin Reply-To: john@mccalpin.com X-Mailer: Mozilla 4.08 (Macintosh; I; 68K) MIME-Version: 1.0 To: mccalpin@austin.ibm.com Subject: STREAM benchmark for various Macs Content-Type: multipart/mixed; boundary="------------6AC1080B7408A9EA88878007" X-Mozilla-Status: 8001 X-Mozilla-Status2: 00000000 Status: RO This is a multi-part message in MIME format. --------------6AC1080B7408A9EA88878007 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit http://www.cloudmaster.com/pulsar/stream/ --------------6AC1080B7408A9EA88878007 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Content-Base: "http://www.cloudmaster.com/pulsar/stre am/" STREAM benchmark for various Macs

This page lists runs of the STREAM benchmark for testing memory bandwidth. The version run on these machines is the "604opt" Mac executable. If you would like to see the results for the other executables, just email me and I'll send them to you. I might put them up on the page, but that'll be later.

The PowerBook had MathLibMoto which may or may not have affected the scores any. Maybe I'll find out if it actualy does have an affect. There are LinuxPPC scores on the PowerBook as well, but I haven't gotten around to putting them up yet. I might eventually put up a Dell 233MHz PPro (EDO 66MHz), a 75MHz EDO based 375MHz AMD K6-2 system, and the same system with 100MHz SDRAM too.

I will rearrange the table so it looks better when I get other scores in. I just whipped this page up in a few minutes based on the relatively few scores I've collected.

Test values are in MB/s.

Machine processor:bus PowerMac G3 400:100MHz PowerMac G3 266:66MHz PowerBook G3 266:66MHz iMac Rev C 266:66MHz iMac Rev A 233:66MHz
Copy 233.4914 146.9103 132.3599 138.2230 137.8894
Scale 222.5236 143.3756 126.6549 138.1573 135.1665
Add 230.6473 132.2879 130.2525 128.8331 128.8833
Triad 212.9736 131.5087 121.2259 126.6056 125.6874

 

Send comments, suggestions, and criticisms to Ryan Tokarek

Back to my home page.

--------------6AC1080B7408A9EA88878007-- From fjohn@us.ibm.com Wed Nov 3 11:57:31 1999 X-Lotus-FromDomain: IBMUS To: mccalpin@austin.ibm.com From: Frank Johnston Message-ID: <85256823.007ACC33.00@d54mta05.raleigh.ibm.com> Date: Wed, 3 Nov 1999 11:57:31 -0500 Subject: IBM 43P-260 base and T70 base Status: RO From: Frank Johnston on 11/03/99 11:57 AM Here is the data: (no pretech directive or hand-tuning) Machine CPUs Copy Scale Add Triad 43P-260 1 597.5 569.3 757.3 776.1 43P-260 2 960.9 965.7 1138.7 1112.3 T70 1 413.0 421.2 587.4 614.2 T70 2 820.7 800.4 1182.1 1165.4 T70 4 1603.5 1535.9 2218.3 2187.5 T70 8 2954.1 2821.0 3889.0 3872.2 Frank Johnston From - Mon Dec 13 17:50:56 1999 Received: from aux153.plano.net ([209.217.36.11]) by mail.austin.rr.com with Microsoft SMTPSVC(5.5.1877.197.19); Mon, 13 Dec 1999 10:48:15 -0600 Received: (qmail 4823 invoked by uid 10000); 13 Dec 1999 16:53:32 -0000 Delivered-To: mccalpin.com-john@mccalpin.com Received: (qmail 4547 invoked from network); 13 Dec 1999 16:53:18 -0000 Received: from ares.cs.virginia.edu (128.143.136.41) by aux153.plano.net with SMTP; 13 Dec 1999 16:53:18 -0000 Received: from syseng.hstc.necsyl.com (root@syseng.hstc.necsyl.com [198.170.1.20]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with SMTP id LAA21573 for ; Mon, 13 Dec 1999 11:47:35 -0500 (EST) Received: by syseng.hstc.necsyl.com (4.1/YDL1.7-940222.14) id AA04243(syseng.hstc.necsyl.com); Mon, 13 Dec 99 10:47:28 CST Message-Id: <009b01bf4589$f81f9a40$2502aac6@hstc.necsyl.com> Reply-To: "Philip Tannenbaum" From: "Philip Tannenbaum" To: Subject: SX-5 STREAM Date: Mon, 13 Dec 1999 10:49:03 -0600 Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-Msmail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2919.6600 X-Mimeole: Produced By Microsoft MimeOLE V5.00.2919.6600 Return-Path: philt@ieee.org X-Mozilla-Status: 9003 X-Mozilla-Status2: 00000000 X-UIDL: 5b000000.eml Status: RO Dear John, Per your request at SC99, I am enclosing formal STREAM results for the SX-5. The formal machine name is SX-5/16A Per the naming below, the digit after the "/" is the number of processors used in each run. So you have 1, 2, 4, 8, and 16 cpus represented. Remember that the SX-5 is an SDRAM memory, so it is not directly comparable with the SX-4 original with SSRAM, but is comparable to the SX-4A which also had SDRAM but which was never reported. BTW, I will put this into an HPCWire article in the next few days unless you want to release it as the keeper of the stream ;) Best Regards Philip Tannenbaum HNSX Supercomputers Inc. --------------------------------------------- PA24-260C N9312P ----- Original Message ----- Subject: Re: STREAM > > Machine SX5#4 16A 128GB > Date Dec 7 ,1999 > > [1] SX-5/1A > ---------------------------------------------- > Double precision appears to have 16 digits of accuracy > Assuming 8 bytes per DOUBLE PRECISION word > ---------------------------------------------- > Array size = 80000000 > Offset = 0 > The total memory requirement is 1831 MB > You are running each test 10 times > The *best* time for each test is used > ---------------------------------------------------- > Your clock granularity/precision appears to be 44 microseconds > The tests below will each take a time on the order > of 30108 microseconds > (= 684 clock ticks) > Increase the size of the arrays if this shows that > you are not getting at least 20 clock ticks per test. > ---------------------------------------------------- > WARNING -- The above is only a rough guideline. > For best results, please be sure you know the > precision of your system timer. > ---------------------------------------------------- > Function Rate (MB/s) RMS time Min time Max time > Copy: 42544.6479 0.0301 0.0301 0.0302 > Scale: 42546.1651 0.0301 0.0301 0.0301 > Add: 47780.3278 0.0402 0.0402 0.0402 > Triad: 47779.0521 0.0402 0.0402 0.0402 > Sum of a is = 9.226406249984801D+19 > Sum of b is = 1.845281249988019D+19 > Sum of c is = 2.460375000016015D+19 > > > > [2] SX-5/2A > ---------------------------------------------- > Double precision appears to have 16 digits of accuracy > Assuming 8 bytes per DOUBLE PRECISION word > ---------------------------------------------- > Array size = 80000000 > Offset = 0 > The total memory requirement is 1831 MB > You are running each test 10 times > The *best* time for each test is used > ---------------------------------------------------- > Your clock granularity/precision appears to be 44 microseconds > The tests below will each take a time on the order > of 15122 microseconds > (= 344 clock ticks) > Increase the size of the arrays if this shows that > you are not getting at least 20 clock ticks per test. > ---------------------------------------------------- > WARNING -- The above is only a rough guideline. > For best results, please be sure you know the > precision of your system timer. > ---------------------------------------------------- > Function Rate (MB/s) RMS time Min time Max time > Copy: 84852.6042 0.0151 0.0151 0.0151 > Scale: 84852.6042 0.0151 0.0151 0.0151 > Add: 95351.6151 0.0201 0.0201 0.0201 > Triad: 95327.9119 0.0201 0.0201 0.0202 > Sum of a is = 9.226406249985602D+19 > Sum of b is = 1.845281249991997D+19 > Sum of c is = 2.460375000000000D+19 > > > > [4] SX-5/4A > ---------------------------------------------- > Double precision appears to have 16 digits of accuracy > Assuming 8 bytes per DOUBLE PRECISION word > ---------------------------------------------- > Array size = 80000000 > Offset = 0 > The total memory requirement is 1831 MB > You are running each test 10 times > The *best* time for each test is used > ---------------------------------------------------- > Your clock granularity/precision appears to be 44 microseconds > The tests below will each take a time on the order > of 7620 microseconds > (= 173 clock ticks) > Increase the size of the arrays if this shows that > you are not getting at least 20 clock ticks per test. > ---------------------------------------------------- > WARNING -- The above is only a rough guideline. > For best results, please be sure you know the > precision of your system timer. > ---------------------------------------------------- > Function Rate (MB/s) RMS time Min time Max time > Copy: 168485.5912 0.0076 0.0076 0.0076 > Scale: 168509.3886 0.0076 0.0076 0.0076 > Add: 189555.2133 0.0101 0.0101 0.0101 > Triad: 189517.2955 0.0101 0.0101 0.0101 > Sum of a is = 9.226406249987201D+19 > Sum of b is = 1.845281249999994D+19 > Sum of c is = 2.460375000000000D+19 > > > > [8] SX-5/8A > ---------------------------------------------- > Double precision appears to have 16 digits of accuracy > Assuming 8 bytes per DOUBLE PRECISION word > ---------------------------------------------- > Array size = 80000000 > Offset = 0 > The total memory requirement is 1831 MB > You are running each test 10 times > The *best* time for each test is used > ---------------------------------------------------- > Your clock granularity/precision appears to be 44 microseconds > The tests below will each take a time on the order > of 3869 microseconds > (= 88 clock ticks) > Increase the size of the arrays if this shows that > you are not getting at least 20 clock ticks per test. > ---------------------------------------------------- > WARNING -- The above is only a rough guideline. > For best results, please be sure you know the > precision of your system timer. > ---------------------------------------------------- > Function Rate (MB/s) RMS time Min time Max time > Copy: 332551.3578 0.0039 0.0038 0.0039 > Scale: 332551.3578 0.0039 0.0038 0.0039 > Add: 371160.2378 0.0053 0.0052 0.0054 > Triad: 366690.0567 0.0053 0.0052 0.0054 > Sum of a is = 9.226406249990401D+19 > Sum of b is = 1.845281249999999D+19 > Sum of c is = 2.460375000000000D+19 > > > > [16] SX-5/16A > ---------------------------------------------- > Double precision appears to have 16 digits of accuracy > Assuming 8 bytes per DOUBLE PRECISION word > ---------------------------------------------- > Array size = 80000000 > Offset = 0 > The total memory requirement is 1831 MB > You are running each test 10 times > The *best* time for each test is used > ---------------------------------------------------- > Your clock granularity/precision appears to be 45 microseconds > The tests below will each take a time on the order > of 2157 microseconds > (= 48 clock ticks) > Increase the size of the arrays if this shows that > you are not getting at least 20 clock ticks per test. > ---------------------------------------------------- > WARNING -- The above is only a rough guideline. > For best results, please be sure you know the > precision of your system timer. > ---------------------------------------------------- > Function Rate (MB/s) RMS time Min time Max time > Copy: 607491.8382 0.0021 0.0021 0.0022 > Scale: 590389.7421 0.0022 0.0022 0.0023 > Add: 607411.6518 0.0032 0.0032 0.0033 > Triad: 583069.4479 0.0034 0.0033 0.0034 > Sum of a is = 9.226406249996797D+19 > Sum of b is = 1.845281249999999D+19 > Sum of c is = 2.460374999999999D+19 > > > From - Fri Dec 17 07:21:22 1999 Received: from aux153.plano.net ([209.217.36.11]) by Mail.austin.rr.com with Microsoft SMTPSVC(5.5.1877.197.19); Fri, 17 Dec 1999 05:51:44 -0600 Received: (qmail 27295 invoked by uid 10000); 17 Dec 1999 12:05:40 -0000 Delivered-To: mccalpin.com-john@mccalpin.com Received: (qmail 27184 invoked from network); 17 Dec 1999 12:05:34 -0000 Received: from ares.cs.virginia.edu (128.143.136.41) by aux153.plano.net with SMTP; 17 Dec 1999 12:05:34 -0000 Received: from mta4.snfc21.pbi.net (mta4.snfc21.pbi.net [206.13.28.142]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id GAA06456 for ; Fri, 17 Dec 1999 06:59:19 -0500 (EST) Received: from photon ([63.195.90.12]) by mta4.snfc21.pbi.net (Sun Internet Mail Server sims.3.5.1999.09.16.21.57.p8) with SMTP id <0FMV00H7YVYAKJ@mta4.snfc21.pbi.net> for mccalpin@cs.virginia.edu; Fri, 17 Dec 1999 03:59:13 -0800 (PST) Date: Fri, 17 Dec 1999 04:43:31 -0800 From: Paul Hsieh Subject: Stream score submission for Athlon system. To: mccalpin@cs.virginia.edu Cc: mccalpin@austin.ibm.com Message-id: <0FMV00H80VYAKJ@mta4.snfc21.pbi.net> MIME-version: 1.0 X-Mailer: Pegasus Mail for Win32 (v3.01b) Content-type: Multipart/Mixed; boundary=Message-Boundary-10580 Priority: normal Return-Path: qed@pobox.com X-Mozilla-Status: 8001 X-Mozilla-Status2: 00000000 X-UIDL: 0d000000.eml Status: RO --Message-Boundary-10580 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Content-description: Mail message body I performed this port to the WATCOM C/C++ compiler on my Athlon 500Mhz system. Here is a description of the files. 6763 stream_d.c - My modifications to the stream_d.c source. The most important modification was to change OFFSET from 0 to 70464. This number was determined experimentally. It also contains a conditionally compiled ASM hack to speed up just the COPY loop. 1645 p55clock.c - A different port of the timer code. It is far more accurate than previous timers, in particular the win32_clock.c timer code. 219 make.bat - Just a batch file for compiling everything up. 49152 stream.exe - The straight compile of the sources. 49152 fastream.exe - A compile of the sources with the assembly hack turned on. 1217 500.txt - The results without the assembly hack. 1217 500f.txt - The results with the assembly hack. -- Paul Hsieh qed@pobox.com --Message-Boundary-10580 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Content-description: Text from file '500.txt' ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 70464 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 36126 microseconds. (= 36126 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 373.5327 0.0435 0.0428 0.0460 Scale: 351.8294 0.0460 0.0455 0.0466 Add: 459.3388 0.0531 0.0522 0.0555 Triad: 453.4303 0.0533 0.0529 0.0539 --Message-Boundary-10580 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Content-description: Text from file '500f.txt' ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 1000000, Offset = 70464 Total memory required = 22.9 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds. Each test below will take on the order of 37978 microseconds. (= 37978 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 452.9012 0.0358 0.0353 0.0363 Scale: 351.0894 0.0467 0.0456 0.0500 Add: 454.1782 0.0552 0.0528 0.0586 Triad: 454.6508 0.0533 0.0528 0.0539 --Message-Boundary-10580 Content-type: text/plain; charset=US-ASCII Content-disposition: inline Content-description: Attachment information. The following section of this message contains a file attachment prepared for transmission using the Internet MIME message format. If you are using Pegasus Mail, or any another MIME-compliant system, you should be able to save it or view it from within your mailer. If you cannot, please ask your system administrator for assistance. ---- File information ----------- File: fastream.exe Date: 17 Dec 1999, 4:31 Size: 49152 bytes. Type: PCEXE --Message-Boundary-10580 Content-type: Application/Octet-stream; name="fastream.exe"; type=PCEXE Content-disposition: attachment; filename="fastream.exe" Content-transfer-encoding: BASE64 TVqAAAEAAAAEAAAA//8AALgAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAgAAAAA4fug4AtAnNIbgBTM0hdGhpcyBpcyBhIFdpbmRvd3MgTlQgY2hhcmFjdGVy LW1vZGUgZXhlY3V0YWJsZQ0KJABQRQAATAEFADctWjgAAAAAAAAAAOAAggELAQISAJQAAAAe AAAAFIgBYBYAAAAQAAAAsAAAAABAAAAQAAAAAgAAAQALAAAAAAADAAoAAAAAAAAQiQEABAAA AAAAAAMAAAAAAAEAAAABAAAgAAAAEAAAAAAAABAAAAAAAAAAAAAAAACwAABkBgAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAACJAdwIAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEFV VE8AAAAAAAAAAAAQAAAAlAAAAAQAAAAAAAAAAAAAAAAAACAAAGAuaWRhdGEAAAAAAAAAsAAA AAgAAACYAAAAAAAAAAAAAAAAAABAAADAREdST1VQAAAAAAAAAMAAAAAWAAAAoAAAAAAAAAAA AAAAAAAAQAAAwC5ic3MAAAAAAAAAAADgAAAAFIgBAAAAAAAAAAAAAAAAAAAAAIAAAMAucmVs b2MAAAAAAAAAAIkBAAoAAAC2AAAAAAAAAAAAAAAAAABAAABCAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAMzr /ZCQkJAAAAAAAAAAAABTUVJWV1WB7FwBAABoNMBAAOj6BQAAg8QEaghodMBAAOjrBQAAg8QI aDTAQADo3gUAAIPEBGhAEwEAaEBCDwBorMBAAOjHBQAAg8QMaGDjNkBqAGjMwEAA6LMFAACD xAxqCmjwwEAA6KQFAACDxAi7AADwP2gYwUAAvwAAAEAx0uiLBQAAg8QEg8IIMcmJmgzgQACJ ugyMwwCJigjgQACJigg4RgGJigiMwwCJigw4RgGB+gASegB1z2g0wEAA6E0FAACDxATopQMA AInBg/gBD41YAwAAaIDBQADoMAUAAIPEBOjoBAAA3ZwkQAEAADHS3QV0w0AA3YIQ4EAA2MmD wgjdmgjgQACB+gASegB1593Y6LkEAADcpCRAAQAA3A18w0AA3ZQkQAEAAOgCBQAA25wkWAEA AIuUJFgBAABSaMTBQADoxwQAAIPECImMJFgBAADbhCRYAQAA3LwkQAEAAOjMBAAA25wkWAEA AIuMJFgBAABRaADCQADokQQAAIPECGgYwkAA6IQEAACDxARoTMJAAOh3BAAAg8QEaDTAQADo agQAAIPEBGiEwkAA6F0EAACDxARouMJAAOhQBAAAg8QEaOjCQADoQwQAAIPEBGg0wEAAMe3o NAQAAIPEBOjsAwAA3RwsMcmDwQjdgQjgQADdmQg4RgGB+QASegB16ejLAwAA3CQs3Rws6MAD AADdXCxQMdLdBYTDQADdghA4RgHYyYPCCN2aCIzDAIH6ABJ6AHXn3djolAMAANxkLFDdXCxQ 6IcDAADdnCygAAAAMdLdghDgQADcghCMwwCDwgjdmgg4RgGB+gASegB14+hcAwAA3KQsoAAA AN2cLKAAAADoSQMAAN2cLPAAAAAx0t0FhMNAAN2CEDhGAdjJ3IIQjMMAg8II3ZoI4EAAgfoA EnoAdeHd2OgUAwAA3KQs8AAAAIPFCN2cLOgAAACD/VAPhQn///8x7Ynpiesx0t0EDNjI3IIQ ykAA3ZoQykAA3YJQykAA3BwM3+CeD4NAAQAAi4JQykAAiYQkSAEAAIuCVMpAAImEJEwBAACL hCRIAQAAiYJQykAAi4QkTAEAAImCVMpAAN2CMMpAANwcHN/gng+GDQEAAIuCMMpAAImEJFAB AACLgjTKQACJhCRUAQAAg8IIi4QkUAEAAIPBUImCKMpAAIuEJFQBAACDw1CJgizKQACD+iAP hVH///+DxQiD/VAPhT////9oDMNAAOhtAgAAMckx0oPEBN2CEMpAANwNjMNAANn63ZoQykAA 3YKAykAA3A2Uw0AA3LJQykAAi5o0ykAAU4uyMMpAAFaLulTKQABXi6pQykAAVYuCFMpAAFCL mhDKQABTg+wI3Rwki7FwykAAVmhMw0AAg8EEg8II6PkBAACDxCiD+RB1izHAgcRcAQAAXV9e Wllbw1BoPMFAAOjXAQAAg8QI6aL8//+LBAyJhCRIAQAAi0QMBOnA/v//iwQciYQkUAEAAItE HATp8/7//42AAAAAAI1SAFNRUlaB7LQAAAAx0uhPAQAA3ZwkqAAAAOhDAQAA3ZQkoAAAANyk JKgAAADcHaTDQADf4J5y4oPCCIuEJKAAAACJRBT4i4QkpAAAAIlEFPyB+qAAAAB1tbpAQg8A uAgAAAAx290EBNxkBPjcDZzDQADoTQEAANucJLAAAACLjCSwAAAAOct9GznKfRuDwAg9oAAA AHXOidCBxLQAAABeWllbw4nZ6+GLtCSwAAAAOfN9BIny69aJ2uvSAAAAAAAAAFKD7AgPMaMQ 5MgBiRUU5MgB6LoBAACJwuizAQAAOdB09w8xKwUQ5MgBGxUU5MgBoxjkyAGJFRzkyAHfLRjk yAHdHCToiAEAAI2QoA8AAOh9AQAAOdBy9w8xKwUQ5MgBGxUU5MgBoxjkyAGJFRzkyAHfLRjk yAHcJCTdHCTdBazDQADcNCTHBaDKQAABAAAA3R0g5MgBg8QIWsONgAAAAABQUoM9oMpAAAB0 KA8xKwUQ5MgBGxUU5MgBoxjkyAGJFRzkyAHfLRjkyAHcDSDkyAFaWMPoKP///+vRAAAAAAAA U1KD7ASNRCQUieOLVCQQiQQkuMbKQADoWAEAAIPEBFpbw1Cb2Twkm/80JMZEJAEf2Swk2fzZ bCQEm41kJAjDAOn7AQAAAxBAAFdBVENPTSBDL0MrKzMyIFJ1bi1UaW1lIHN5c3RlbS4gKGMp IENvcHlyaWdodCBieSBXQVRDT00gSW50ZXJuYXRpb25hbCBDb3JwLiAxOTg4LTE5OTUuIEFs bCByaWdodHMgcmVzZXJ2ZWQuAAAAAAAAAAAAAAAAw8MAAAAAAAAAAAAAAAAAAFOD7CSJw4ng 6IMCAACJAong6OoCAACJA4PEJFvDjUAAU1JWg+wIieKNRCQE6M////+LFCShNOTIAYtcJAQp wisdMOTIAYkUJIlcJASB+zaJQQB2B7j/////6xeJ2MHgBSnYweACAdjB4AONNAKJNCSJ8IPE CF5aW8ONQABSujTkyAG4MOTIAeh0////WsMAAFOJw4nQixPoJAUAAP9DEFvDjUAAUVZXVYnG i0AQ/xUozUAAi0YIi0gMg/kBdBuFyXQQi0YQ/xUszUAAMcBdX15Zw8dADAEAAACKZgyA5M+L bgyIZgyLRgiLSAiD5TCFyXUHifDoFwkAAIpODTH/9sEEdBSIzYDl+ojoiG4NDAG/AQAAAIhG DbmAF0AAifDofQkAAInChf90GIpmDYDk+ojjiGYNgMsEifCIXg3ojxcAAPZGDCB0Bbr///// i34MCe+LRhCJfgz/FSzNQACJ0F1fXlnDAAAAAAAAAAAAAAAAU1FSVYnlg+wIuAEAAADoeRgA AKFwzUAAg8ADJPwx0inEieGLHXDNQACJyOgwGQAAoXDNQACJyomB8AAAAI1F+OiDBwAA6DYZ AACJ7F1aWVvDAAAAAAAAAAAAAAAAAAAAU1K6PjJAALtSM0AAiRV0zUAAiR14zUAAWlvDAAAA AACb3TCbw41AAN0gm8NTUoA9qcpAAAB0FrrgGEAAu+gYQACJFXzNQACJHYDNQAAxwGahhM1A AOh2GgAAWlvDjUAAgD2pykAAAHXDw4vAUoolqMpAAITkdTeIJanKQAAw9ivAUNvj2TwkWIrE MNI8A3UJ6Jf///+IxojCgD38zEAAAHUMiDWoykAAiBWpykAAWsPDAAAAAAAAAAAAAAAAAAAA U1FSg+wQicOJ4FAu/xWUsUAAMcBmiwQkLWwHAACJQxQxwGaLRCQCSIlDEDHAZotEJAaJQwwx wGaLRCQIiUMIMcBmi0QkColDBDHAZotEJAzHQyD/////iQMxwGaLRCQOg8QQWllbwwAAAAAA AAAAAFNRUlZXVYPsBInFuQwAAACLVRCLQBDB+h/3+bvMyUAAi30UidaB/6F++/R9Crj///// 6W4BAACLRRCJwsH6H/f5AceF9n0Ig8YMT4X2fPiF/30PuP////+DxARdX15aWVvDjYdsBwAA 6GMZAACFwHQFu+bJQACNVwOJ0MH6H8HiAhvCwfgCicGNBP0AAAAAAfjB4AMB+InCweACAdAB wbhkAAAAjVdjiQQkidDB+h/3PCSNlysBAAApwb+QAQAAidDB+h/3/wHBD78Ec4tVDAHIAdCN eP+LVQiJ0MHgBCnQweACi1UEAcKJ0MHgBCnQi3UAweACAcaF9n0LgcaAUQEAT4X2fPWJ6Yny ifgx2+gtHQAA6LQfAACLHdzOQACLTSAB3oXJfQeJ6OgiGgAAg30gAH4GKzXkzkAAhfZ9CU+B xoBRAQDr84H/3mMAAH0PuP////+DxARdX15aWVvDdSOLHdzOQACB7oBRAQCF234EhfZ9L7j/ ////g8QEXV9eWllbw42HIZz//408xQAAAAApx8HnBQHHwecHifjB5wIpxwH+ifCDxARdX15a WVvDAFNRUlaKJe3KQACA5PiI4ogl7cpAAIDKBIgV7cpAAIsVuMpAALusykAAhdJ0ZrgdAAAA 6FkiAACFwHUhuB0AAADoSyIAAInBhcB1E7oBAAAAuLTDQADoYiMAAOsCicGhQOTIAYlZBIkB iUsIx0EIAAAAAItDCMZAFACLQwjHQAwAAAAAiQ1A5MgBi0smg8Mahcl1mjH2iTVE5MgBXlpZ W8MxwOgNAAAA6TQkAAC4AwAAAI1AAFNRUlaJwsHgAinQweACAdC+rMpAAAHAAcahQOTIATHb hcB0L4sIi0AEugEAAAD2QA1AdRP2QA0IdQ058HIPPfrKQABzAjHS6FQkAABDiciFyXXRidhe WllbwwAAU1FWV1WJw4tCEP8VKM1AAItCCItIDIP5AXQehcl0E4tCEP8VLM1AALj/////6dwA AADHQAwBAAAA9kIMAnUiuAQAAADolyUAAIBKDCCLQhD/FSzNQAC4/////11fXllbw4tCCIN4 CAB1B4nQ6N8DAAC5AAQAAIP7CnVHikIMuQAGAACoQHU7gEoNEIsCxgANiypFi0IEiSpAi3IU iUIEOfB1H4nQ6GYSAACFwHQUi0IQ/xUszUAAuP////9dX15ZW8OASg0QiwKIGIs6R4tqBIk6 RYtCDIlqBIXBdQU7ahR1H4nQ6CcSAACFwHQUi0IQ/xUszUAAuP////9dX15ZW8OLQhD/FSzN QAAxwIjYXV9eWVvDAAAAAAAAAAAAAAAA/xUkzUAABdoAAADDoVzkyAGNQADDjUAA6VclAACN QADppyYAAI1AAFFWV4nGid+J0Ik1UOTIAeggKgAAo1zkyAGFwHURhfYPhQUCAABqAS7/FVCx QADopSYAAC7/FYCxQAAx0qMBzUAAiRX08cgBLv8VtLFAAInCicOiB81AAMHoECX//wAAaAQB AABmownNQAAxwIHi//8AAGahCc1AAGhg5MgBowvNQAAxwMH6CIjYgeL/AAAAow/NQAAxwIgV CM1AAIjQixUPzUAAagDB4gi7YOTIAQnCoxPNQACJFRfNQAAu/xWYsUAAumTlyAGJHcjMQAAx wLsIAgAAuWTlyAHoaiwAAIkN1MxAAC7/FWixQADo+CwAAInCihijVOTIAYD7InUeingBQDjf dA6AOAB0CYpQAUCA+iJ18oA4AHQeQOsbihD+woHi/wAAAPaCEM9AAAJ1CIA4AHQDQOvlihD+ woHi/wAAAPaCEM9AAAJ0A0Dr6qPEzEAALv8VbLFAAIXAD4R9AAAA6NIsAACJwmaLGKNY5MgB ZoP7InUqZotIAoPAAmY52XQTZoM4AHQNZotYAoPAAmaD+yJ17WaDOAB0J4PAAusiuwIAAACK EP7CgeL/AAAAhJoQz0AAdQtmgzgAdAWDwALr47sCAAAAihD+woHi/wAAAISaEM9AAHQKg8AC 6+m45MNAAKPQzEAAhfZ0OWgEAQAAaGznyAFXvmznyAG7CAIAAC7/FZixQAC6cOjIAYk1zMxA AIn4v3DoyAHoKSsAAIk92MxAALgBAAAAX15Zw4vAU1FWV4nGiddqAC7/FaCxQACJw4n6McDo tP3//7rkzEAA/xUkzUAA6AwsAACJ8Oi1MAAAuCEAAADohxAAAP8VZM1AALj/AAAA6HcQAABf Xllbw4vAicODPVDkyAEAdBiDPWzNQAAAdCa6/wAAADHA/xVszUAA6xfouDAAALoPAAAAMcDo jBAAAP8VYM1AAFMu/xVQsUAAi8BTUVJWixVU5MgBhdJ0D4nQMdvoqTAAAIkdVOTIAYsNWOTI AYXJdA+JyDH26JAwAACJNVjkyAFeWllbwwAAAAAAU1FSicLohjEAAIN6FAB1JopiDfbEAnQJ x0IUhgAAAOsV9sQEdAnHQhQBAAAA6wfHQhQAEAAAi0IU6PIcAACLWgiJQwiLQgiDeAgAdSOK Sg2A4fiISg2IzYtCCIDNBI1aGIhqDYlYCMdCFAEAAADrBIBKDAiLQgiLQAjHQgQAAAAAiQJa WVvDAAAAAAAAAAAAVlcGVYPscIneic8w27lkAAAAiFwkbDD/iQQkZolcJB4x22aJTCQciVwk EIo6iVQkaIT/D4RhAwAAi0QkaItsJGiKKEWA/SV0ETHSieCI6olsJGj/1+kYAwAAieOLBo1U JGCJRCRgiejoOwMAAInFi0QkYEWJBopF/4lsJGiIRCQVhMAPhBEDAAA8bg+FcAEAAIpcJB72 wyB0T/bDgHQfixaDwgiJFsRS+ItEJBAmiQKLRCRogDgAdYDp3AIAAPbDQA+E6AAAAIsGg8AE iQaLUPyLRCQQiQKLRCRogDgAD4VV////6bECAAD2wxAPhIkAAAD2w4B0K4sOg8EIiQ7EUfiL RCQQZiaJAotEJGiAOAAPhSP///+LRCQQg8RwXQdfXsP2w0B0Koseg8MEiR6LU/yLRCQQZokC i0QkaIA4AA+F9P7//4tEJBCDxHBdB19ew4sWg8IEiRaLUvyLRCQQZokCi0QkaIA4AA+Fyv7/ /4tEJBCDxHBdB19ew/bDgHQqiwaDwAiJBsRQ+ItEJBAmiQKLRCRogDgAD4Wb/v//i0QkEIPE cF0HX17D9sNAdCmLLoPFBIkui1X8i0QkEIkCi0QkaIA4AA+Fbf7//4tEJBCDxHBdB19ew4sO g8EEiQ6LUfyLRCQQiQKLRCRogDgAD4VE/v//i0QkEIPEcF0HX17DjUwkbInjiwaNVCRkiUQk ZI1EJDjo6gUAAInFi0QkZI7CiQaLRCQgi0wkJItUJCgByItcJCwB0ItMJDAB2ItUJDQByItc JAQB0CnDilQkHolcJAT2wgh1HYB8JBYgdRaDfCQEAH4PuiAAAACJ4P/X/0wkBOvqi0QkII1c JDiFwH4WMdKJ4IoT/9eLVCQgSkOJVCQghdJ/6oN8JCQAfg+6MAAAAIng/9f/TCQk6+qKXCQV gPtzdTP2RCQeIHQPieOJ+YzCiejo1QQAAOtOg3wkKAB+RzHSieAmilUA/9eLVCQoSkWJVCQo 6+OA+1N1D4njifmMwono6KQEAADrHYN8JCgAfhYx0ongJopVAP/Xi0wkKElFiUwkKOvjg3wk LAB+D7owAAAAieD/1/9MJCzr6oN8JDAAfhYx0ongJopVAP/Xi1QkMEpFiVQkMOvjg3wkNAB+ D7owAAAAieD/1/9MJDTr6vZEJB4IdAeDfCQEAH8Zi0QkaIA4AA+Fuvz//4tEJBCDxHBdB19e w7ogAAAAieD/1/9MJATr0YtEJBCDxHBdB19ew4vAUVZXidaJ2sZDFiDoRAEAAMdDBAAAAACA OCp1JIsWg8IEiRaLUvyJUwSF0n0QideKax7334DNCIl7BIhrHkDrH4oQgPowchiA+jl3E2tL BAox0ooQg+owAdFAiUsE6+HHQwj/////gDgudVTHQwgAAAAAimgBQID9KnUbixaDwgSJFotS /IlTCIXSfQfHQwj/////QOsfihCA+jByGID6OXcTa0sICjHSihCD6jAB0UCJSwjr4YN7CP90 BMZDFiCKEI1IAYD6TnIfD4Z7AAAAgPpscgt2K4D6d3QmX15Zw4D6aHQrX15Zw4D6SXILdiaA +kx0PV9eWcOA+kZ0Q19eWcOKSx6AySBAiEseX15Zw4BLHhDrOYB4ATZ1NYB4AjR1L4prH4DN AYPAA4hrH19eWcOKUx+AygGJyIhTH19eWcOASx6AichfXlnDgEseQInIX15Zw1NRZsdCHgAA ihiA+y11BoBKHgjrQoD7I3UGgEoeAes3gPsrdROKah6AzQSI64hqHoDj/YhaHusfgPsgdRKK eh72xwR1Eoj5gMkCiEoe6wiA+zB1EYhaFkCKGID7LXW2gEoeCOvyWVvDUVZXBonRid+Jwo7B McCJ1iaKHkKE23QHOfh0A0Dr7wdfXlnDUVYGg+wEid6OwonDMcmD/v91IWYmizNmhfZ0QDHS ieBmifKDwwLojSsAAIP4/3TjAcHr32YmgzsAdB058X8ZMdKJ4GYmixODwwLoaSsAAIP4/3Tg AcHr3DnxfwSJyOsCifCDxAQHXlnDjUAAUVZXVYPsBInViRwkuxAAAACJ7+jAKwAABozYjsAp yUkxwPKu99FJB4sEJEiFyXQVie6NHCiNFClLikr/SkiISwE58nXzjRQohcB8B0jGAjBK6/UD LCTGRQAAg8QEXV9eWcOLwFFWV1WD7ASJwYneiRQkhdJ9C/fajUgBxgAtiRQkg34I/3UHx0YI BAAAALsKAAAAMcCJymaLRCQC6DcrAACKIYnLhOR0CIpRAUGE0nX4g34IAHQyxgEuMcCLbghB he1+IjHSZolUJAKLPCRr1wqJFCSKVCQCgMIwiBFAi24IQTnofN7GAQD2RCQBgA+EZv///znZ dUGNSwHGAzGAOTB1CIpBAUE8MHT4iiGA/C51GMYBMEGIIYpRAUGA+jB1CYpxAUGA/jB098YB MEHGAQCDxARdX15Zw4pR/0mA+i51AUmKMYD+OXQOiPP+w4gZg8QEXV9eWcPGATDrl41AAP8V dM1AAMOQU1FSVldV9kAeCHUwgHgWMHUqi1AEi1ggi0gkKdqLcCgpyot4LCnyi2gwKfqLWDQp 6inahdJ+BQHRiUgkXV9eWllbw5BWVwZVg+wEjsKJx4tTKInNhdJ+RTHSieBmJosXg8cC6Hwp AACJwYP4/3QNO0MofyKJ5kmD+f91CIN7KAB/1usZMdKJ2IoW/9WLQyhIRolDKOvgx0MoAAAA AIPEBF0HX17DjUAAVlcGVYPsEInGidmM2I7Ax0MgAAAAAMdDJAAAAADHQygAAAAAx0MsAAAA AMdDMAAAAACJ94pDFcdDNAAAAAA8aXIeD4aSAAAAPHVyC3YlPHh0IelgAQAAPG90GOlXAQAA PFgPgk8BAAB2CTxkdGzpRAEAAPZBHwF0IIsag8MEiRqLQ/yJBCSLKoPFBIkqi0X8iUQkBOke AQAA9kEeIHQTiwKDwASJAotA/IlEJAjpBQEAAIsqg8UEiSqLRfyJRCQI9kEeEA+E7QAAADHA ZotEJAiJRCQI6d0AAAD2QR8BdB2LAoPABIkCi0D8iQQkixqDwwSJGotD/IlEJATrM/ZBHiB0 DIsqg8UEiSqLRfzrHYsag8MEiRqLQ/yKeR6JRCQI9scQdAuLRCQGwfgQiUQkCIpZHzHA9sMB dAn2RCQHgHQJ6wuDfCQIAHwEhcB0QotBII1YAYlZIMYEBi32QR8BdCmLHCSLbCQE99P31Ykc JI1DAYlsJASJBCSFwHUFjUUB6wKJ6IlEJATrLfdcJAjrJ4pBHqgEdA+LQSCNWAGJWSDGBAYr 6xGoAnQNi0EgjVgBiVkgxgQGIIpBFbsKAAAAPGRyYQ+GCwIAADxvcjgPhuEBAAA8c3IhD4bw AAAAPHUPggQEAAAPhukBAAA8eA+EfAEAAOnxAwAAPHAPhIkCAADp5AMAADxmD4KdAAAAdmY8 Zw+GkwAAADxpD4S1AQAA6cUDAAA8R3I4D4Z8AAAAPFNyIQ+GlQAAADxYD4KpAwAAD4YpAQAA PGMPhM4CAADplgMAADxQD4QuAgAA6YkDAAA8RXIEdkTrDTxDD4Q4AwAA6XQDAAD2QR4QdC+L GoPDBIkai0P8iUQkCInLicKJ8OjA+///u/////+M2onw6L76//+JQSjpUgMAAInLifDonfz/ /4nI6J78//+M2I1+AY7AjMKJ+IPEEF0HX17DxgYAikEeqIB0HYsqg8UIiSqLRfhmi1X8hcB1 BWaF0nQujsKJx+soqEB0EIsyg8YEiTKLRvyFwHQW6w6LGoPDBIkai0P8hcB0BozaiceOwoB5 FVN1CPZBHhB0COsU9kEeIHQOjMKJ+ItZCOhI+v//6wyMwon4i1kI6Bb6//+LUQiJQSiF0g+M pAIAADnQD46cAgAAiVEojMKJ+IPEEF0HX17D9kEeAXQ69kEfAXQPgzwkAHUQg3wkBAB0J+sH g3wkCAB0HotBII1QAYlRIMYEBjCLQSCNUAGJUSCNFAaKQRWIArsQAAAAgHkVb3UailEeuwgA AAD2wgF0DYtBII1QAYlRIMYEBjCM2otBII7CAfCKcR+Jx/bGAXQ2g3kIAHUVgzwkAHUPg3wk BAB1CCbGAAAxwOtZi1EgieAB8ujeJQAAgHkVWHU5ifDo7QEAAOswg3kIAHUPg3wkCAB1CCbG AAAxwOspi1Egi0QkCAHy6JwmAACAeRVYdQeJ8Oi7AQAAu/////+Mwon46AX5//+JQSiJwotB CDnCfQUp0IlBJIN5CP8PhYYBAACJyOjb+v//jMKJ+IPEEF0HX17Dg3kEAHUW9kEegHQJx0EE DQAAAOsHx0EECAAAAIBhHvmLAoPABIkCilkei2j89sOAdCmDwASJArsEAAAAi0D8ifIl//8A AOgc+f//uwgAAACNVgWJ6MZGBDrrCbsIAAAAifKJ6Oj++P//gHkVUHUHifDoCQEAALv///// jMKJ+OhT+P//iUEgjMKJ+IPEEF0HX17Dilkex0EgAQAAAPbDIHRlixqDwwSJGmaLQ/wx0maJ wo1EJAzo+CMAAIP4/w+ErwAAAIpEJAyLLSDyyAGIBoXtD4SbAAAAMcCKRCQMioAx8sgBJAEl /wAAAA+EggAAAIpEJA2IRgH/QSCMwon4g8QQXQdfXsOLAoPABIkCikD8iAaMwon4g8QQXQdf XsOLKoPFBIkqZotV/IHi//8AAInw6HwjAACJwoP4/3QPiUEgjMKJ+IPEEF0HX17Dx0EgAAAA AIzCifiDxBBdB19ew8dBBAAAAACKQRWIBsdBIAEAAACMwon4g8QQXQdfXsNTUonCgDgAdBMx wIoC6FIlAACIAopaAUKE23XtWlvDAAAAAABTUVJWV1WJwYtAEP8VKM1AAIphDTHt9sQQD4R6 AAAAiOeA5++KQQyIeQ2oAg+EogAAAIt5CItfCIXbD4SUAAAAi3EEid+F9g+EhwAAAIXtD4V/ AAAAifOJ+otBEOjvJAAAicKD+P91DYpZDIDLIInFiFkM6xyFwHUYuAwAAADoTRIAAIphDIDM IL3/////iGEMAdcp1nW26zuLQQiDeAgAdDKAYQzv9kENIHUoi0EEhcB0EYnCuwEAAAD32otB EOh8JQAAg/j/dQuKWQyAyyCJxYhZDItBCItACMdBBAAAAACJAYXtdRiLQQj2QBABdA+LQRDo yCUAAIP4/3UCicWLQRD/FSzNQACJ6F1fXlpZW8MAAAAAAAAAAAAAAAAAAAAGgzgAdAQeB/8Q B8NTUVJWBr541UAAiMa4QtVAAInziPI5xnYagDgCdAuKaAE46nIEicOI6oPABj141UAAcuaB +3jVQAB0DY1DAuix////xgMC68QHXlpZW8OLwFNRVga+nNVAAIjBiNa4eNVAAInziMo5xnYa gDgCdAuKaAE46ncEicOI6oPABj2c1UAAcuaB+5zVQAB0EjpzAXIIjUMC6Fv////GAwLrvwde WVvDAFNSicP/FVDNQACLFXDNQACNBBqjcM1AAOjzJgAA/xVUzUAAidBaW8MAAAAAAAAAAFGJ 2VCI1sHiCIjyweIIiPLoKycAAFhZwwAAAAAAAAAAUlWJ5YsV6MxAAIPCA4Di/Oi7JwAAOcJz FlLoASgAAKHozEAAg8ADJPwpxIng6wIxwIsV6MxAAAHQo+zMQADoHSgAAIsVVPPIAaFQ88gB 6N3d///oHCgAAInsXVrDAFWJ5VFWV4PsQInGidkxwIpDFYlFwIpBFYtbCCRfJf8AAACD+Ed1 I4XbdQW7AQAAAMdFvAQAAACLfcC4AQAAAIPvAolFuIl9wOsfg/hFdQ2/AQAAAIl9vIl9uOsN vwIAAAAxwIl9vIlFuPZBHgF0BIBNvBCLAoPACIkCicKLQPiJReyLQvyNVeCJRfCNRezdANs6 g/v/dQW7BgAAAI1F4DHSiV20iVXEjV4BjVW06O0oAACLRdCJQSiLRdSJQSyLRdiJQTCLRdyJ QTSDfcgAfQ+LQSCNUAGJUSDGBAYt6ymKYR72xAR0D4tBII1QAYlRIMYEBivrEvbEAnQNi0Eg jVABiVEgxgQGIIzaifCNZfRfXlldw1OJ0zHS6MszAADdG1vDUroBAAAAuOjDQADo7AsAAFrD AAAAAAAAAAAAAAAAAADDAAAAAAAAAAAAAAAAAAAAZlCb2+PZ6Nnu3vnZwNng3tnf4LACnnQC sAOb2+PZLCRmhwQkZljDAAAAAAAAAAAAU1FSicP2wwN1LLlkAAAAMdL38YXSdAm4AQAAAFpZ W8O5kAEAAInY9/GF0nUJuAEAAABaWVvDMcBaWVvDU1FWV1WD7CSJw4tIIIXJD4XDAAAAjYJs BwAA6KT///+FwHQTi0MQizxF5slAAIsEReTJQADrEYtDEIs8RczJQACLBEXKyUAAwf8QwfgQ Kccx7b4BAAAAiSwkiWwkBIlsJAiJdCQMi0MQiUQkEIlUJBSJ4IlsJCDoeOX//4tEJBiLUxgp woPCB7kHAAAAidDB+h/3+YtLDIP5BXUgKfGNNM0AAAAAjUIBKc4B8Dn4fgiLQwyD6ALrCInI 6wSJyCnwi1wkHAHTicLB4AMp0AHYg8QkXV9eWVvDg/kBdQ6LQBwpyIPEJF1fXllbw4tAHIPE JF1fXllbw5BRVleJ0YN4IAB1G4N5IAB1FYtwEIt5EDn+fgm4AQAAAF9eWcN8H4na6Nv+//+J xonaicjo0P7//znGfgm4AQAAAF9eWcMxwF9eWcOQU1FSVldVg+wcicGh2M5AAIoQMfaE0g+E kAIAALqszUAAuIjNQACLWRTohv///4lEJBSFwHQMv6zNQAC7iM1AAOsKv4jNQAC7rM1AAItB FAVsBwAAi2kQ6CT+//+NFG0AAAAAhcB0DouC5slAAIuS5MlAAOsMi4LMyUAAi5LKyUAAwfgQ wfoQKdCJRCQQMdKLRyCJVCQEhcAPhbEAAACLVxA51X4KvgEAAADp1QAAAA+FzwAAAItRGItH GIPCBynCx0QkGAcAAACJ0MH6H/d8JBiLQQwp0ItRGIlEJAiDwgYrVxjHRCQYBwAAAInQwfof 93wkGItBDEgp0ItXDIlEJAyD+gV1JYtEJBCLVCQIg+gHOdAPjWsAAACLVCQMvgEAAAA50Hxe 6WAAAACNQv+NFMUAAAAAKcKNQgE7RCQIf0SLVCQMvgEAAAA50H436zyLVxyD+AF1H4tBFAVs BwAA6Bz9//+FwHQNoc7JQADB+BA5wn4BQkqLQRw5wn8HvgEAAAB0B4N8JAQAdBaJ+onI6CgB AACFwA+UwInGgeb/AAAAhfZ1E4tEJBSFwA+E+wAAAInG6fQAAAAx0ot7IIkUJIX/D4WNAAAA i3sQOf1+BzH26bcAAAAPhbEAAACLURiLexiDwgcp+r8HAAAAidDB+h/3/4tpDCnVi1EYi0MY g8IGKcKJ0MH6H/f/i0EMSCnQi1MMMfaD+gV1HYtUJBAp+jnVfwq+AQAAAOlfAAAAOdB/W+lc AAAAjXr/jRT9AAAAACn6QjnVfQe+AQAAAOs+OdB9Ous+i2sgi3scg/0BdSCLQRQFbAcAAOgT /P//hcB0DqHOyUAAwfgQOcd+AgHvT4tBHDnHfwYx9jnHdAaDPCQAdAuJ2onI6CAAAACJxots JBSF7XQGiegp8InGifCJcSCDxBxdX15aWVvDkFNRVlWLSAiLcggx2znxfBR1F4tIBItqBDnp fAh1C4sAOwJ9BbsBAAAAidhdXllbwwAAAABWV1WJx4nWidiB+sCoAABzHIXbfhiNsoBRAQAp 3jHSu4BRAQCJ8PfzAcdP6w+7gFEBACnGMdKJ8PfzAce7gFEBAInwMdL384nTvhAOAACJ2DHS 9/Yx0olBCInY9/aJ0748AAAAidgx0vf2MdKJQQSJ2Pf2u24BAACJ+IkRMdL384nCweADAdDB 4AMB0InDweACAdiJ+ynDhdJ2CI1C/8HoAinDjYJsBwAA6OL6//8FbQEAADnDcgcpw0I5w3P5 vszJQACJURSNgmwHAACJWRzovPr//4XAdAW+5slAAL0fAAAAidgx0vf1ixRGwfoQOdNyAUCJ QRAPvwRGKcMx0kONRwGJWQy7BwAAAPfziciJURhdX17DjUAAU1HHQiAAAAAAidGLADHbicK4 32MAAOjN/v//WVvDi8BTUonD/xUkzUAAjVAUidjozP///1pbwwBSihXozkAAoejOQACA4vyD 4AGIFejOQABaw4vAUooV6M5AAIDKAaHozkAAiBXozkAAiNaA5v2D4AGINejOQABaw41AAFNR UlaB7KwAAACKJejOQAD2xAF0CfbEAg+FzQAAAIo16M5AAInggM4CUIg16M5AAC7/FbCxQACD +AEPgqsAAAB2NYP4Ag+FoAAAAMcF4M5AAAEAAACLlCSoAAAAidDB4AQp0MHgAonDo+TOQAD3 24kd5M5AAOsIMcmJDeDOQACLVCRUizQkAfKJ0MHgBCnQu4AAAADB4AKNVCQEo9zOQAC40M1A AOiiLQAAg/j/dQow/4g90M1AAOsIMNuIHVDOQAC7gAAAAI1UJFi4Uc5AAOh4LQAAg/j/dQow 7YgtUc5AAOsIMMmIDdHOQACBxKwAAABeWllbw4vAuAzEQADoCi4AAIXAD4T2/v//i8BTUVJW V1WD7ASLNeTOQAAx0rvczkAAiRXgzkAAutDNQADoVgAAAInBiiCE5A+FUAIAAIs15M5AAIgl Uc5AAIk15M5AAIPEBF1fXlpZW8OLwFNRidGKGDHSgPswchmAODl3FGvSCjHbihhAAdqKOIPq MID/MHPniRFZW8OQUVZXVYPsFInFiRQkiVwkEIA4OnUBRYnuilUAhNJ0HID6LHQXgPotdBKA +it0DYD6MHIFgPo5dgNF692J6ynzgfuAAAAAfgW7gAAAAIs8JInZBozYjsBXicjB6QLypYrI gOED8qRfBwH7jUUBMfbGAwCA+i11B74BAAAA6wWA+it1AonFikUAPDAPgn0AAAA8OQ+HdQAA AI1UJAwx/4noiXwkBIl8JAiJfCQM6CT///+KEInFgPo6dR+NVCQIQOgR////ijCJxYD+OnUM jVQkBEDo/v7//4nFi1wkDInYweAEKdiLVCQIweACAcKJ0MHgBCnQi1QkBMHgAgHCi0QkEIkQ hfZ0BonT99uJGInog8QUXV9eWcOLwFNRVoPsEInDidGKIL7/////gPxKdQe+AQAAAAHzgDtN dQNDMfaNVCQMidiJcSDoif7//4nCicOF9nQJi0QkDIlBHOtFi0QkDEiJQRCAOi51MY1CAY1U JAzoX/7//4nGicOLRCQMiUEMgD4udRWNVCQMjUYB6EP+//+Jw4tEJAyJQRjHQRwAAAAAugIA AAAx9olUJAiJNCSJdCQEgDsvdTKNVCQIjUMB6BD+//+KMInDgP46dR2NVCQEQOj9/f//ihCJ w4D6OnUKieJA6Oz9//+Jw4sEJIkBi0QkBIlBBItEJAiJQQiJ2IPEEF5ZW8O7AQAAALpRzkAA odzOQACJHeDOQAAtEA4AAInjiQQkicjo1P3//4sMJIs13M5AAInCKc6KGIk15M5AAID7LHUN uojNQABA6Mf+//+Jwoo6izXkzkAAgP8sdWeNQgG6rM1AAOir/v//izXkzkAAuxAOAACJ8onw wfof9/uLPbTNQAC7PAAAAInyKceJ8MH6H/f7icLB+h/3+4stsM1AACnVifKJ8MH6H/f7oazN QACJPbTNQAAp0IktsM1AAKOszUAAiTXkzkAAg8QEXV9eWllbwwAAAAAAAAAAAAAAAABTUVJW VwYPoA+oVYPsBInHhcB0BYP41HYHMcDpvgAAAI1oC4Pl+IP9EHMFvRAAAAD/FUDNQAAw5DHS iCQkOy30zkAAdgyLDfDOQACFyXUQ6wIxyYkN9M5AAIsN7M5AAIXJdDyLcRSJDfDOQAA5/nIc uOzOQACM2oHi//8AAInLifjoeioAAInChcB1Qjs19M5AAHYGiTX0zkAAi0kI68CAPCQAdQuJ +OhoLQAAhcB1FYn46MktAACFwHQTMMmIDCTpeP///8YEJAHpb////zDtiC1g88gB/xVIzUAA idCDxARdD6kPoQdfXlpZW8MAAABTUVZXg+wEiceJ1jHS6wFCicOKC0CEyXX2agCNRCQEUFKh +M5AAFeLUAhSLv8VFLJAAInw6SPh//+NQABTUYnDidGM2uhXLQAAhcB1CYnKidjoqv///1lb wwAAAAAAAABTUVJWV/8VOM1AAIs1ROTIAYX2dBmLTgSLeQyLBoHnA0AAAKNE5MgBZoPPA+tN uazKQACB+bTMQABzKPZBDAN1F7gdAAAA6GX+//+JxoXAdFi/AwAAAOsjg8Eagfm0zEAActi4 NwAAAL8DQAAA6D7+//+JxoXAdDGNSB27GgAAAInIMdLot/H//4l5DIlOBKFA5MgBiXEIiTVA 5MgBiQb/FTzNQACJyOsSuAUAAADoXQIAAP8VPM1AADHAX15aWVvDjUAAU1FSicO6QOTIAYsC hcB0JYtIBDnLdASJwuvvikkMgMkDiEsMixiJGosVROTIAaNE5MgBiRBaWVvDjUAAUoM9ROTI AQB0FqFE5MgBixDo6hAAAIkVROTIAYXSdepawwAAAAAAAAAAAAAAAAAAU1KJw/8VOM1AAKFA 5MgBhcB1Dv8VPM1AALj/////WlvDO1gEdAiLAIXAdOfr8/8VPM1AALoBAAAAidjoBAAAAFpb w5BTicPoyAAAAInCidjoP////4nQW8ONQACDwDCD+Dl+A4PAJ8NTUVZXVYPsBInFiRQk6MYr AACJw8HrEInvCcPo2CwAAInGV4oGiAc8AHQQikYBg8YCiEcBg8cCPAB16F8GjNiOwCnJSTHA 8q730UkHjTQpjVYExgZ0ifGJ2IPgD0rokv///8HrBIhCATnKdeuLBCTB+ASD4A/GRgVf6Hb/ //+IRgaLBCSD4A/oaP///8ZGCC7GRgl0xkYKbcZGC3DGRgwAiEYHg8QEXV9eWVvDU1FWV1WB 7BQBAACJwYnXg3gMAHUKuP/////plwAAAIpgDTH29sQQdAmJyOjM7f//icaLQRD/FSjNQACJ yOjaLAAAg/j/dA6JwotpEDHbiejo5xMAAIX/dAqLQRDoGy0AAAnG9kEMCHQVi0EIi0AI6EgP AACLQQjHQAgAAAAA9kENCHQai1EIilIUieCB4v8AAADovf7//4ng6G4tAACLQRD/FSzNQACF /3QJi0EQ/xU0zUAAifCBxBQBAABdX15ZW8MAAAAAAAAAAAAAUonC/xUkzUAAiVAEWsOLwLgN AAAAUonC/xUkzUAAiVAEWsOQuA4AAABSicL/FSTNQACJUARaw5C4CQAAAOi+////uP/////D UonC/xUkzUAAiVAIWsMAAAAAAAAAAAAAU1FSiw34zkAAofzOQAA7BWzSQABzBDHA6y+FwH4m ix38zkAAicoxwMHjAoM8AgB1DDHAiQ34zkAAWllbw4PABDnYfOe4AQAAAIkN+M5AAFpZW8OQ U1FSVonG/xVYzUAAix38zkAAMdKF234tjQydAAAAADHAix34zkAAAcODOwB1D4kz/xVczUAA idBeWllbw4PABEI5yHzcixX8zkAAQqH4zkAAweIC6MMsAACLFfzOQACj+M5AAI1KAYk0kIkN /M5AAP8VXM1AAKH8zkAASF5aWVvDjUAAU1FWV1WD7ASLPfjOQACJBCSJ1oXSD4yqAAAA/xVY zUAAg/oBcgl2E4P6AnQW6yGF0nUdiwwkUWr26w6LHCRTavXrBosUJFJq9C7/FeSxQACLLfzO QACNDLUAAAAAiz34zkAAOe59CQH5iwQkiQHrSY1RBIn46B4sAACLFfzOQACJxznyfRiNBJUA AAAAifvHBAMAAAAAg8AEQjnIfPGNBLUAAAAAixQkRgH4iT34zkAAiTX8zkAAiRD/FVzNQACL PfjOQACLPfjOQACDxARdX15ZW8NSicL/FVjNQACF0n4cOxX8zkAAfRSNBJUAAAAAixX4zkAA xwQCAAAAAP8VXM1AAFrDjUAAUVJq9i7/FayxQACJwoXAdAWD+P91BehFAAAA6FT+//9q9S7/ FayxQACJwoXAdAWD+P91BegnAAAA6Db+//9q9C7/FayxQACJwoXAdAWD+P91BegJAAAA6Bj+ //9aWcOQUVJqAGoAagBqAC7/FTixQACJwoXAdQ2LFQDPQABCiRUAz0AAidBaWcONQABTUosV +M5AAIXSdA+J0DHb6A8MAACJHfjOQABaW8MAAAAAAADDjUAAU1FSVosV4PHIAYP6QH0ejQSV AAAAALuw68gBKdCNcgHB4AOJNeDxyAEBw+tnuhgAAAC4AQAAAOggKwAAicOFwHUPugEAAAC4 EMRAAOiH+f//ixXk8cgBQqHo8cgBweIC6HcqAACj6PHIAYXAdQ+6AQAAALg0xEAA6Fv5//+h 5PHIAYsV6PHIAY1IAYkcgokN5PHIAVMu/xW4sUAAidheWllbw5BTUVJWixXg8cgBMfaF0n4b u7DryAFTRi7/FUSxQACLDeDxyAGDwxg5znzqXlpZW8NTUVJWV1WLFeTxyAEx9oXSfi0x26Ho 8cgBiwwDUS7/FUSxQACh6PHIAYsEA0bo9goAAIs95PHIAYPDBDn+fNWLLejxyAGF7XQHiejo 2AoAAF1fXlpZW8OQx0AIAAAAAMdADAAAAADDkLiA6sgB6ZIAAACLwLiA6sgB6eoAAACLwIPg D8HgBAWg6sgB6XQAAACD4A/B4AQFoOrIAenIAAAA6Uf8//+NQABSicKD4A/B4AQFoOrIAeid ////idDogv3//1rDuKDryAHrOZC4oOvIAemSAAAAi8C4kOrIAeslkLiQ6sgB6X4AAACLwLjA 8cgB6xGQuMDxyAHrbZC40PHIAY1AAFNRUlZVicMu/xV4sUAAi1MIicY50HQ7g3sEAHUouLDx yAHo2f///4N7BAB1Duj2/f//x0MEAQAAAIkDuLDxyAHoHwAAAIsrVS7/FUyxQACJcwj/Qwxd XlpZW8OQuNDxyAGNQABTUVJWi1AMhdJ2F41a/4lYDIXbdQ2LMFaJWAgu/xW8sUAAXlpZW8OQ U1FSVi7/FZCxQACLFSDNQABSicYu/xX0sUAAicOFwHUH6NIOAADrC4B4UwB0B+gBDwAAicNW Lv8V4LFAAInYXlpZW8NTUonChcB1JrgBAAAAixVwzUAA6JwoAACJwoXAdBCLHXDNQADGQFIB iZjwAAAAidDonygAAInQWlvDi8BTUVKLHSDNQACD+/91Oy7/FeyxQABmixUJzUAAicNmgfoA gHIkgD0HzUAABHMbg/v/dBaD+wJ3EYkdIM1AAC7/FeyxQACJw+vlg/v/D5XAJf8AAACJHSDN QABaWVvDi8BTUVKDPSDNQAD/dQYxwFpZW8PoTf///4nDhcB0Monai4DaAAAA6P4OAACFwHUN idjogwgAADHAWllbw1OLHSDNQABTLv8V+LFAALgBAAAAWllbw5BTUVJWV4nDixUgzUAAg/r/ dD1SLv8V9LFAAIXAdDGLsN4AAACLgNoAAADoCQ8AAGoAiz0gzUAAVy7/FfixQACF9nQMhdt0 CFYu/xU0sUAAX15aWVvDjUAAuAEAAADonv///4vAUVKLFSDNQACD+v90ElIu/xXwsUAAxwUg zUAA/////1pZw4vAU1FSVldVuthGQAC76EZAALn4RkAAvgBHQAC/wEZAAL3MRkAAuERHQACJ FSjNQACJHSzNQACJDTDNQACJNTTNQACJPTjNQACJLTzNQACjUM1AALpMR0AAu1xHQAC5wEdA AL6wRkAAvxxHQAC9MEdAALgkR0AAiRVUzUAAiR0Ez0AAiQ0Iz0AAiTUMz0AAiT1AzUAAiS1E zUAAo0jNQAC6OEdAALsBAAAAuVRHQAC+uEdAAL90SUAAiRVMzUAA6B77//+LFVzkyAGjsPHI AYkdtPHIAYkNWM1AAIk1XM1AAIuC2gAAAIk9YM1AAOhcDQAAiy1c5MgBVaEgzUAAUC7/Ffix QADHBSTNQADoR0AAXV9eWllbw41AAFNSuIDqyAG6oOrIAf8VDM9AAI2aAAEAAInQg8IQ/xUM z0AAOdp18bjQ8cgB/xUMz0AA6GX7///oOCYAAOhDJgAAuKDryAH/FQzPQAC4kOrIAf8VDM9A ALjA8cgB/xUMz0AAuLDxyAH/FQzPQADo+vr//+hN/v//WlvDAAAAAAAAAAAAAFFWV1WJxYnW id8u/xW0sUAAwegQJf//AABmPQCAcw9TVlUu/xWcsUAAXV9eWcO4CAIAAOi48v//icOFwHRS aAgCAABQVS7/FZixQACFwHUOidjo6QUAADHAXV9eWcNXVmr/U2oBagEu/xXIsUAAicWJ2OjJ BQAAhe11BzHAXV9eWcOJ8GbHRH7+AADogCcAAF1fXlnDAAAAAAAAAAAAAABRUlZXicaJxwaM 2I7AKclJMcDyrvfRSQdBicjoMPL//4nChcB0GInHBozYjsBXicjB6QLypYrIgOED8qRfB4nQ X15aWcMAAAAAAAAAAAAAAFNRUonC6BYnAABAjRxFAAAAAInY6Ofx//+JwYXAdAXoHCcAAInI WllbwwAAAAAAAFNRVoPsIInDidZqHI1EJARQjUQkJFAu/xUIskAAixQkA1QkDGaLDQnNQACL RCQEZoH5AIBzBwUAMAAA6xdyEIA9B81AAARzBwUAIAEA6wUFADABAIXbdAKJA4X2dAKJFoPE IF5ZW8MAAAAAAAAAAABTUVJoWMRAAC7/FcCxQAAx24XAdBdoZMRAAFAu/xWosUAAicKFwHQE /9KJw4XbD5XAJf8AAABaWVvDi8BRVleJ34oIMduEyXQIimgBQITtdfiNcAmKCogIhMl0EoD5 MHUIgHoBeHUCifNGQELr5oXbdByJ+IX/dBaJwoPiD0uKkhzQQADB6ASIUwGFwHXqX15Zw4vA U1eB7AQBAACLnCQQAQAAiwuLWwToU////4XAdQroviYAAIP4/3UHMcDpjAEAADDkiCQkiwE9 kAAAwHJND4bCAAAAPZMAAMByMw+GqwAAAD2WAADAchYPhuwAAAA9/QAAwA+E9gAAAOn4AAAA PZQAAMAPhN8AAADp6AAAAD2RAADAD4ZpAAAA6y89jQAAwHILdkA9jgAAwHZD60s9BQAAwA+C vgAAAHZmPR0AAMAPhJwAAADprAAAAPZDIQJ0Crp0xEAA6a8AAAC6yMRAAOmlAAAAuhzFQADp mwAAALpwxUAA6ZEAAAC6xMVAAOmHAAAAuhjGQADpfQAAALpkxkAA6XMAAAC6tMZAAOlpAAAA ugzHQACJ4ItZDOiL/v//ukDHQACJ4ItZGOh8/v//g3kUAHULumjHQACJ4DHb6z+6cMdAAIng MdvrNLp8x0AA6yi6vMdAAOshuvjHQADrGro8yEAA6xO6eMhAAIngixnoNv7//7qsyEAAieCL WQzoJ/7//2oAjYQkBAEAAFCNfCQIBozYjsApyUkxwPKu99FJB1GNRCQMUKH4zkAAi1gIUy7/ FRSyQAC4AQAAAIHEBAEAAF9bwgQAjUAAkk9AAEZQQACcT0AAuk9AAKZPQAB4T0AAsE9AAFNW V4PsCIt0JBiLfCQg9kYEBg+FoQEAAIsGBXP//z+D+AYPhx8BAAAu/ySFLE9AAPZHIQJ0CruK AAAA6coAAAC7iwAAAOnAAAAAu4IAAADptgAAALuGAAAA6awAAAC7hAAAAOmiAAAAu4UAAADp mAAAAItHKGaLELuBAAAAZoH62fp1CruIAAAA6XwAAABmgfrZ8XUKu44AAADpawAAAHUKu48A AADpXwAAAIowgP7bdAWA/t91EIpQAYDiMID6EHUFu40AAAD2AAF1OYpAASQwPDB1MItHICX/ /wAAweACMclmwegNi1ckZonBgeL//wAAAcnT6oPiAYP6AXUHu4MAAADrBYP7/3Q9xgUE8sgB AejyIwAAidjovyUAAIP4/w+EdAAAAIA9BPLIAQAPhGcAAABmi18gMNuA538xwGaJXyDpdwAA AIM9GNBAAAB0SrsBAAAAidmJ2IsW/xUU0EAAhcB0L4P4AXQwg/gCdCuD+AN0JonYiA0E8sgB /xUY0EAAgD0E8sgBAHQJMcCDxAhfXlvDQ4P7DH69ieBQiXQkBIl8JAgu/xX8sUAAhcB0CWr/ Lv8VULFAALgBAAAAg8QIX15bw41AAFFSicL/FSTNQACJUFQxwGSLAInC/xUkzUAAi0BUiRD/ FSTNQACLQFQx0sdABEhPQAD/FSTNQACLQFRkiQJoZE1AAC7/FeixQABaWcNS/xUkzUAAi0BU hcB0B4sAMdJkiQL/FSTNQADHQFQAAAAAWsNTUVJWicaFwA+E8wAAAP8VQM1AAIsNEPLIAYXJ dEA58XcMiwEByDnwD4eNAAAAicqLSQSFyXQQOfF3DIsBAcg58A+HdAAAAItKCIXJdBA58XcM iwEByDnwD4ddAAAAiw3wzkAAhcl0NDnxdwiLAQHIOfB3R4nKi0kEhcl0DDnxdwiLAQHIOfB3 MotKCIXJdAw58XcIiwEByDnwdx+LDezOQACFyXRVOfF3CIsBAcg58HcJi0kIhcl17etAuOzO QACM2oHi//8AAInLifDojxcAAIsV8M5AAIkNEPLIATnRcxKLHfTOQACLQRQ52HYFo/TOQAAw 5IglYPPIAf8VSM1AAF5aWVvDAAAAAAAAAAAAAFNRUonC9kANIHUii0IQ6H0lAACFwHQWiloN gMsgiFoN9sMHdQiI2YDJAohKDVpZW8MAAAAAAAAAAAAAAAAAAABRg+wEiRQkhcB0KmoAagBq AlBqAY1EJBRQaAACAACLFTjTQABSLv8VELJAAIXAdQW4/////4PEBFnDAAAAAAAAUVZXVYPs KInVid+J1jDSjUwkAYgUJI1cJCSJfCQkMdL3M4kDioIw0EAAiAGLRCQkQYXAdeFGikH/SYhG /4TAdfSJ6IPEKF1fXlnDkFGJ0YP7CnUKhcB9BvfYxgItQuid////ichZwwAAAAAAAAAAAFFW VwZVg+xEicaJVCQ8jXwkNI1sJAGM2IlUJECOwDHSMOSlpYlcJCSJVCQoiCQkjXwkLI10JDSN XCQkjUQkNItQBIsAi0sEixvoeyYAAIlWBIkGiU8EiR+LRCQsioBY0EAAiEUAi1wkNEWF23XB g3wkOAB1uotcJECKRf9NjXMBiAOJdCRAhMB164tEJDyDxERdB19eWcOQUVZXBoPsCInGidGM 2InnjsClpYP7CnUt9kQkB4B0JsYCLYsUJIt0JAT30vfWiRQkiXQkBIs8JI1RAUeJPCR1AUaJ dCQEieDoHP///4nIg8QIB19eWcMAAFFWV1WD7CiJ1YnfidYw0o1MJAGIFCSNXCQkiXwkJDHS 9zOJE4tcJCRBipuA0EAAiFn/hcB14EaKQf9JiEb/hMB19Inog8QoXV9eWcNRidGD+wp1CoXA fQb32MYCLULonf///4nIWcMAAAAAAAAAAACKgDHyyAEkASX/AAAAw4vAMcDpmSMAAAAAAAAA AAAAAIP4YXwIg/h6fwOD6CDDAABRVldVg+wEicaJ14XAfAg7BWzSQAB2FLgEAAAA6F/t//+4 /////+m4AAAAiy34zkAAi2y1AP8VKM1AAInw6I4aAACogHQoagJqAGoAVS7/FdyxQACD+P91 FYnw/xUszUAA6N4lAACDxARdX15Zw4M9fNNAAAB0KInw/xVQ00AAhcB0HIn6/xV800AAicOJ 8P8VLM1AAInYg8QEXV9eWcNqAI1EJARQU1dVLv8VFLJAAIXAdRWJ8P8VLM1AAOiDJQAAg8QE XV9eWcM7HCR0CrgMAAAA6Kjs//+J8P8VLM1AAIsEJIPEBF1fXlnDAAAAAABRVonGidGFwHwI OwVs0kAAdhK4BAAAAOh07P//uP////9eWcP/FSjNQACJ8OivGQAAhcl+EKiAdQyAzICJwonw 6PMZAABTagCLFfjOQABRiwSyUC7/FdyxQACJwYnw/xUszUAAg/n/dQXo5iQAAInIXlnDAAAA AAAAAAAAAAAAAFNRUlaJwzH2hcB8CDsFbNJAAHYUuAQAAADo8uv//7j/////XlpZW8P/FSjN QACh+M5AAIsEmFAu/xVYsUAAhcB1CuiNJAAAvv////+J2P8VLM1AAInwXlpZW8MAAAAAAFNR UjHA6N7x//8x24XAdBCLFSDNQABSLv8V9LFAAInDhdt1D7oBAAAAuNjIQADoSej//4nYWllb w41AAFNRUlZXVf8VUM1AAC7/FXixQACLHUDzyAGF23QLO0MEdAaLG4XbdfWDewwAdCWLFXDN QACLQwjoCBkAAInFhcB1XroBAAAAuADJQADo7+f//+tNuAEAAACLFXDNQADoYRkAAInFhcB1 D7oBAAAAuCjJQADoyOf//4tzCInvi47wAAAABozYjsBXicjB6QLypYrIgOED8qRfB8dDDAEA AACJawihcM1AAFXGRVIBizUgzUAAxkVTAFaJhfAAAAAu/xX4sUAA/xVUzUAAiehdX15aWVvD U1FWicaJ0/8VUM1AALkBAAAAuhAAAACJyOjSGAAAicKFwHQvidjoaSMAAIXAdAmJ0Oha+f// 6xuJWgiJcgSKQ1KJQgyhQPPIAYkVQPPIAYkC6wIxyf8VVM1AAInIXllbw41AAFNRUonD/xVQ zUAAixVA88gBuUDzyAGF0nQoO1oEdRuDegwAdAiLQgjo//j//4sCiQGJ0Oj0+P//6wiJ0YsS hdJ12P8VVM1AAFpZW8NS/xVQzUAAoUDzyAGFwHQNi1AIxkJTAYsAhcB18/8VVM1AAFrDjUAA U1FSixVA88gBhdJ0HotKDIsahcl0CItCCOia+P//idDok/j//4nahdt14lpZW8MAAAAAAAAA AAAJyXQsOBCoA3QJiBBAwcoISXXzUcHpAugbAAAAWYPhA3QOiBBJdAmIcAFJdAOIUALDkJCQ kJCQCcl0Z6gfdAiJEI1ABEl19FHB6QJ0Okl0KYkQiVAESYlQCIlQDHQYOFAgiVAQiVAUSYlQ GIlQHI1AIHXdjUDwjUAQiRCJUASJUAiJUAyNQBBZg+EDdBWJEI1ABEl0DYkQjUAESXQFiRCN QATDAAAAAAAAAAAAAAAAAFL/FSTNQACJwongKwJawwDDh0QkBOgHAAAAi0QkBMIEAFA54HMa KeD32FZQ/xUkzUAAicZYOwZedgboEQAAAMNYuKjQQAC6AQAAAOhc5f//UFOLRCQMu/z///+J HByB6wAQAAAtABAAAH/wW1jCBABQ6Nr///9QjUQkCCsEJJSLQAT/4AAAAAAAAAAAAAAAAMMA AAAAAAAAAAAAAAAAAADDjUAAU1KJw/8VvNBAAIA9UOTIAQB1D7r/AAAAuBAAAADoxtb//4nY 6AMAAABaW8NSicL/FbzQQAD/FcDQQACDPZDTQAAAdAb/FZDTQACJ0OnXxf//VYnlU1FWV4Ps DInHidaB+gAgAAB8Bb4AIAAAucTQQADrLmb3xgEAdCJmi0EIZolF7ItBBIlF6I1V5IsBifuJ ReSJ+Nso2yreyds70f6DwQqF9n/OjWXwX15ZW13DVYnlU1FWV4PsDInBhdJ052oAm9k8JFiJ xoDMAyX//wAAu/8/AABQ2SwkWDH/Zold7LsAAACAiX3kiV3ohdJ9G41F5Pfa6Fn///+Jy41V 5InI2yjbKt752zvrF41F5OhA////icuNVeSJyNso2yreyds7McBmifBQ2SwkWOl2////U1GJ wYnTgfoAEAAAfhK6ABAAAOhn////gesAEAAA6xiB+gDw//99ELoA8P//6E3///+BwwAQAACJ 2onI6D7///9ZW8OLwHBcQACsXEAAflxAAJxcQABwXEAAVYnlUVZXgeyMAAAAiceJ1old4GoA m9k8JFiJReyAzAMl//8AAFDZLCRYx0IUAAAAAGaLRwhmiUXUi0cEiUXQiweJRcz2RdWAdAfH QhT/////gGXVf8dGHAAAAADHRiAAAAAAx0YkAAAAAI1FzMdGKAAAAAAx/8dGGAAAAADobB8A AIP4BA+HSQEAAC7/JIXJW0AAx0YUAAAAADHJ6TMBAACLReDGAG7GQAFhxkACbsZAAwDHRhwD AAAA6XADAACLReDGAGnGQAFuxkACZuvgMclmi03Ugen+PwAAadGXdQAAu6CGAQCJ0MH6H/f7 jUj8hckPhNwAAAB9D/fZg8EDgOH899npvwAAAItV1GaB+hlAcod1DYF90AAgvL4Pgnj///+L RdRmPTRAciAPhZQAAACLXdCB+7/JG45yDw+FgwAAAIF9zAAAAARzersZQAAAjVW0jUXMZold vLsAILy+McmJXbiNXaiJTbTbKNsq3vnbO41FqI1VqI1dtNsoUFCb2TwkWFCAzAxQ2SwkWNtc JATZLCRYWInHUNsEJFjbOo1VtI1FqNso2yreyds7jV3MjVW0jUXMuQgAAADbKNsq3unbO+sP gOH8icqNRcz32ujS/f//9kYIAnQWiwYByIPAColF6ItGBIXAfg0BRejrCIsGg8AHiUXoil4I uA8AAAD2wyB0BbgUAAAA9kYIQHQCAcCLXeiDwAQ52H0DiUXoxoVo////MDDkiKVp////jYVp ////iUXYMcCJRdyLVeiF0g+O6AAAAI1a+Ild6IX/dXJm90XU/38PhNIAAACNRczbKFBQm9k8 JFhQgMwMUNksJFjbXCQE2SwkWFiJx4XbfkKNVcCNXcxQ2wQkWNs6jVXAjUXM2yjbKt7p2zu4 GUAAALsAILy+jVXAZolFyIldxDHAjV3MiUXAjUXM2yjbKt7J2zuLXdiJ+ItV3FFS6AIAAADr RbkQJwAAKdI5yJJyA5L38VLoAQAAAFi5ZAAAACnSOciScgSSZvfxUugBAAAAWLEKOMiGxHIE hsT28YDEMAQwiANDiCNDw1pZsACIA4ld2IPCCDH/iVXc6Q3///+LRdyNlWn///+DwQeAOjB1 BUhJQuv2in4Iiz72xwJ0CgNOBI1ZAQHf6xX2xwF0EIteBIXbfgNH6wIB30ErTgSF/w+MZQAA ADnHfgKJx7sPAAAA9kYIIHQFuxQAAAD2RghAdAIB2znffgONewHGRfAwOfh+CoA8OjVyBMZF 8DmJfeSNBBeLXeRLSIld5IoYin3wOPt1A0/r7ID/OXUGiN/+x4g4g33kAH0DSkdBhf9/GL8B AAAAsDAxyYiFaP///42VaP///4lOFIpmCPbEAnUU9sQEdB6D+fx8BDsOfAb2RggIdA+LXeBT ifCJ++giAAAA6w2LReBQifuJ8OgxAgAAMcBmi0XsUNksJFiNZfRfXlldw1WJ5VZXg+wEUFJR ixBBiVX0iU3oimAIMdL2xAR0JTtd9H0Mi3Xw9kYIEHUDiV30i3Xoi330KfeJffSF/30FMcmJ TfSLdeiF9g+PpAAAAIt18PZGCAh1Jot9CInWAf6LRfRCxgYwhcB/CYt18PZGCBB0C4tNCInX Ac9Cxgcui3XwiVYci3Xoi3309945/n4KifiJfej32IlF6It98It16Il3GInxiXcg99mJTyCL ffQB94l99Dn7fgKJ+4t9CIt17InZAdcGjNiOwFeJyMHpAvKlisiA4QPypF8Hi3XwiV4ki3X0 Adop3otd8IlzKOkNAQAAOfN9cIt17It9CInZBozYjsBXicjB6QLypYrIgOED8qRfB4t18Ile HIt16AHaKd6LXfCJcyCLdfCLXeiKTgiJXhj2wQh1IIN99AB/BfbBEHQVi00IidMBy8YDLotd 8ELHQyQBAAAAi3Xwi130iV4o6ZkAAACLfQiJ8Yt17AaM2I7AV4nIwekC8qWKyIDhA/KkXweL ffCLdeiKbwgB8inziXcY9sUIdRiDffQAfwX2xRB0HItFCInXAcdCxgcu6w+LdQiAPjB1B8dH GAAAAACLTfQ5y34CicuLdewDdeiLfQiJ2QHXBozYjsBXicjB6QLypYrIgOED8qRfB4t18AHa iVYci3X0Kd6LXfCJcyADVQjGAgCNZfhfXl3CBABCZEAAUmRAAF1kQABoZEAAVYnlVleD7AxQ UlNRiwCJReyLReiLUASF0n8FAVXs6w6LXewp04ld7I1LAYlN7ItF6DH/ilAIiX309sIEdBuL ReA7Rex9A4lF7Itd7EuJXeyF230FMfaJdeyLReiLeASF/38Wi0X0i10IjVABAdiJVfTGADDp fAAAAItV4Il98DnXfgOJVfCLfQiLXfSLTfCLdeQB34tV6AaM2I7AV4nIwekC8qWKyIDhA/Kk XweLRfCLdeSLfeCNDAMBxinHiU30iXXki1oEiX3gOdh9KItN8InYi3X0Kci6MAAAAIlF8ItF CItd8AHw6JfO//+LRfCNPAaJffSLVeiLRfSJQhj2QggIdSCDfewAfwmLRej2QAgQdBGLRfSL fQiNcAEB+Il19MYALotF6ItQBIXSfSWLRQiLTfSJ04lV8LowAAAA99sByIld8Og5zv//i0Xw jTQBiXX0i33shf9+STt94H0DiX3gi13ghdt0Jot9CItN9It15AHPidkGjNiOwFeJyMHpAvKl isiA4QPypF8HAV30i1Xoi0X0i33giUIci0XsKfiJQiCLReiDeAwAdBaLRfSLTQiNWAGLVegB yIpSDIld9IgQi0X0i3XcQIX2fA+LVfSJRfSLRQjGBAIr6xSLVfSJ94lF9Pffi0UIiX3cxgQC LYtd6ItbEIP7A3c+idgu/ySFLWJAAIF93OgDAAB9JrsDAAAA6ySDfdwKfAW7AgAAAIN93GR8 BbsDAAAAgX3c6AMAAHwFuwQAAACLReiJWBCD+wR8SzH/i0XciX3wPegDAAB8JonCv+gDAADB +h/3/4lF8InCweAFKdDB4AIB0ItV3MHgAynCiVXci0X0i3UIjUgBilXwAfCAwjCJTfSIEIP7 A3xJMf+LRdyJffCD+GR8JonCv2QAAADB+h/3/4lF8InCweACKdDB4AMB0ItV3MHgAinCiVXc i0X0i3UIjUgBilXwAfCAwjCJTfSIEIP7AnxDMf+LRdyJffCD+Ap8IInCuwoAAADB+h/3+4lF 8InCweACAdCLVdwBwCnCiVXci0X0i00IjVgBilXwAciAwjCJXfSIEItF9It9CI1wAYpV3AH4 gMIwiBCLReiJ8otAHCnCi0XoiVAkjQQ3iXX0xgAA6ZH8//9VieVRVleD7DSJ14ld6IlF4IoQ gPogdAqA+glyCID6DXcDQOvsjVgBMMmA+it0B4D6LXUEsQGJ2DHbtjCJXeSKEECA+i51CvbB CHUtgMkI6+6A+jByI4D6OXce9sEIdAP/ReQI1oD+MHQKg/sTfQSIVCvAQ4DJBOvGMfb2wQQP hGUAAACA+mV0BYD6RXVXjVD/iiiJVfCNUAGA/St0CID9LXUFgMkCidCA4fuKEID6MHIigPo5 dx2B/ugDAAB9D2v2Col17A+28gN17IPuMIDJBEDr1/bBAnQC9972wQR1BotF8OsBSIlF4IN9 6AB0CItV6ItF4IkCK3Xkg/sTfgqD6xMB3rsTAAAAhdt+C4B8K78wdQRGS+vxhdt1EWbHRwgA AIlfBIn4iR8x+OtfMNKNRcCIVCvAjVXU6EsVAACF9nQKjUXUifLopPT///bBAXQEgE3dgItF 3GaJRwiLRdiJRwSLRdSJB41EHv89NAEAAH4HuAMAAADrEz3M/v//fQe4AgAAAOsFuAEAAACN ZfRfXlldw1WJ5VNRg+wkidONVdToWv7//4nDhcB1C4lF6IlF7OmtAAAAi0XcgOR/Jf//AAA9 /0MAAHwq6FPb///2Rd2AdBDdBQDKQADZ4N1d6OmAAAAAoQDKQACJReihBMpAAOu8Pc07AAB9 QoP4zH0PMcnoG9v//4lN6IlN7OtWjVXojUXU2yjdGotF6IlF8ItF7Itd6IlF9Kn///9/dQSF 23QtZvdF9vB/dSrrI41V6I1F1Nso3RqD+wN0FItF6IlF4ItF7IlF5Gb3Rebwf3UF6L3a///d ReiNZfhZW13DAAAAAAAAAAAAAAAAAAAAUVZXVYPsDIlEJASJ1YkcJDHbhcAPhHEAAACDPCQA D4aJAAAAZotNAGaFyXQgMdKNRCQIZonK6IXq//+JwoP4/w+EaQAAADsEJHdi6wmLRCQExgAA 61eNdCQIi3wkBInBBozYjsBXicjB6QLypYrIgOED8qRfB4PFAgHTiwQkAdcp0Il8JASJBCTr j2aDfQAAdBsx0o1EJAhmi1UA6B/q//+D+P90CYPFAgHD696J2IPEDF1fXlnDAAAAAAAAAAAA U1FSVldVicWLNfTxyAGF9nQ/hcB0O4nHBozYjsApyUkxwPKu99FJB4nP6x+J+4nqicjouRMA AIXAdQ2APDk9dQeNRwEByOsLg8YEiw6FyXXbMcBdX15aWVvDAAAAAAAAUVZXVTHthcAPhJoA AACNSAs5wQ+CjwAAAInfgOH4i0MUg/kQcwW5EAAAADnBD4d1AAAAi18Qi0cMOdl3BYtHKDHb jXcgixA50XYSOdp2AonTi0AIOfB17YlfFOtLiV8Qi18YQynKiV8Yg/oQch6NHAiJXwyJE4kI i0gEiUsEi1AIiVMIiVkIiVoE6xL/TxyLUASJVwyLWAiJWgiJUwSKEIDKAY1oBIgQiehdX15Z w5BRVldVidmFwA+EHQEAAI1Y/PYDAQ+EEQEAAIsTgOL+jQQTic72AAF1Ios4AfqJEztBDHUD iVkMi1AEi0AIiUIIiVAE/04c6ZQAAACJE4tBDDnDcxY7WAQPh4IAAACLQSg5ww+CdwAAAOsZ i0AIOcMPgmoAAACNQSCLUAQ50w+HXAAAAIt+HItGGDHSjU8B9/GJwTn4cyiLRhgp+AHJOfh3 Bbn/////iwMB2IsQ9sIBdC6D+v90CIDi/gHQSXXsi0YMOcNzA4tGKDnDchOLQAg5w3IMi0AI OcNyBYtACOvpi1AEizoB14sLOd91EosqAemJCjteDHUDiVYMidPrD/9GHIlDCIlTBIlaCIlY BItWGEqLfgyJVhg5+3MIO04QdgOJThA7ThR2A4lOFF1fXlnDU1KJwqHszkAAMduFwHQNOcJy CYnDi0AIhcB184laBIlCCIXbdAWJUwjrBokV7M5AAIXAdAOJUASNWiCDwizHQvQAAAAAx0Lk AAAAAMdC7AAAAADHQvAAAAAAiVr4i0LUiVr8g+gsiVrgiQLHBAL/////idBaW8OQU1FSVleD 7ASJBCSDPaDTQAAAdQcxwOmYAAAAgz3AzEAA/nULMcCDxARfXlpZW8OJ4OiaAAAAhcAPhHUA AABqQGgAEAAAi0wkCFFqAC7/FQCyQACJwoXAdFqLBCSLNCSD6AQ58HYLMcCDxARfXlpZW8OJ BCSD+DhzCzHAg8QEX15aWVvDiQKJ0Oj7/v//icOLAIkEJAwBiQOLehjHQhQAAAAAR41DBIl6 GOhK5f//uAEAAACDxARfXlpZW8NSicLo9AQAAInQ6DX///9aw41AAFFSicKLAIPAByT4hcB0 PYkCg8A8OwJzBTHAWlnDiw2k00AAOchzBInIJP6JAgX/DwAAOwJzBTHAWlnDMMCA5PCJAoXA D5XAJf8AAABaWcMAAAAAAAAAAAAxwMMAAAAAAAAAAAAAAAAAgD1Q0UAAAHQageL//wAAUlDM 6wZXVklERU+4AQAAAIPECMMxwMMAAAAAAAAAAAAAUVIu/xV0sUAAWlnDAAAAAFNRUlZVgewo AgAA/xUkzUAAvXDJQACLcAQxyY2EJBQBAACJyujv0///ugIAAACNhCQUAQAAQeg1EAAAhcB0 3I2EJBQBAACJ6ujLEgAAhcB1G+jaEwAAgzgLdAro0BMAAIM4BnW2McDpmAAAADHb6D3T//+K HbTMQACJ4Ina6JbT//+J4o2EJBQBAADowBMAAIXAdVGJ4Inq6HsSAACJwoXAdCqKYA2AzAiI Yg2LQgiIWBSJ8IgdtMxAAOjR1P//idCBxCgCAABdXlpZW8PoXhMAAIM4C3UOMcCBxCgCAABd XlpZW8O6AgAAAI2EJBQBAABD6HkPAACFwA+FHP///+l1////gcQoAgAAXV5aWVvDU1FSVleA PWjRQAAAD4WrAAAAu1TRQADrOeii+v//icKFwHQriccGjNiOwCnJSTHA8q730UkHgfkDAQAA dxG7AwEAALho0UAA6BMTAADrCoPDBIsDgDgAdcCAPWjRQAAAdSox0jHAv2jRQADooBMAAInG V4oGiAc8AHQQikYBg8YCiEcBg8cCPAB16F+/aNFAAAaM2I7AKclJMcDyrvfRSQeNQf8FaNFA AIoYgPtcdA2A+y90CEDGAFxAxgAAuGjRQABfXlpZW8MAAAAAAAAAAFNRUlaJwvZADIB0DfZC DRB0B4nQ6McTAACLQhDozxMAAInDicGD+P90KotCEP8VKM1AAItyBIX2dA/2Qg0QdAWNDB7r BInZKfGLQhD/FSzNQACJyF5aWVvDAAAAAAAAAFNRUlZXVYnDhcB8CDsFbNJAAHYRuAQAAADo MtP//7j/////62GLFfjOQACLLVzTQAAxyTH2izyahe10Hv8VUNNAAInChcB0EonY/xVU00AA idD/FVzTQADrIYXJdR1XLv8VNLFAAIXAdRG4BAAAAL7/////6NfS///rCYnYMdLodAAAAInw XV9eWllbwwAAAOlbEwAAAAAAAAAAAAAAAABTUVKJwjsFbNJAAHIGMcBaWVvDg/gDfTONHIUA AAAAocDSQAAB2IpIAfbBQHUdiM2AzUCIaAGJ0OgDCAAAhcB0CqHA0kAAgEwDASChwNJAAIsE kFpZW8OQU8HgAoXSdA6LHcDSQACAzkCJFANbw4sdwNJAAIkUA1vDAAAAAAAAAFNRVldVicOJ 0YXAdQmJ0Oi8zf//62KF0nUN6AHh//8xwF1fXllbw+jEEgAAicaJ2OjLEgAAicWFwHU7icjo js3//4nFhcB0JYnxiceJ3gaM2I7AV4nIwekC8qWKyIDhA/KkXweJ2Oi14P//6wmJ8onY6IoS AACJ6F1fXllbwwAAUw+vwonD6EXN//+FwHQHMdLoysD//1vDAAAAAAAAAABTUVKJw4XAdBsx 0sdADAEAAADoSdv//y7/FXixQACJg9oAAABaWVvDAAAAAAAAAADpk+f//wAAAAAAAAAAAAAA U1FSVv8VQM1AAKHszkAAhcB0HIsYi0gkg+ssizGLUAg583UF6HMAAACJ0IXSdeT/FUjNQAAx wF5aWVvDU1FSVleJw2gAgAAAagBQi3AILv8VBLJAAIXAdQe4/////+s2Ox3wzkAAdRyF9nQI iTXwzkAA6xCh7M5AAIk19M5AAKPwzkAAOx0Q8sgBdQgx/4k9EPLIATHAX15aWVvDU1KLWASL UAjok////4XAdRaF23UIiRXszkAA6wOJUwiF0nQDiVoEWlvDAAAAAAAAAAAAAOkbzP//jUAA U1FSVldViz308cgBhf8PhcEAAACLLQHNQAAxyYpVAInohNJ0FoowjVgBhPZ0BInY6/NBijOJ 2IT2deqJPfTxyAEp6HUFuAEAAADoycv//4nCicOFwA+EdQAAAKNw88gBjQSNAAAAAIPABAHI 6KfL//+FwHRUiceKVQAx9jHJieiE0nQaifqJHBFDihBAiFP/hNJ19YPBBIoQRoTSdeaJ+McE AQAAAACDwQSJ840EDzHSo/DxyAGJPfTxyAHo577//4s99PHIAesHidDomN7//+hzEgAAiz30 8cgBXV9eWllbw4vAU1FSVlfoehMAAIsV9PHIAYXSdA+J0DHb6Gfe//+JHfTxyAGLDXDzyAGF yXQPicgx9uhO3v//iTVw88gBiz0BzUAAhf90CFcu/xVcsUAAX15aWVvDUVKJwmaDOAB0DGaL SAKDwAJmhcl19CnQ0fhaWcMAAABRVleJ2YnWiccGjNiOwFeJyMHpAvKlisiA4QPypF8Hifhf XlnDAAAAAAAAAAAAAABmgzgBdRyDeAQAdBZmi0AKZj0QAHIGZj0SAHYGuAEAAADDMcDDjUAA UVIxwP8VKM1AAIM9xNJAAP91I2oAaIAAAABqA2oAagFoAAAAgGh0yUAALv8VPLFAAKPE0kAA gz3I0kAA/3UjagBogAAAAGoDagBqAmgAAABAaHzJQAAu/xU8sUAAo8jSQAAxwP8VLM1AAFpZ w41AAOiL////ocTSQADDkOh/////ocjSQADDANviwwAAAAAAAAAAAAAAAABTUYnDg/gHdAWD +AR1GI0E3QAAAACLmMzSQACJkMzSQACJ2Flbw40M3QAAAAD/FSTNQACLXAFY/xUkzUAAiVQB WInYWVvDi8BSicKD+Ad0BYP4BHUJiwTVzNJAAFrD/xUkzUAAi0TQWFrDi8BSicKD+Ad0BYP4 BHUJiwTV0NJAAFrD/xUkzUAAi0TQXFrDi8BTicPo1P///znCdQmJ2Oil////W8MxwFvDjUAA i0QkBIXAdgqD+AF0JDHAwgQAuAQAAADogP///4nChcB1A8IEALgEAAAA6N0BAADrHbgHAAAA 6GH///+JwoXAdQPCBAC4BwAAAOi+AQAAg/oCdAWD+gN1BTHAwgQAuAEAAADCBACNQABSuAQA AADoKf///4nCuAcAAADoHf///4P6AnQFg/oDdQqD+AJ0DIP4A3QHuAEAAABawzHAWsOLwFFS gD0000AAAHUZagFoDHVAAC7/FcyxQACFwHQHxgU000AAATHAoDTTQABaWcOLwFFSgD0000AA AHQaagBoDHVAAC7/FcyxQACFwHQIMNKIFTTTQAAxwKA000AAhcAPlMAl/wAAAFpZw41AALgB AAAA6fYAAACLwFNRUonDuAIAAADodf7//4nBg/gBdCWD+AJ0IIP4A3QbugIAAACJ0OgQ/v// uAIAAACJ2v/RMcBaWVvDuP////9aWVvDU1FWicOJ0YP4AXwFg/gMfhO4CQAAAOgNzP//uAMA AABeWVvDxwWo00AAGHZAAIP5AnQfg/kDdBroJ/7//4XAdBGD+wJ1DLqfAAAAMcDohhAAAInY 6Of9//+JyonGidjolP3//+ij/v//hcB0B+jS/v//6wXo+/7//4nwXllbw5BAd0AANHdAAEp3 QABKd0AASndAAEp3QABKd0AASndAAEp3QABKd0AASndAAEp3QABTUVKJw+iG/f//jVP/icGD +gt3TS7/JJXodkAAuIwAAADo5v7//+tCg/gCdQXo3g8AAIP5AXQag/kCdBWD+QN0ELoCAAAA idjo+/z//4nY/9HoBv7//4XAdRDoZf7//+sJuP////9aWVvDMcBaWVvDU1JWVwa6CAAAAP8V JM1AAI00EIzYjsCNfliNsszSQACDwgilpYP6aHXguvB0QAC7GHdAAIkVFNBAAIkdGNBAAAdf Xlpbw4vAUuie/f//hcB0I+j9/f//ugIAAAC4BAAAAOhy/P//ugIAAAC4BwAAAOhj/P//WsOQ U1K6iHdAALvQd0AAiRVkzUAAiR1ozUAAWlvDAAAAAABTUVKJw/8VKM1AAIM9UNNAAAB0HYnY /xVQ00AAhcB0EYnY/xUszUAAuAEAAABaWVvDofjOQACLBJhQLv8ViLFAAIP4AnURidj/FSzN QAC4AQAAAFpZW8OJ2P8VLM1AADHAWllbwwAAAAAAAAAAAAAAU1FSVlWD7BSJxoP4/3UOLv8V YLFAAInG6ZgAAACD+P51Di7/FaSxQACJxumFAAAAg/j9dSa7AQEAALgw8sgBMdLo6rj//zHS McCJFSDyyAGJFTjTQADp/wAAAIP4/HVVuwEBAAC4MPLIATHS6L+4//+4gQAAALIBQIiQMPLI AT2fAAAAfvK44AAAALYBQIiwMPLIAT38AAAAfvK9AQAAALikAwAAiS0g8sgBozjTQADpowAA AIX2dQW+AQAAAIngUFYu/xVksUAAhcB1DrgBAAAAg8QUXV5aWVvDuwEBAAC4MPLIATHS6ES4 //8xyYpkJAaJDSDyyAGE5HQKxwUg8sgBAQAAADHbsQHrHDHAikQcBusHQIiIMPLIATHSilQc BznQfu+DwwKAfBwGAHXdgHwcBwB11oP+AXUXLv8VpLFAAKM400AAMcCDxBRdXlpZW8OJNTjT QAAxwIPEFF1eWllbwwAJ0ngeCcl4BuhIAAAAw/fZ99uD2QDoOwAAAPfa99iD2gDD99r32IPa AAnJeRT32ffbg9kA6BwAAAD32ffbg9kAw+gPAAAA99n324PZAPfa99iD2gDDCcl1Gkt0FkM5 03cJicGJ0CnS9/OR9/OJ04nKKcnDOdFyHHUSOcN3DinYicMpySnSuAEAAADDKckp25OHysNV Vlcp9on3ifUB2xHJchNFOdFy9XcEOcN27/gR9hH/TXgi0dnR2ynYGcr1cu4B9hH/TXgM0enR 2wHYEcpz7+vbAdgRyonDidGJ8In6X15dwwAAAAAAAAAAAIXSdQQxwMOQhdJ0+lKJwujkx/// g/p7dQe4AQAAAOtFgfrOAAAAdRG4CQAAAOh+x///uP////9aw4H6twAAAHURuAcAAADoZcf/ /7j/////WsOD+hN2BboTAAAAi4I500AAwfgY6EbH//+4/////1rDjUAAUVIu/xWQsUAA6Ir/ //9aWcMAAAAAAAAAAAAAADHAw5D/FZTTQADDkP8VmNNAAMOQ/xWc00AAw5BRVoXAdBCLCIs1 lNNAAIkNlNNAAIkwhdJ0D4sCiw2Y00AAo5jTQACJCoXbdA+LA4sVnNNAAKOc00AAiRNeWcNS ZotQCIDmf2aB+v9/dRyBeAQAAACAdQyDOAB1B7gDAAAAWsO4AgAAAFrDZvdACP9/dRaDeAQA dQmDOAB1BDHAWsO4BAAAAFrDuAEAAABawwBVVldRU1KJxinSKckp7SnAigY8AHQvideJy4no Ae0RyRHSAe0RyRHSAcUR2RH6Ae0RyRHSKcCKBiQPAcWD0QCD0gBG68uJyL9eQAAA6BEAAABd iVUEiUUAZol1CFtZX15dwyn2CcYJ1gnudDYJ0nUJicKJ6Cntg+8gCdJ1CYnCiegp7YPvIAnS eAlPAe0RwBHS6/MB7YPQAIPSAHMD0dpHif7DAFFWicGJ1jHShdt2PInI6G0MAACFwHUxifDo YgwAAIXAdSaJ8onI6JUMAACJwoXAdT2JyOjoDAAAicGJ8Evo3gwAAInGhdt3xIXbdiCJyOgt DAAAhcB1C4nw6CIMAACFwHQKMdIxwIoRigYpwonQXlnDAAAAAAAAAAAAAAAAAAAAU1KJw/8V QM1AAIsVoNNAAIkdoNNAAP8VSM1AAInQWlvDAAAAAAAAAAAAAAAAAAAAU1GJ01Au/xWEsUAA g/j/dQjozv3//1lbw/bDAnQRqAF0DbgFAAAA6FD9//9ZW8MxwFlbwwAAAAAAAAAAAAAAAFNR VldVg+wEicGJ1roBAAAAMdsx/zHtiRwkhfZ0DjsVrNNAAHUEiRbrAokeigE8cnIIdgw8d3QN 6xU8YXQM6w+AywHrG4DLAusWgMuC6xG4CQAAAOiHxP//MdLpzwAAAIphAUGE5A+ErgAAAIXS D4SmAAAAigE8Y3IdD4ZlAAAAPG4PgoYAAAAPhmwAAAA8dHQy6XcAAAA8Kw+CbwAAAHYJPGJ0 LulkAAAAgzwkAHQHMdLpVwAAALgBAAAAgMsDiQQk60qF/3QEMdLrQr8BAAAA6zuF/3QEMdLr M78BAAAAgMtA6ymF7XQEMdLrIYoGDAG9AQAAAIgG6xSF7XQEMdLrDIomgOT+vQEAAACIJopB AUGEwA+FUv///4X/dQ+BPbjMQAAAAgAAdQOAy0CJ2onQg8QEXV9eWVvDi8BWV1WJxojQicqL TCQUgGEM/AlZDCX/AAAA6BYLAAA8cnUiMcD2wwJ0BbgCAAAA9sNAdAWAzALrA4DMAWoAi3wk FFfrMPbDAQ+VwCX/AAAAg8Ah9sOAdAQMEOsCDED2w0B0BYDMAusDgMwBaIABAACLbCQUVVBW 6O0KAACDxBCJQRCDeRD/dQuJyOjKwP//McDrTsdBBAAAAACLQQjHQRQAAAAAx0AMAAAAAItB CIlQEItBCMdAFQAAAACLQQjHQAgAAAAA9sOAdA67AgAAAInIMdLoRQ0AAInI6MbS//+JyF1f XsIIAIvAUVZXg+wEicaJ0YnfieKJyOjH/f//icOFwHQiMcDoir///4nChcB0E1AxwleJ8IoR i0wkCOjf/v//icKJ0IPEBF9eWcNTMdvotP///1vDi8BTUVKJw/8VOM1AAKFA5MgBhcB0K4tQ BDnTdR72QgwDdAy6AQAAAInY6HPB////FTzNQACJ2FpZW8OLAIXAddW6ROTIAYsChcB0JjtY BHUdiwiJCosVQOTIAaNA5MgBiRD/FTzNQACJ2FpZW8OJwuvUuAQAAADo8sH///8VPM1AADHA Wllbw4vAUVZXVYPsBInHidGJ4onI6PD8//+JxYXAdFWLcxCJ8P8VKM1AAIM9VNNAAAB0CInw /xVU00AAidjoO////4nDhcB0IlCLWAwx0oHjAEAAAGoAiVgMieuJ+IoRi0wkCOjd/f//icOJ 8P8VLM1AAInYg8QEXV9eWcMAAAAAAAAAAAAAAAAA/xUkzUAAg8AEw4vA/xUkzUAAg8AIwwAA AAAAAAAAAABRUlAu/xXEsUAAhcB1B+jx+f//WcMxwFnDAAAAAAAAAFFWV4PsBInGidGFwHUl uwQBAACJ2OimvP//iceFwHURuAUAAADo9sD//zHA6XQAAACJxoXJdAWAOQB1EInaifDoawAA AIPEBF9eWcO6hMlAAInI6LgMAACFwHUog/sEcxqJ+Oioz///uA4AAADorsD//zHAg8QEX15Z w6GEyUAAiQbrH4ngUFZTUS7/FYyxQACFwHUO6Er5//8xwIPEBF9eWcOJ8IPEBF9eWcMAAAAA U1FWV4HsBAEAAInDidaJ4FBoBAEAAC7/FXCxQACJwYXAdQnoDPn//zHA62WF23UpQDnGdgKJ 8OjUu///icOFwHUyuAUAAADoJMD//zHAgcQEAQAAX15ZW8M58HYXuA4AAADoCcD//zHAgcQE AQAAX15ZW8OJ5kGJ3waM2I7AV4nIwekC8qWKyIDhA/KkXweJ2IHEBAEAAF9eWVvDAAAAAIXA D4X4rP//6PMLAAAxwMNTUVKJw4XAfAg7BWzSQAB2E7gEAAAA6KW///+4/////1pZW8OLFfjO QACLFJr/FSjNQABqAWoAagBSLv8V3LFAAInCidj/FSzNQACD+v91Begv+P//idBaWVvDAAAA AABRUlAu/xVIsUAAhcB1COgR+P//WlnDMcBaWcMAAAAAAIPoBIsAJP6D6ATDAAAAAABTUVaD 7ASJxonT/xVAzUAAuOzOQACM2jHAieFmidCJ8ugXAAAAhcAPhbsBAAD/FUjNQACJ8IPEBF5Z W8NWV1WD7BCJRCQMideJzY1DCyT4OdhzB7j/////6wqD+BBzBbgQAAAAjVf8iRQkixKA4v45 0A+GCwEAAIsMJCnQAdGJRQCLGYP7/3UKuwIAAADpTQEAAPbDAQ+F2QAAAItBCIlEJAiLQQSM 2olEJAS47M5AAGY7VCQMdSGLNezOQACDfggAdBU5/ncIiwYB8Dn4dwmLdgiDfggAdeuLVgw5 0XUGi0IEiUYMO10AcgyJ2ItVACnQg/gQczSLRCQEi1QkCIlQCInQi1QkBIlQBIsEJAEYi0Yc SDDSiUYciBVh88gBi1UAOdNyROmzAAAAAdGJAYtEJASJQQSLRCQIiUEIi0QkBIlICItEJAiJ SASLFCSLRQCLCjHbAcEw5IkKiCVh88gBidiDxBBdX17DidAp2AHZ6Qr///+7AQAAAInYg8QQ XV9ewynCg/oQclSJw4sMJIDLAYDKAYkZjRwBuOzOQACJHCSJE4zai1wkDGY52nUhizXszkAA g34IAHQVOf53CIsGAfA5+HcJi3YIg34IAHXr/0YYiwQkg8AE6EPM//8x24nYg8QQXV9ew/8V SM1AADHAg8QEXllbwwAAAAAAAAAAAFNRUlZXVYPsELiIyUAA6F3j//+JwYXAD4T5AAAAgDkA D4TmAAAAsjqJzooGOsJ0EjwAdAxGigY6wnQHRjwAdeor9ongifUpzonKifPoXwkAALsQAAAA MNKJ4IgUNDHS6NwKAACNdQGyOonHifGKBjrCdBI8AHQMRooGOsJ0B0Y8AHXqK/aJ4In1Kc6J yonz6BwJAAC7EAAAADD2ieCINDQx0uiZCgAAjXUBsiqJRCQMifGKBjrCdBI8AHQMRooGOsJ0 B0Y8AHXqK/aJ4In1Kc6Jyonz6NcIAACJ4DDbMdKIHDS7EAAAAOhUCgAAicGJ+otEJAzob73/ /4nKifiNdQHo0+n//4nx6RH///+4lMlAAOiaCgAAg8QQXV9eWllbw1NRUlaLFfTxyAGF0g+E gAAAAOswgz3w8cgBAHQkizX08cgBidEp8Ysd8PHIAcH5AoA8GQB0BejFyv//xwIAAAAAg8IE iwKFwHXKgz3w8cgBAHUMuAUAAADoU7f//+sPugUAAACh9PHIAehy6f//hcB1Crj/////XlpZ W8Oj9PHIAccAAAAAAIPABKPw8cgBxgAAMcBeWllbwwAAAAAAAAD/FajTQACLwFK6/wAAALik yUAA6CS4//9awwAAAAAAAFNRVleD7ASJwzHJiiWoykAAiQwkhOR0JonnNtk/m4XSdByJ0Is0 JPfQIdoh8AnQieeJBCQ22S+biec22T+bMcBmiwQkg8QEX15ZW8MAAAAAU1ZXBg+gD6hVieWD 7AyLdSCLFVDkyAGLBot+BIlF/IXSdTGhcM1AAIPAAyT8KcSJ4YsdcM1AAInI6PKp//+hcM1A AImB8AAAAInI6OjA//+FwHQx/xUkzUAAi14MBd4AAABTiUYQLv8V2LFAAI1F9OgLyf///xVk zUAAifj/VfzoZ/P//4nsXQ+pD6EHX15bw41AAFFWVwZVg+w4icGJ1YM9IM1AAP91EugqwP// hcAPhL4AAADoWcH//4nniUwkIIlcJCQu/xV8sUAAiUQkKIzYMduOwL68yUAAiVwkMLsQAAAA paWkLv8VeLFAAI1UJAjo1cr//4ngUGoAagBqAC7/FTixQACJRCQsjUQkNFBqAI1EJChQgcX/ DwAAaJCHQACB5QDw//9VagAu/xVAsUAAicOJxYXAdQrHRCQ0/////+siav+LTCQwUS7/FQyy QACLdCQwhfZ0BIke6whTLv8VNLFAAItEJCxQLv8VNLFAAInog8Q4XQdfXlnDjUAAUVL/FWjN QADoP8j//4M9UOTIAQB1CrgBAAAA6Oy///9qAC7/FVSxQABaWcMAAAAAUoA4AHUHuAEAAABa w4M9IPLIAQB0IjHSihCKkjHyyAGA4gGB4v8AAAB0DYB4AQB1B7gCAAAAWsMxwFrDAAAAAFNR g+wIicOJ0YnCieDo7gsAAInYMNLoJQwAAIgUBI1EJASJyujXCwAAicgw9ugODAAAiHQEBIng 6DMMAACNRCQEjVQkBOgmDAAAieDobwwAAIPECFlbwwAAAAAAAAAAAFKDPSDyyAEAdCAx0ooQ ipIx8sgBgOIBgeL/AAAAdAuAeAEAdAWDwAJaw0BawwAAAMcFrNNAAAEAAADDAAAAAACD+EF8 CIP4Wn8Dg8AgwwAAU1KNRCQUg8AEi0D8UGoAi1QkGFKLXCQYU+gGAAAAg8QQWlvDU1FSg+wE jUQkIInhi1wkHItUJBiJBCSLRCQU6AgAAACDxARaWVvDkFZXVYPsIInHiVQkHInd6GG4//+F wHQUuAsAAADo87f//7j/////6f4BAACJ1o1cJBSD5geNVCQQifDoFQwAAI1UJAyJ6DHbCfC9 gAAAAOg1DAAAugwAAACKZCQciVwkBIkUJPbEgA+UwCX/AAAAiUQkCIM9WNNAAAB0NLrIyUAA ifjofQMAAIXAdSToJLr///8VMM1AAGr/UFO6ACAAAInB/xVY00AAg8QM6SoBAACKVCQc9sIg D4RyAAAAixGDwgSJEYtC/IlEJBTHAQAAAAChPNRAAItcJBT30CHDiVwkFPZEJBUBdAz2RCQU gHUFvQEAAAD2RCQdBHQNugEAAACJVCQYidDrN/ZEJBxAdA/HRCQYAgAAALgBAAAA6yG5BAAA ALgDAAAAiUwkGOsR9sJAdAe4BQAAAOsFuAMAAABqAFVQjUQkDFCLXCQcU4tMJCRRVy7/FTyx QACJw4P4/3U29kQkHCB0HmoAVYtsJCBVagCLRCQcUItUJCRSVy7/FTyxQACJw4P7/3UM6Ejv //+DxCBdX17Didj/FTDNQAAx0os9bNJAAInBOfhyHlMu/xU0sUAAuAUAAADoVbb//7j///// g8QgXV9ew+jU6///hcB0BboAIAAAg/4CdQWAygPrEYX2dQWAygHrCIP+AXUDgMoC9kQkHBB0 A4DKgInQilwkHQxA9sMDdAf2wwJ0EOsMgT24zEAAAAIAAHUCicKJyOiV4///iciDxCBdX17D AAAAAFNRVlc7QgR/LItaCIsKi1sIKcs52HweiloMizKLegSA4+8BxohaDCnHiTIxwIl6BF9e WVvDuAEAAABfXllbw5BSgGAM74tQCItSCMdABAAAAACJEFrDi8BRVleJwYnWi0AQ/xUozUAA imEM9sQGD4R1AAAA9kENEHQvicjoj6L//4XAdDuF23UOhfZ9CrgJAAAA6Em1//+LQRD/FSzN QAC4/////19eWcOD+wF1AytxBItBCItACMdBBAAAAACJAYBhDOuJ8otBEOiCyP//g/j/D4Xk AAAAi0EQ/xUszUAAuP////9fXlnDg/sBcg12EYP7Ag+EigAAAOuShdt0ReuMicqJ8It5BOjz /v//hcAPhKYAAACJ8otBECn66C/I//+D+P91EotBEP8VLM1AALj/////X15Zw4nI6AH////p dwAAAItBEOjU9P//i1EEifcp0CnHicqJ+Oii/v//hcB0WYnyi0EQ6OTH//+D+P91x4tBEP8V LM1AALj/////X15Zw4jigOLvi0EIiFEMi0AIiQGJ8otBEMdBBAAAAADorMf//4P4/3USi0EQ /xUszUAAuP////9fXlnDi0EQ/xUszUAAMcBfXlnDAAAAAAAAU4nDigOKIjxBcgY8WncCBCCA /EFyCID8WncDgMQgOOB1CITkdARDQuvZMdKIwojgJf8AAAApwonQW8MAAAAAALj/////jUAA U1FSicH/FTjNQACLFUDkyAEx24XSdBqLQgSFSAx0DEP2QA0QdAXozaD//4sShdJ15v8VPM1A AInYWllbwwAAAAAAAAAAAAAAUVaJxoXbdA+AOgB0CkCKCkKISP9LdfEx0oXbdAdAS4hQ/+v1 ifBeWcMAAAAAAAAAVldVg+wQiUQkDIlUJASJ3olMJAiF0nQCiQKLVCQMigL+wCX/AAAA9oAQ z0AAAnQDQuvriiqA/St0BYD9LXUBQoX2dSyAOjB1FIpKAYD5eHQFgPlYdQe+EAAAAOszgDow dQe+CAAAAOs8vgoAAADrNYP+AnwFg/4kfhG4DQAAAOjMsv//McDpwQAAAIP+EHUVgDowdRCK egGA/3h0BYD/WHUDg8ICiRQkjSy1AAAAADDJMdsxwIoC6LQAAACJxznwfRo7najTQAB2ArEB idgPr94B+znDcwKxAULr1zsUJHUEi1QkDIt8JASF/3QCiReDfCQIAXUPgfsAAACAcgd1CYD9 LXUEhMl0OrgOAAAA6D2y//+DfCQIAHUMuP////+DxBBdX17DgP0tdQy4AAAAgIPEEF1fXsO4 ////f4PEEF1fXsOA/S11AvfbidiDxBBdX17DkFExyeik/v//WcOLwFG5AQAAAOiV/v//WcON QABSPDByDjw5dwol/wAAAIPoMFrDJf8AAADoU/n//4nCPGFyDTxpdwkxwIjQg+hXWsM8anIO PHJ3CiX/AAAAg+hXWsM8c3IOPHp3CiX/AAAAg+hXWsO4JQAAAFrDAAAAAABTUVJWV1WD7BiJ xboCAAAAuwEAAACJFCS6PQAAAIlcJAToKgYAAInDhcB1Crj/////6XwBAAA56HUPuP////+D xBhdX15aWVvDKeiJRCQQQOjLrP//icKJRCQIhcB1D7j/////g8QYXV9eWllbw4tMJBCJ7onH BozYjsBXicjB6QLypYrIgOED8qRfB4tEJBCNcwHGBAIAifDoEgYAAIlEJBSFwHRLQOh0rP// icOJRCQMhcB1FonQ6LO///+4/////4PEGF1fXlpZW8OLTCQUiccGjNiOwFeJyMHpAvKlisiA 4QPypF8HA1wkFMYDAOsEiUQkDItEJAxQi1QkDFIu/xXQsUAAicKLRCQI6F+///+LRCQM6Fa/ //+F0nUPuP////+DxBhdX15aWVvDiejoiAAAAIXAdA+4/////4PEGF1fXlpZW8ODPfjxyAEA dGKJ6OhZBQAAjVgBiwQkD6/D6Lur//+JwYXAdRm4BQAAAOhn6P//uP////+DxBhdX15aWVvD D69cJASJ6uhQBQAAg/j/dRaJyOjUvv//uP////+DxBhdX15aWVvDicjoRgcAAIPEGF1fXlpZ W8NTUVJWV1WD7AyJRCQEhcB1Crj/////6YkBAACAOAB0FY1QAYA6AHQNgDo9dAiKSgFChMl1 84A6AHUPuP////+DxAxdX15aWVvDgHoBAA+UwA+26KH08cgBiUQkCIXAdUGF7Q+FPAEAALgJ AAAA6PKq//+JRCQIhcB1D7j/////g8QMXV9eWllbw6P08cgBg8AIiWj4iWj8o/DxyAHp5QAA AItEJASJ6ugEAQAAhe0Phe4AAACFwA+PxwAAAInF992NXQGNDK0AAAAAjUEIizXw8cgBiQQk AdiF9nVP6IOq//+JwolEJAiFwHUPuP////+DxAxdX15aWVvDizX08cgBiccGjNiOwFeJyMHp AvKlisiA4QPypF8HiwQkAdAx0qPw8cgB6Mud///rPInCi0QkCOhe3P//iUQkCIXAdQ+4//// /4PEDF1fXlpZW8OLDCSLFfDxyAEBwYnricjocwgAAIkN8PHIAYt8JAiNBK0AAAAAAfiJPfTx yAHHQAQAAAAA6wONaP+LVCQIjQStAAAAAAHQi1QkBIkQofDxyAHGBCgAMcCDxAxdX15aWVvD i8BTUVZXVYnHidWLNfTxyAHpywAAAIonifuE5A+EvAAAADHAigLoW8D//4nBMcCKA+hQwP// OcEPhaAAAACAOj0PhYoAAACLFfTxyAGJ9ynXwf8Che0PhGwAAACLFonxhdJ0D4tBBIkBi1kE g8EEhdt18Ys18PHIAYX2dEOJ8IA8BwB0B4nQ6Hq8//+LLfTxyAGJzinuixXw8cgBwf4CiciJ 8+h+BwAAiQ3w8cgBOfd9Do0ED0CKEEeIUP8593z1McBdX15ZW8ONRwFdX15ZW8NCiksBQ4TJ D4VE////g8YEixaF0g+FK////6H08cgBKfDB+AJdX15ZW8MAAAAAAAAAAAAAAFODPSDyyAEA dCEx24oaipsx8sgBgOMBgeP/AAAAdAyKGogYilIBiFABW8OKEogQW8MAAAAAAAAAAAAAAAAA AACDPSDyyAEAdByKACX/AAAAioAx8sgBJAEl/wAAAHQGuAIAAADDuAEAAADDAAAAAABTUVKD 7ASJwYnDidjoP/P//4XAdTGJ2Ini6OIGAADoHQcAAOiIBwAAieAw0uif////iBQEieKJ2OhT ////idjorPP//+vCiciDxARaWVvDAFNRihg6GnVBgz0g8sgBAHQfMduKGIqbMfLIAYDjAYHj /wAAAHQKilgBikoBOMt1BTHAWVvDiNgl/wAAAIjKgeL/AAAAKdBZW8Mx24oYMcCKAinDidhZ W8MAAAAAAAAAAKgBdA7HAgAAAIDHAwEAAADrDMcCAAAAwMcDgAAAAKgCdAOACwKoBHQDgAsE w41AAIP4AnUNxwIAAADAxwOAAAAAw4P4AXUNxwIAAABAxwOAAAAAw8cCAAAAgMcDAQAAAMON QABTicOD4HCD4weD+CByEHY4g/gwckF2KYP4QHQ0W8OFwHYHg/gQdBFbw8cCAQAAAIXbdSOA CgJbw8cCAAAAAFvDxwICAAAAW8PHAgEAAABbw8cCAwAAAFvDAAAAAFNRg+wEicOJ0Yniicjo LgYAAIngMNLoRf7//4gUBInY6Lvx//+FwHUYieKJ2Oiu/v//hcB0C4nY6EPy//+Jw+vdidjo mPH//4XAdASFyXUEidjrAjHAg8QEWVvDAAAAAFNSicIx24nQ6HPx//+FwHUMidDoCPL//0OJ wuvpidhaW8MAAAAAAAAAAAAAAAAAAFFWV1WJxonRid0x/4XAdDiF7XZagDkAdBi7AgAAAInK ifDoqQUAAIP4/3UMXV9eWcNmxwYAAOs2ichN6LDx//+DxgJHicHryL4CAAAAgDkAdByJ84nK McDocwUAAIP4/3QOicjoh/H//0eJwevfifhdX15ZwwAAAAAAAAAAAAAAU1FSVldVg+wQicG6 AgAAAIlUJASJFCS6PQAAAOjPBQAAiceFwHUKuP/////pWgEAADnIdQ+4/////4PEEF1fXlpZ W8MpyNH4jSxFAAAAAI1FAuhppf//icaJRCQIhcB1D7j/////g8QQXV9eWllbw4nricroh9r/ /41XAonQZscELgAA6Ffa//+FwHRCAcCJRCQMg8AC6CWl//+Jx4nFhcB1Fonw6Ga4//+4//// /4PEEF1fXlpZW8OLdCQMi1wkDAH36Dja//9mxwcAAOsCMe2LRCQIieroRAUAAInCi0QkCOgp uP//iejoIrj//4XSdQ+4/////4PEEF1fXlpZW8ODPfjxyAEAdQXoEQYAAInI6IIAAACFwHQP uP////+DxBBdX15aWVvDicjosNn//41YAYtEJAQPr8PogaT//4nGhcB1GbgFAAAA6C3h//+4 /////4PEEF1fXlpZW8MPrxwkicroN87//4P4/3UWifDom7f//7j/////g8QQXV9eWllbw4nw 6NH4//+DxBBdX15aWVvDjUAAU1FSVldVg+wIiceFwHUKuP/////pYwEAAGaDOAB0G41QAmaD OgB0EmaDOj10DGaLcgKDwgJmhfZ17maDOgB1D7j/////g8QIXV9eWllbw2aDegIAD5TAMduL DfjxyAGIw4XJdUGF2w+FDQEAALgJAAAA6LGj//+JwYXAdQ+4/////4PECF1fXlpZW8Oj+PHI AYPACIlY+DH2iVj8o/DxyAHpyAAAAInaifjo2QAAAIXbD4XDAAAAhcAPj6wAAAD32Ist8PHI AYnGQI0ctQAAAACJBCSNQwiLFCSJRCQEAdCF7XU+6D+j//+JxYnBhcB1D7j/////g8QIXV9e Wllbw4sV+PHIAehd2P//i0QkBIscJAHoMdKj8PHIAeiYlv//6zmJwonI6C3V//+JwYXAdQ+4 /////4PECF1fXlpZW8OLbCQEixXw8cgBAcWJ84no6EMBAACJLfDxyAGJDfjxyAHHRLEEAAAA AOsDjXD/ofDxyAGJPLHGBAYAMcCDxAhdX15aWVvDi8BTUVZXVYnHidWLNfjxyAHp1wAAAGaL D4n7ZoXJD4TGAAAAMcBmiwLoNAQAAInBMcBmiwPoKAQAAGY5wQ+FpwAAAGaDOj0PhYoAAACL FfjxyAGJ9ynXwf8Che0PhGwAAACLFonxhdJ0D4tBBIkBi1kEg8EEhdt18Ys18PHIAYX2dEOJ 8IA8BwB0B4nQ6GC1//+LLfjxyAGJzinuixXw8cgBwf4CiciJ8+hkAAAAiQ3w8cgBOfd9Do0E D0CKEEeIUP8593z1McBdX15ZW8ONRwFdX15ZW8ODwgJmi0MCg8MCZoXAD4U6////g8YEixaF 0g+FH////6H48cgBKfDB+AJdX15ZW8MAAAAAAAAAAAAAAFFWV4nZOcJ0QHMojTwaOcd2IY13 /408GIzaTwaOwv1OT9HpZvOlEclGR2bzpAf8X15Zw4nWiceM2gaOwlHB6QLzpVmD4QPzpAdf XlnDAAAAUoM9IPLIAQB0KDHSihCKkjHyyAGA4gGB4v8AAAB0EzHSihDB4giKQAEl/wAAAAnQ WsOKACX/AAAAWsMAAAAAAFNRUoPsBInDieLoYQAAAIngMNLoePj//4gUBIE9ONNAAKQDAAB1 JoM9IPLIAQB0HTHAigQkioAx8sgBJAEl/wAAAHQJidjohQIAAOsVagGNRCQEUC7/FSyxQACJ 4Ohe////g8QEWllbwwAAAAAAAABTicOJ0PbH/3QMidqIWAHB6giIEFvDiBpbwwAAAAAAAFFW V1WJ5YPsBInGhdJ1BzHA6YMAAACF2w+GdgAAAIA6AHUShfZ0BWbHBgAAMcCJ7F1fXlnDgz0g 8sgBAHQZMcCKAoqAMfLIASQBJf8AAAB0BoB6AQB0PYnQ6KL3//9qAY1N/InHUTnDcwKJ2FBS agihONNAAFAu/xXIsUAAhcB0E4X2dAaLRfxmiQaJ+InsXV9eWcO4/////4nsXV9eWcNTUVdm iwiJ02Y5ynQWicJmizqDwAJmhf90B2Y7GHQE6+wxwF9ZW8MAAAAAAAAAAABTUVZXVYPsBIkE JInXZoE9Cc1AAACAcw5XUC7/FdSxQADpwwAAAOiV1P//AcCNSAGJyOhpn///icKJxYXAD4Sm AAAAhf91BDH26yuJ+Ohu1P//AcCNSAGJyOhCn///icaFwHUSidDohbL//zHAg8QEXV9eWVvD ixQkicuJ6Oj+yP//g/j/dR2J6Ohisv//hfZ0B4nw6Fey//8xwIPEBF1fXllbw4X2dCKJy4n6 ifDozcj//4P4/3USifDoMbL//zHAg8QEXV9eWVvDVlUu/xXQsUAAicKJ6OgUsv//hfZ0B4nw 6Amy//+J0IPEBF1fXllbwwAAAAAAAAAAAAAAAAAAU1FSVos19PHIAYX2dEGLFoPGBIXSdDiJ 0OgS+P//jVgBjQRdAAAAAOhznv//icGFwHTb6Cj4//+D+P90CYnI6DT6///ryInI6KOx///r v15aWVvDAAAAAAAAAAAAAAAAZj1hAHIJZj16AHcDg+ggw1KJwugYAAAAhcB0A4PqIYnQWsMA AAAAAAAAAAAAAAAAUonC6EgAAACFwHQfgfqBggAAchGB+pqCAAB3CboBAAAAidBawzHSidBa w4jQ/sAl/wAAAIqAEM9AADHSJICIwonQWsMAAAAAAAAAAAAAAABSgz0g8sgBAHQxicLB6giB 4v8AAACKkjHyyAGA4gGB4v8AAAB0FSX/AAAA6BAAAACFwHQHuAEAAABawzHAWsMAgz0g8sgB AHQpgT0400AApAMAAHQKhcB0G7gBAAAAwyX/AAAAioBB1EAAJAgl/wAAAMMxwMMAAAAAAAAA AAAAAP8loLFAAP8lbLFAAP8laLFAAP8lmLFAAP8ltLFAAP8lgLFAAP8lULFAAP8llLFAAP8l 1LFAAP8lLLFAAP8l0LFAAP8lVLFAAP8lDLJAAP8lQLFAAP8lfLFAAP8l2LFAAP8lSLFAAP8l cLFAAP8ljLFAAP8lxLFAAP8lhLFAAP8lZLFAAP8lpLFAAP8lYLFAAP8liLFAAP8lzLFAAP8l PLFAAP8lXLFAAP8lBLJAAP8ldLFAAP8lALJAAP8lWLFAAP8l3LFAAP8lELJAAP8l6LFAAP8l /LFAAP8lqLFAAP8lwLFAAP8lCLJAAP8lyLFAAP8lnLFAAP8l8LFAAP8lNLFAAP8l+LFAAP8l 7LFAAP8l4LFAAP8l9LFAAP8lkLFAAP8lvLFAAP8lTLFAAP8leLFAAP8lRLFAAP8luLFAAP8l OLFAAP8lrLFAAP8l5LFAAP8lFLJAAP8lsLFAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA8sAAAAAAAAAAAAAAcsgAALLEAAESw AAAAAAAAAAAAACeyAAA0sQAAAAAAAAAAAAAAAAAAAAAAAAAAAAA0sgAAAAAAAEayAABUsgAA ZLIAAHKyAACCsgAAmrIAAKiyAADAsgAAzrIAANyyAADwsgAACrMAABSzAAAgswAAMrMAAESz AABcswAAcrMAAIizAACcswAAtLMAAMqzAADYswAA7LMAAPyzAAAMtAAAIrQAADi0AABMtAAA WLQAAGq0AAB6tAAAlLQAAKK0AAC+tAAA1rQAAOa0AADytAAACLUAACC1AAA6tQAAVLUAAGC1 AABytQAAgrUAAJK1AACwtQAAvLUAAMa1AADUtQAA4rUAAP61AAAOtgAAHLYAACy2AABCtgAA WLYAAAAAAAA0sgAAAAAAAEayAABUsgAAZLIAAHKyAACCsgAAmrIAAKiyAADAsgAAzrIAANyy AADwsgAACrMAABSzAAAgswAAMrMAAESzAABcswAAcrMAAIizAACcswAAtLMAAMqzAADYswAA 7LMAAPyzAAAMtAAAIrQAADi0AABMtAAAWLQAAGq0AAB6tAAAlLQAAKK0AAC+tAAA1rQAAOa0 AADytAAACLUAACC1AAA6tQAAVLUAAGC1AABytQAAgrUAAJK1AACwtQAAvLUAAMa1AADUtQAA 4rUAAP61AAAOtgAAHLYAACy2AABCtgAAWLYAAAAAAABVU0VSMzIuZGxsAEtFUk5FTDMyLmRs bAABAENoYXJVcHBlckJ1ZmZBAAABAENsb3NlSGFuZGxlAAIAQ3JlYXRlRXZlbnRBAAADAENy ZWF0ZUZpbGVBAAQAQ3JlYXRlVGhyZWFkAAAFAERlbGV0ZUNyaXRpY2FsU2VjdGlvbgAGAERl bGV0ZUZpbGVBAAcARW50ZXJDcml0aWNhbFNlY3Rpb24AAAgARXhpdFByb2Nlc3MACQBFeGl0 VGhyZWFkAAAKAEZsdXNoRmlsZUJ1ZmZlcnMAAAsARnJlZUVudmlyb25tZW50U3RyaW5nc0EA DABHZXRBQ1AAAA0AR2V0Q1BJbmZvAA4AR2V0Q29tbWFuZExpbmVBAA8AR2V0Q29tbWFuZExp bmVXABAAR2V0Q3VycmVudERpcmVjdG9yeUEAABEAR2V0Q3VycmVudFByb2Nlc3NJZAASAEdl dEN1cnJlbnRUaHJlYWRJZAAAEwBHZXRDdXJyZW50VGhyZWFkAAAUAEdldEVudmlyb25tZW50 U3RyaW5ncwAVAEdldEZpbGVBdHRyaWJ1dGVzQQAAFgBHZXRGaWxlVHlwZQAXAEdldEZ1bGxQ YXRoTmFtZUEAABgAR2V0TGFzdEVycm9yAAAZAEdldExvY2FsVGltZQAAGgBHZXRNb2R1bGVG aWxlTmFtZUEAABsAR2V0TW9kdWxlRmlsZU5hbWVXAAAcAEdldE1vZHVsZUhhbmRsZUEAAB0A R2V0T0VNQ1AAAB4AR2V0UHJvY0FkZHJlc3MAAB8AR2V0U3RkSGFuZGxlAAAgAEdldFRpbWVa b25lSW5mb3JtYXRpb24AACEAR2V0VmVyc2lvbgAAIgBJbml0aWFsaXplQ3JpdGljYWxTZWN0 aW9uACMATGVhdmVDcml0aWNhbFNlY3Rpb24AACQATG9hZExpYnJhcnlBAAAlAE1vdmVGaWxl QQAmAE11bHRpQnl0ZVRvV2lkZUNoYXIAJwBTZXRDb25zb2xlQ3RybEhhbmRsZXIAKABTZXRF bnZpcm9ubWVudFZhcmlhYmxlQQApAFNldEVudmlyb25tZW50VmFyaWFibGVXACoAU2V0RXZl bnQAACsAU2V0RmlsZVBvaW50ZXIAACwAU2V0TGFzdEVycm9yAAAtAFNldFN0ZEhhbmRsZQAA LgBTZXRVbmhhbmRsZWRFeGNlcHRpb25GaWx0ZXIALwBUbHNBbGxvYwAAMABUbHNGcmVlADEA VGxzR2V0VmFsdWUAMgBUbHNTZXRWYWx1ZQAzAFVuaGFuZGxlZEV4Y2VwdGlvbkZpbHRlcgAA NABWaXJ0dWFsQWxsb2MAADUAVmlydHVhbEZyZWUANgBWaXJ0dWFsUXVlcnkAADcAV2FpdEZv clNpbmdsZU9iamVjdAA4AFdpZGVDaGFyVG9NdWx0aUJ5dGUAOQBXcml0ZUZpbGUAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEBAQBDb3B5OiAgICAgIABTY2FsZTogICAg IABBZGQ6ICAgICAgIABUcmlhZDogICAgIAAtLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tCgAAVGhpcyBzeXN0ZW0gdXNlcyAl ZCBieXRlcyBwZXIgRE9VQkxFIFBSRUNJU0lPTiB3b3JkLgoAAABBcnJheSBzaXplID0gJWQs IE9mZnNldCA9ICVkCgAAAFRvdGFsIG1lbW9yeSByZXF1aXJlZCA9ICUuMWYgTUIuCgAAAEVh Y2ggdGVzdCBpcyBydW4gJWQgdGltZXMsIGJ1dCBvbmx5CgAAAAB0aGUgKmJlc3QqIHRpbWUg Zm9yIGVhY2ggaXMgdXNlZC4KAABZb3VyIGNsb2NrIGdyYW51bGFyaXR5L3ByZWNpc2lvbiBh cHBlYXJzIHRvIGJlICVkIG1pY3Jvc2Vjb25kcy4KAAAAAFlvdXIgY2xvY2sgZ3JhbnVsYXJp dHkgYXBwZWFycyB0byBiZSBsZXNzIHRoYW4gb25lIG1pY3Jvc2Vjb25kLgoAAAAARWFjaCB0 ZXN0IGJlbG93IHdpbGwgdGFrZSBvbiB0aGUgb3JkZXIgb2YgJWQgbWljcm9zZWNvbmRzLgoA ICAgKD0gJWQgY2xvY2sgdGlja3MpCgAASW5jcmVhc2UgdGhlIHNpemUgb2YgdGhlIGFycmF5 cyBpZiB0aGlzIHNob3dzIHRoYXQKAHlvdSBhcmUgbm90IGdldHRpbmcgYXQgbGVhc3QgMjAg Y2xvY2sgdGlja3MgcGVyIHRlc3QuCgAAV0FSTklORyAtLSBUaGUgYWJvdmUgaXMgb25seSBh IHJvdWdoIGd1aWRlbGluZS4KAAAAAEZvciBiZXN0IHJlc3VsdHMsIHBsZWFzZSBiZSBzdXJl IHlvdSBrbm93IHRoZQoAAHByZWNpc2lvbiBvZiB5b3VyIHN5c3RlbSB0aW1lci4KAAAAAEZ1 bmN0aW9uICAgICAgUmF0ZSAoTUIvcykgICBSTVMgdGltZSAgICAgTWluIHRpbWUgICAgIE1h eCB0aW1lCgAlcyUxMS40ZiAgJTExLjRmICAlMTEuNGYgICUxMS40ZgoAAAAAAAAAAAAAAAAA AEAAAAAAgIQuQQAAAAAAAAhAmpmZmZmZuT+N7bWg98awPgAAAACAhC5Bje21oPfGsD4AAAAA AAAQQE5vdCBlbm91Z2ggbWVtb3J5IHRvIGFsbG9jYXRlIGZpbGUgc3RydWN0dXJlcw0KAAAA AABGbG9hdGluZy1wb2ludCBzdXBwb3J0IG5vdCBsb2FkZWQNCgBUWgAAVW5hYmxlIHRvIGFs bG9jYXRlIHNlbWFwaG9yZSBkYXRhDQoAVW5hYmxlIHRvIGFsbG9jYXRlIHNlbWFwaG9yZSBk YXRhDQoAVVNFUjMyLkRMTAAAR2V0QWN0aXZlV2luZG93AFRoZSBpbnN0cnVjdGlvbiBhdCAw eDAwMDAwMDAwIGNhdXNlZCBhIHN0YWNrIG92ZXJmbG93IGZsb2F0aW5nIHBvaW50CmV4Y2Vw dGlvbi4KAAAAAFRoZSBpbnN0cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNhdXNlZCBhIHN0YWNr IHVuZGVyZmxvdyBmbG9hdGluZyBwb2ludApleGNlcHRpb24uCgAAAFRoZSBpbnN0cnVjdGlv biBhdCAweDAwMDAwMDAwIGNhdXNlZCBhIGRlbm9ybWFsIG9wZXJhbmQgZmxvYXRpbmcgcG9p bnQKZXhjZXB0aW9uLgoAAFRoZSBpbnN0cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNhdXNlZCBh IGRpdmlzaW9uIGJ5IHplcm8gZmxvYXRpbmcgcG9pbnQKZXhjZXB0aW9uLgoAAFRoZSBpbnN0 cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNhdXNlZCBhbiBpbmV4YWN0IHZhbHVlIGZsb2F0aW5n IHBvaW50CmV4Y2VwdGlvbi4KAAAAAFRoZSBpbnN0cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNh dXNlZCBhbiBvdmVyZmxvdyBmbG9hdGluZyBwb2ludCBleGNlcHRpb24uCgBUaGUgaW5zdHJ1 Y3Rpb24gYXQgMHgwMDAwMDAwMCBjYXVzZWQgYW4gdW5kZXJmbG93IGZsb2F0aW5nIHBvaW50 IGV4Y2VwdGlvbi4KAAAAAFRoZSBpbnN0cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNhdXNlZCBh biBpbnZhbGlkIG9wZXJhdGlvbiBmbG9hdGluZyBwb2ludApleGNlcHRpb24uCgAAAABUaGUg aW5zdHJ1Y3Rpb24gYXQgMHgwMDAwMDAwMCByZWZlcmVuY2VkIG1lbW9yeSAAAAAAYXQgMHgw MDAwMDAwMC4KVGhlIG1lbW9yeSBjb3VsZCBub3QgYmUgAHJlYWQuCgAAd3JpdHRlbi4KAAAA QSBwcml2aWxlZ2VkIGluc3RydWN0aW9uIHdhcyBleGVjdXRlZCBhdCBhZGRyZXNzIDB4MDAw MDAwMDAuCgAAAEFuIGlsbGVnYWwgaW5zdHJ1Y3Rpb24gd2FzIGV4ZWN1dGVkIGF0IGFkZHJl c3MgMHgwMDAwMDAwMC4KAEFuIGludGVnZXIgZGl2aWRlIGJ5IHplcm8gd2FzIGVuY291bnRl cmVkIGF0IGFkZHJlc3MgMHgwMDAwMDAwMC4KAAAAQSBzdGFjayBvdmVyZmxvdyB3YXMgZW5j b3VudGVyZWQgYXQgYWRkcmVzcyAweDAwMDAwMDAwLgoAAAAAVGhlIHByb2dyYW0gZW5jb3Vu dGVyZWQgZXhjZXB0aW9uIDB4MDAwMDAwMDAgYXQgAAAAAGFkZHJlc3MgMHgwMDAwMDAwMCBh bmQKY2Fubm90IGNvbnRpbnVlLgoAAAAAVGhyZWFkIGhhcyBubyB0aHJlYWQtc3BlY2lmaWMg ZGF0YQ0KAAAAAFVuYWJsZSB0byByZXNpemUgdGhyZWFkLXNwZWNpZmljIGRhdGENCgBVbmFi bGUgdG8gcmVzaXplIHRocmVhZC1zcGVjaWZpYyBkYXRhDQoAVE1QAFRFTVAAAAAAVE1QRElS AABURU1QRElSAAAAAAB3YisAY29uaW4kAABjb25vdXQkAGNvbgBDX0ZJTEVfSU5GTwBDX0ZJ TEVfSU5GTz0AAAAAQUJOT1JNQUwgVEVSTUlOQVRJT04NCgAAX19iZ250aGQAAAAAY29uAAAA HwA7AFoAeACXALUA1ADzABEBMAFOAW0BAAAfADwAWwB5AJgAtgDVAPQAEgExAU8BbgEAAAAA AADwfwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAADVw8N///+9HNXDw3///70c1cPDf///vRzVw8N///+9H BMBAABDAQAAcwEAAKMBAAAAAAACAhG5BAAAAAICEbkEAAAAAYON2QQAAAABg43ZBAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACAAAAAQAAAAAA AAAAAAAAAAAAAAAAAAAAAAIAAAACAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAGEAAAAAAQAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABwGUAAAP/////cHUAA5B1AAOQd QADoHUAA8B1AAOQdQADkHUAA5B1AAOQdQADkHUAA5B1AAOQdQADkHUAA5B1AAOQdQADkHUAA 5B1AAOQdQAAAAAAA9AAAAGAzQABgM0AAgDNAAIAzQAB/EgAAAAAAAAAAAAACAAAAAQAAAAMA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABAAAABQAAAAkAAAAAAAAAAAAAAAAAAAAAAAAA RVNUAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAARURUAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAADQzUAAUc5AAFBG AAABAAAAEA4AAAEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAICARUAAgEVAAIBFQAAAAQEB AQEBAQEBAwMDAwMBAQEBAQEBAQEBAQEBAQEBAQEKDAwMDAwMDAwMDAwMDAwMODg4ODg4ODg4 OAwMDAwMDAxYWFhYWFhISEhISEhISEhISEhISEhISEhISAwMDAwMDJiYmJiYmIiIiIiIiIiI iIiIiIiIiIiIiIiIDAwMDAEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAMDEyMzQ1 Njc4OWFiY2RlZgAAAAAwMTIzNDU2Nzg5YWJjZGVmZ2hpamtsbW5vcHFyc3R1dnd4eXoAAAAA MDEyMzQ1Njc4OWFiY2RlZmdoaWprbG1ub3BxcnN0dXZ3eHl6AAAAADAxMjM0NTY3ODlhYmNk ZWZnaGlqa2xtbm9wcXJzdHV2d3h5egAAAABTdGFjayBPdmVyZmxvdyENCgAAAFBaQABQWkAA AAAAAAAAAKACQAAAAAAAAADIBUAAAAAAAABAnAxAAAAAAAAgvL4ZQAAAAAS/yRuONECetXAr qK3FnWlA1abP/0kfeMLTQOCM6YDJR7qTqEGO3vmd++t+qlFDx5EOpq6gGeOjRhcMdYGGdXbJ SE3lXT3FXTuLnpJam5cgigJSYMQldQAAAAAAAACA/38AAAAAUMlAAFTJQABcyUAAZMlAAGzJ QAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABQAAAABAAAA AgAAAAIAAAADAAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAHDSQAD//////////wEAAAD/////AgAAAP////8BAAAA /////wIAAAAdAADAAgAAADoBAMACAAAABQAAwAIAAAA6AQDAAgAAADoBAMABAAAA/////wEA AAD/////AQAAAP////8CAAAAlAAAwAEAAACVAADAAAAAAAEAAAAACQEBCwYEBQUFAgP//wcJ BggB/wAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAJIhAABCJQABwe0AAAQAAAAAAAQAoh0AAAAAAAP///39VVVVV ////PzMzMzOqqqoqJEmSJP///x8cx3EcmZmZGXTRRRdVVVUVEzuxE5IkSRIRERER////Dw8P Dw+O4zgOXkN5DczMzAwMwzAMuuiiCyxkIQuqqqoKo3A9Comd2AleQnsJSZIkCbDc0wiIiIgI hBBCCP///wd88MEHh4eHBwd1UAfHcRwHAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACAgICAgICAgICAgICAgI CAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAAgMDAwM DAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMCAoKCgoKCQkJCQkJCQkJCQkJCQkJCQkJCQkJ CQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQwMDAwMDAwMDAwMDAwMDAwM DAwMDAwMDAwMDAwMAAAAAAAgwBhAAAAB4BZAAAACLBlAAAAgbBdAAAAgoBtAAAAgGHJAAAAg 8FRAAAAgwFlAAAAgAHhAAAAB4RZAAAAgQBxAAAAgwCBAAAAfXEVAAAAKvEpAAAAg/HJAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEAAAeAEAAB0w LDA5MFAwZDBzMIUwnzClMKswsTC3ML0wyjDnMAMxCTEUMTAxUDGGMZMxoDGtMboxxzHUMeEx /jEEMigyLjI5MmQyajJzMqIyqDKwMrky9TL7MgEzEzMgMzQzQTNHM1kzZjOAM5AzqjO8M8Iz yjPQM9Yz3DPiM+kz8DP3M/4zBTQSNBg0QDStNOw0RzVNNWU1azVwNXY1fDWbNaE1pjWsNbI1 vjXHNdE15DXvNfU1+jUANgY2DDYzNmU2JTcxN243czefN7c3CDhJOHQ4hTiROMM4yDjOONQ4 8Dj4OP04AzkJORE5IjkvOTk5WzlkOWo5jjkMOmI6CjsmO1I7pjuxO7o7wDvFO/E7/zslPDc8 ZTxuPJM8vDzUPAI9Yz2iPbY90j3dPQM+DT4iPi4+NT47PkI+Sz5ePmw+cT52Pog+kD6YPqI+ qT6vPrY+uz7BPs0+2D7fPu0+ID87P0Y/TT9kP6k/yz/XP9w/6j/wP/w/ACAAADQAAAABMAcw DjAZMDUwRTBLMGcwgDCJMJkwsjC6MMYw2TDfMPIw5jjnPv0+zT8AAAAwAADcAAAAxDDyMPkw GTEhMUExSjFqMXIxljGcMaQxrzHmMQAyEzIaMiUyKjJnMyY0LTQ5NEA0TDVdNWI1dzV8NYM1 iDWpNa81tzW9Nak2sjfvOAk5bjmDOYg5lDmfOac5rTm7OdA55DnwOfc5ETotOjU6PzpgOmU6 dzqBOo86oTqrOr062zriOug67ToEOwo7EDtYPV09Yz19PYk9kz2jPbA9uz3PPew9/D0CPgo+ Dz4VPmI+bz53PoU+iz6YPqE+vj7GPgI/CD9CP04/hz+NP6Q/rz+1P90/AAAAQAAAyAEAABAw GTAhMDUwSjBuMHMwgzCLMJgwtjC7MMUw4DD7MWcydDKVMqoyvjLdMvUy+jIAMxAzJTM7M0wz UjNnM3YzizORM58zpDOwM7YzuzPSM+UzFDQaNCc0RDR1NHs0gzSJNI80oTSrNLo0xzTXNPU0 EzU9NUk1UDVgNXM1ijWbNak1zDXXNd016jX4NQI2CDYUNhw2LjY5NkI2SDZgNm02eDZ9Now2 mTbBNs023zbvNgo3HTclNzE3OTdFN003VTdmN3o3mDenN7k33jfvN/U3/zchODs4TDhtOHk4 gDiPOKI4qTi+OM04BjkOOSU5MjlPOVc5ZzmEOZE5lzmrObA5tTm6Ob85xDnJOc851TnbOeE5 5zntOfI59zn8OQE6BjoLOhA6FTobOiE6JzotOjM6OTo+OkM6TTpSOlc6XTpoOm06czp5On86 izqWOpw6pDqqOq46vzrEOso62zrkOuo6/joEOwk7DzsUOxo7HzslO007ZTuIO6o7eTyHPKE8 1DzbPOY87jxQPSM+LT43PkE+Sz5VPl8+aT5zPoI+lz6iPq0+tD67PsI+yT7XPgo/FT8sPzA/ ND84Pzw/QD9EP3Q/AAAAUAAA0AAAAFQwcDCRMKUwwDDGMMww8DD9MBYxJjExMT0xQzFOMVUx XzFzMZAxljHgMR4yPDJTMlkyYzJvMncyfTLwMvgyNjPeM5s04jQhNT01RzVgNW01gDWLNZc1 oTW8Ncg17DUMNiY2SjZWNmA2jjaqNq82ujbQNvI2+jYKNyQ3KzcxN0w3ZDd2N4s3wDfLN903 4zf5NzU4OzhHOFs4YThmOJg4ozioOL84zTizOd458jlaOmA6hTqLOpE6mjrBOsk7zTvRO9U7 2TtsPAAAAGAAAFgAAAAtMjEyNTI5Mj40aTd4N4A32jgVOz07lTulO9c7dzzCPPU8DT0SPXU9 tj0XPiM+VD5rPnc+nT61Ps4+Dz8uP04/Zz9tP34/jD+UP6I/5z8AAABwAAAQAQAAADAiMCww QjBQMBoxRjFLMXExkTGiMa4xtTG7McAxxjHQMfMxIDIuMlUydDLKMtAy2zLvMggzGzMhMzQz OjNGM84z1DPuM/Uz+jMANBo0ITQmNC40PjRKNHc0fTSPNJk0uDTANNw05DSwNbo1wTXLNdM1 4DXqNfE1/TUENo42kjboNuw28Db0Nvg2/DYANwQ3CDcMNxA3FDcwN5Q3pDezN7g3vjfENwM4 CDgOOBQ4JzgtODg4RDhSOF04ajh7OKI4tTjLONw44jj2OAs5IDk3OTw5VTlxOYQ5jjmnOc45 0znkOT47WTt2O347hjuWO5w7qjuvO707wjtGPUw9Uj1YPXg90D3QPgAAAIAAAMgAAAArMDAw UzBkMHkwfjCGMKAwzTDTMN4wFTEyMT4xVjG9Mewx/TE5Mtsy9jL/Mg0zFzM2M2wzcTOPMww0 GTR6NLw0+jQQNUo1ajVtNoY2ljafNqk2zDblNv02CzciNy83TTejN7Q3wzfPN+Y3+TcHODI4 WDhnOHo4kziuOL443jjyOP44FDkfOTU5TzlcOeM58DkSOvA6+DoOOx87TTvSO/w7FzwfPC08 kDwUPUw9jD3bPSY+Xj5wPs8+1T77Pmk/AAAAkAAAfAAAAAQwEjJZMiAzVjNkM5szyDPsMx40 LzRCNGY0hzTDNO80CDUSNSQ1ZjWDNZA1wjXSNUo2VzZzOWM6lTqlOss6CjsfO047XztlO3c7 mzvdOwk8IjwsPD48hjzzPAA9TT1ZPWc9hz33PQQ+MT45PqA+rT5fP5Y/AKAAAIgAAABWMHMw hzCyMLsw1jBIMkIyPDI2MjAyKjIkMh4yGDISMgwyBjIAMvox9DHuMegx4jHcMdYx0DHKMcQx vjG4MbIxrDGmMaAxmjGUMY4xiDGCMXwxdjFwMWoxZDFeMVgxUjFMMUYxQDE6MTQxLjEoMSIx HDEWMRAxCjEEMf4w+DDyMADAAABIAAAAcDp0Ong6fDobPSQ9KD0sPTA9ND04PTw9QD1EPUg9 TD1QPVQ9WD1cPWA9ZD1oPXQ9eD18PYA91D7YPgQ/CD8MPwDQAABAAAAARDVKNXo1UDVWNYA1 XDWGNYw1kjWYNWI1aDVuNbwwwDBUMVgxXDFgMWQxwDJ0NZQzmDOcM6gzAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAA --Message-Boundary-10580 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Content-description: Text from file 'make.bat' wcl386 /zq /otexan stream_d.c p55clock.c @copy stream_d.exe stream.exe > nul stream.exe > 500.txt wcl386 /zq /otexan stream_d.c p55clock.c /dASMHACK=1 @copy stream_d.exe fastream.exe > nul fastream.exe > 500f.txt --Message-Boundary-10580 Content-type: text/plain; charset=US-ASCII Content-disposition: inline Content-description: Attachment information. The following section of this message contains a file attachment prepared for transmission using the Internet MIME message format. If you are using Pegasus Mail, or any another MIME-compliant system, you should be able to save it or view it from within your mailer. If you cannot, please ask your system administrator for assistance. ---- File information ----------- File: p55clock.c Date: 17 Dec 1999, 4:30 Size: 1645 bytes. Type: Program-source --Message-Boundary-10580 Content-type: Application/Octet-stream; name="p55clock.c"; type=Program-source Content-disposition: attachment; filename="p55clock.c" Content-transfer-encoding: BASE64 Ly8jaW5jbHVkZSA8c3RkaW8uaD4NCiNpbmNsdWRlIDx0aW1lLmg+DQoNCnN0YXRpYyBkb3Vi bGUgdGltZXJhdGlvOw0Kc3RhdGljIGludCBmdG1wWzJdOw0Kc3RhdGljIGludCB0dG1wWzJd Ow0KDQojaWZkZWYgX19XQVRDT01DX18NCg0KLyogR2V0IGluaXRpYWwgVFNDIHZhbHVlICov DQp2b2lkIHJkdHNjMCgpOw0KI3ByYWdtYSBhdXggcmR0c2MwID0gCQlcDQoiIC41ODZwICIJ CQlcDQoiIHJkdHNjICIJCQlcDQoiIG1vdiBkd29yZCBwdHIgdHRtcFswXSxlYXggIglcDQoi IG1vdiBkd29yZCBwdHIgdHRtcFs0XSxlZHggIiBtb2RpZnkgW2VheCBlZHhdOw0KDQovKiBT dWJ0cmFjdCBpbml0aWFsIFRTQyB2YWx1ZSwgdGhlbiBjb252ZXJ0IHRvIGRvdWJsZSAqLw0K ZG91YmxlIHJkdHNjKCk7DQojcHJhZ21hIGF1eCByZHRzYyA9IAkJXA0KIiAuNTg2cCAiCQkJ XA0KIiByZHRzYyAiCQkJXA0KIiBzdWIgZWF4LCB0dG1wWzBdCSAiCVwNCiIgc2JiIGVkeCwg dHRtcFs0XQkgIglcDQoiIG1vdiBkd29yZCBwdHIgZnRtcFswXSxlYXggIglcDQoiIG1vdiBk d29yZCBwdHIgZnRtcFs0XSxlZHggIglcDQoiIGZpbGQgcXdvcmQgcHRyIGZ0bXAgIgkJXA0K bW9kaWZ5IFtlYXggZWR4XSB2YWx1ZSBbODA4N107DQoNCiNlbHNlDQoNCi8qIEdldCBpbml0 aWFsIFRTQyB2YWx1ZSAqLw0Kc3RhdGljIHZvaWQgcmR0c2MwKCkgew0KICAgIF9hc20gew0K CXJkdHNjDQogICAgICAgIG1vdiBkd29yZCBwdHIgdHRtcFswXSxlYXgNCiAgICAgICAgbW92 IGR3b3JkIHB0ciB0dG1wWzRdLGVkeA0KICAgIH0NCn0NCg0KLyogU3VidHJhY3QgaW5pdGlh bCBUU0MgdmFsdWUsIHRoZW4gY29udmVydCB0byBkb3VibGUgKi8NCnN0YXRpYyBkb3VibGUg cmR0c2MoKSB7DQpkb3VibGUgZnRtcDsNCiAgICBfYXNtIHsNCglyZHRzYw0KICAgICAgICBz dWIgZWF4LCB0dG1wWzBdCSANCiAgICAgICAgc2JiIGVkeCwgdHRtcFs0XQkgDQogICAgICAg IG1vdiBkd29yZCBwdHIgZnRtcFswXSxlYXgNCiAgICAgICAgbW92IGR3b3JkIHB0ciBmdG1w WzRdLGVkeA0KICAgICAgICBmaWxkIHF3b3JkIHB0ciBmdG1wDQogICAgICAgIGZzdHAgZnRt cA0KICAgIH0NCiAgICByZXR1cm4gZnRtcDsNCn0NCg0KI2VuZGlmDQoNCnN0YXRpYyBpbnQg Y2FsaWJyYXRlZCA9IDA7DQoNCnN0YXRpYyB2b2lkIGNhbGlicmF0ZSgpIHsNCmludCB0Ow0K ZG91YmxlIHQwOw0KICAgIHJkdHNjMCgpOw0KDQojaWYgMQ0KICAgIHQgPSBjbG9jaygpOw0K ICAgIHdoaWxlKCBjbG9jaygpID09IHQgKSA7DQogICAgdDAgPSByZHRzYygpOw0KICAgIHQg PSBjbG9jaygpICsgNCpDTE9DS1NfUEVSX1NFQzsNCiAgICB3aGlsZSggY2xvY2soKTx0ICkg Ow0KICAgIHQwID0gcmR0c2MoKSAtIHQwOw0KI2Vsc2UNCiAgICB0MCA9IDUwMCo0MDAwMDAw LjA7DQojZW5kaWYNCg0KICAgIHRpbWVyYXRpbyA9IDQuMC90MDsNCg0KLy8gICAgcHJpbnRm KCJcbk1hY2hpbmUgc3BlZWQgKGNsb2Nrcy9zZWNvbmQpID0gJTEwZyBNaHpcblxuIix0MC80 MDAwMDAwLjApOw0KDQogICAgY2FsaWJyYXRlZCA9IDE7DQp9DQoNCmRvdWJsZSBzZWNvbmQo KSB7DQogICAgaWYoICFjYWxpYnJhdGVkICkgY2FsaWJyYXRlKCk7DQoNCiAgICByZXR1cm4g cmR0c2MoKSp0aW1lcmF0aW87DQp9DQoNCg== --Message-Boundary-10580 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Content-description: Text from file 'readme.txt' I performed this port to the WATCOM C/C++ compiler on my Athlon 500Mhz system. Here is a description of the files. 6763 stream_d.c - My minor modifications to the stream_d.c source. The most important modification was to change OFFSET from 0 to 70464. This number was determined experimentally. It also contains a conditionally compiled ASM hack to speed up just the COPY loop. 1645 p55clock.c - A different port of the timer code. It is far more accurate than previous timers, in particular the win32_clock.c timer code. 219 make.bat - Just a batch file for compiling everything up. 49152 stream.exe - The straight compile of the sources. 49152 fastream.exe - A compile of the sources with the assembly hack turned on. 1217 500.txt - The results without the assembly hack. 1217 500f.txt - The results with the assembly hack. -- Paul Hsieh qed@pobox.com --Message-Boundary-10580 Content-type: text/plain; charset=US-ASCII Content-disposition: inline Content-description: Attachment information. The following section of this message contains a file attachment prepared for transmission using the Internet MIME message format. If you are using Pegasus Mail, or any another MIME-compliant system, you should be able to save it or view it from within your mailer. If you cannot, please ask your system administrator for assistance. ---- File information ----------- File: stream.exe Date: 17 Dec 1999, 4:31 Size: 49152 bytes. Type: PCEXE --Message-Boundary-10580 Content-type: Application/Octet-stream; name="stream.exe"; type=PCEXE Content-disposition: attachment; filename="stream.exe" Content-transfer-encoding: BASE64 TVqAAAEAAAAEAAAA//8AALgAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAgAAAAA4fug4AtAnNIbgBTM0hdGhpcyBpcyBhIFdpbmRvd3MgTlQgY2hhcmFjdGVy LW1vZGUgZXhlY3V0YWJsZQ0KJABQRQAATAEFADAtWjgAAAAAAAAAAOAAggELAQISAJQAAAAe AAAAFIgBcBYAAAAQAAAAsAAAAABAAAAQAAAAAgAAAQALAAAAAAADAAoAAAAAAAAQiQEABAAA AAAAAAMAAAAAAAEAAAABAAAgAAAAEAAAAAAAABAAAAAAAAAAAAAAAACwAABkBgAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAACJAeAIAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEFV VE8AAAAAAAAAAAAQAAAAlAAAAAQAAAAAAAAAAAAAAAAAACAAAGAuaWRhdGEAAAAAAAAAsAAA AAgAAACYAAAAAAAAAAAAAAAAAABAAADAREdST1VQAAAAAAAAAMAAAAAWAAAAoAAAAAAAAAAA AAAAAAAAQAAAwC5ic3MAAAAAAAAAAADgAAAAFIgBAAAAAAAAAAAAAAAAAAAAAIAAAMAucmVs b2MAAAAAAAAAAIkBAAoAAAC2AAAAAAAAAAAAAAAAAABAAABCAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAMzr /ZCQkJAAAAAAAAAAAABTUVJWV1WB7FwBAABoNMBAAOgKBgAAg8QEaghodMBAAOj7BQAAg8QI aDTAQADo7gUAAIPEBGhAEwEAaEBCDwBorMBAAOjXBQAAg8QMaGDjNkBqAGjMwEAA6MMFAACD xAxqCmjwwEAA6LQFAACDxAi7AADwP2gYwUAAvwAAAEAx0uibBQAAg8QEg8IIMcmJmgzgQACJ ugyMwwCJigjgQACJigg4RgGJigiMwwCJigw4RgGB+gASegB1z2g0wEAA6F0FAACDxATotQMA AInBg/gBD41mAwAAaIDBQADoQAUAAIPEBOj4BAAA3ZwkQAEAADHS3QV0w0AA3YIQ4EAA2MmD wgjdmgjgQACB+gASegB1593Y6MkEAADcpCRAAQAA3A18w0AA3ZQkQAEAAOgSBQAA25wkWAEA AIuUJFgBAABSaMTBQADo1wQAAIPECImMJFgBAADbhCRYAQAA3LwkQAEAAOjcBAAA25wkWAEA AIuMJFgBAABRaADCQADooQQAAIPECGgYwkAA6JQEAACDxARoTMJAAOiHBAAAg8QEaDTAQADo egQAAIPEBGiEwkAA6G0EAACDxARouMJAAOhgBAAAg8QEaOjCQADoUwQAAIPEBGg0wEAAMcno RAQAAIPEBDH/6PoDAADdHAyJ+oPCCIuCCOBAAImCCDhGAYuCDOBAAImCDDhGAYH6ABJ6AHXd 6M0DAADcJAzdHAzowgMAAN1cDFCJ+t0FhMNAAN2CEDhGAdjJg8II3ZoIjMMAgfoAEnoAdefd 2OiWAwAA3GQMUN1cDFDoiQMAAN2cDKAAAACJ+t2CEOBAANyCEIzDAIPCCN2aCDhGAYH6ABJ6 AHXj6F4DAADcpAygAAAA3ZwMoAAAAOhLAwAA3ZwM8AAAAIn63QWEw0AA3YIQOEYB2MncghCM wwCDwgjdmgjgQACB+gASegB14d3Y6BYDAADcpAzwAAAAg8EI3ZwM6AAAAIP5UA+F/f7//zH2 ifGJ8zHS3QQM2MjcghDKQADdmhDKQADdglDKQADcHAzf4J4Pg0ABAACLglDKQACJhCRIAQAA i4JUykAAiYQkTAEAAIuEJEgBAACJglDKQACLhCRMAQAAiYJUykAA3YIwykAA3Bwc3+CeD4YN AQAAi4IwykAAiYQkUAEAAIuCNMpAAImEJFQBAACDwgiLhCRQAQAAg8FQiYIoykAAi4QkVAEA AIPDUImCLMpAAIP6IA+FUf///4PGCIP+UA+FP////2gMw0AA6G8CAAAxyTHSg8QE3YIQykAA 3A2Mw0AA2frdmhDKQADdgoDKQADcDZTDQADcslDKQACLmjTKQABTi7IwykAAVou6VMpAAFeL qlDKQABVi4IUykAAUIuaEMpAAFOD7AjdHCSLsXDKQABWaEzDQACDwQSDwgjo+wEAAIPEKIP5 EHWLMcCBxFwBAABdX15aWVvDUGg8wUAA6NkBAACDxAjplPz//4sEDImEJEgBAACLRAwE6cD+ //+LBByJhCRQAQAAi0QcBOnz/v//jYAAAAAAjVIAi9tTUVJWgey0AAAAMdLoTwEAAN2cJKgA AADoQwEAAN2UJKAAAADcpCSoAAAA3B2kw0AA3+CecuKDwgiLhCSgAAAAiUQU+IuEJKQAAACJ RBT8gfqgAAAAdbW6QEIPALgIAAAAMdvdBATcZAT43A2cw0AA6E0BAADbnCSwAAAAi4wksAAA ADnLfRs5yn0bg8AIPaAAAAB1zonQgcS0AAAAXlpZW8OJ2evhi7QksAAAADnzfQSJ8uvWidrr 0gAAAAAAAABSg+wIDzGjEOTIAYkVFOTIAei6AQAAicLoswEAADnQdPcPMSsFEOTIARsVFOTI AaMY5MgBiRUc5MgB3y0Y5MgB3Rwk6IgBAACNkKAPAADofQEAADnQcvcPMSsFEOTIARsVFOTI AaMY5MgBiRUc5MgB3y0Y5MgB3CQk3Rwk3QWsw0AA3DQkxwWgykAAAQAAAN0dIOTIAYPECFrD jYAAAAAAUFKDPaDKQAAAdCgPMSsFEOTIARsVFOTIAaMY5MgBiRUc5MgB3y0Y5MgB3A0g5MgB WljD6Cj////r0QAAAAAAAFNSg+wEjUQkFInji1QkEIkEJLjGykAA6FgBAACDxARaW8NQm9k8 JJv/NCTGRCQBH9ksJNn82WwkBJuNZCQIwwDp+wEAAAMQQABXQVRDT00gQy9DKyszMiBSdW4t VGltZSBzeXN0ZW0uIChjKSBDb3B5cmlnaHQgYnkgV0FUQ09NIEludGVybmF0aW9uYWwgQ29y cC4gMTk4OC0xOTk1LiBBbGwgcmlnaHRzIHJlc2VydmVkLgAAAAAAAAAAAAAAAMPDAAAAAAAA AAAAAAAAAABTg+wkicOJ4OiDAgAAiQKJ4OjqAgAAiQODxCRbw41AAFNSVoPsCInijUQkBOjP ////ixQkoTTkyAGLXCQEKcIrHTDkyAGJFCSJXCQEgfs2iUEAdge4/////+sXidjB4AUp2MHg AgHYweADjTQCiTQkifCDxAheWlvDjUAAUro05MgBuDDkyAHodP///1rDAABTicOJ0IsT6CQF AAD/QxBbw41AAFFWV1WJxotAEP8VKM1AAItGCItIDIP5AXQbhcl0EItGEP8VLM1AADHAXV9e WcPHQAwBAAAAimYMgOTPi24MiGYMi0YIi0gIg+Uwhcl1B4nw6BcJAACKTg0x//bBBHQUiM2A 5fqI6IhuDQwBvwEAAACIRg25kBdAAInw6H0JAACJwoX/dBiKZg2A5PqI44hmDYDLBInwiF4N 6I8XAAD2RgwgdAW6/////4t+DAnvi0YQiX4M/xUszUAAidBdX15ZwwAAAAAAAAAAAAAAAFNR UlWJ5YPsCLgBAAAA6HkYAAChcM1AAIPAAyT8MdIpxInhix1wzUAAicjoMBkAAKFwzUAAicqJ gfAAAACNRfjogwcAAOg2GQAAiexdWllbwwAAAAAAAAAAAAAAAAAAAFNSuk4yQAC7YjNAAIkV dM1AAIkdeM1AAFpbwwAAAAAAm90wm8ONQADdIJvDU1KAPanKQAAAdBa68BhAALv4GEAAiRV8 zUAAiR2AzUAAMcBmoYTNQADodhoAAFpbw41AAIA9qcpAAAB1w8OLwFKKJajKQACE5HU3iCWp ykAAMPYrwFDb49k8JFiKxDDSPAN1CeiX////iMaIwoA9/MxAAAB1DIg1qMpAAIgVqcpAAFrD wwAAAAAAAAAAAAAAAAAAAFNRUoPsEInDieBQLv8VlLFAADHAZosEJC1sBwAAiUMUMcBmi0Qk AkiJQxAxwGaLRCQGiUMMMcBmi0QkCIlDCDHAZotEJAqJQwQxwGaLRCQMx0Mg/////4kDMcBm i0QkDoPEEFpZW8MAAAAAAAAAAABTUVJWV1WD7ASJxbkMAAAAi1UQi0AQwfof9/m7zMlAAIt9 FInWgf+hfvv0fQq4/////+luAQAAi0UQicLB+h/3+QHHhfZ9CIPGDE+F9nz4hf99D7j///// g8QEXV9eWllbw42HbAcAAOhjGQAAhcB0BbvmyUAAjVcDidDB+h/B4gIbwsH4AonBjQT9AAAA AAH4weADAfiJwsHgAgHQAcG4ZAAAAI1XY4kEJInQwfof9zwkjZcrAQAAKcG/kAEAAInQwfof 9/8BwQ+/BHOLVQwByAHQjXj/i1UIidDB4AQp0MHgAotVBAHCidDB4AQp0It1AMHgAgHGhfZ9 C4HGgFEBAE+F9nz1iemJ8on4MdvoLR0AAOi0HwAAix3czkAAi00gAd6FyX0HiejoIhoAAIN9 IAB+Bis15M5AAIX2fQlPgcaAUQEA6/OB/95jAAB9D7j/////g8QEXV9eWllbw3Ujix3czkAA ge6AUQEAhdt+BIX2fS+4/////4PEBF1fXlpZW8ONhyGc//+NPMUAAAAAKcfB5wUBx8HnB4n4 wecCKccB/onwg8QEXV9eWllbwwBTUVJWiiXtykAAgOT4iOKIJe3KQACAygSIFe3KQACLFbjK QAC7rMpAAIXSdGa4HQAAAOhZIgAAhcB1IbgdAAAA6EsiAACJwYXAdRO6AQAAALi0w0AA6GIj AADrAonBoUDkyAGJWQSJAYlLCMdBCAAAAACLQwjGQBQAi0MIx0AMAAAAAIkNQOTIAYtLJoPD GoXJdZox9ok1ROTIAV5aWVvDMcDoDQAAAOk0JAAAuAMAAACNQABTUVJWicLB4AIp0MHgAgHQ vqzKQAABwAHGoUDkyAEx24XAdC+LCItABLoBAAAA9kANQHUT9kANCHUNOfByDz36ykAAcwIx 0uhUJAAAQ4nIhcl10YnYXlpZW8MAAFNRVldVicOLQhD/FSjNQACLQgiLSAyD+QF0HoXJdBOL QhD/FSzNQAC4/////+ncAAAAx0AMAQAAAPZCDAJ1IrgEAAAA6JclAACASgwgi0IQ/xUszUAA uP////9dX15ZW8OLQgiDeAgAdQeJ0OjfAwAAuQAEAACD+wp1R4pCDLkABgAAqEB1O4BKDRCL AsYADYsqRYtCBIkqQItyFIlCBDnwdR+J0OhmEgAAhcB0FItCEP8VLM1AALj/////XV9eWVvD gEoNEIsCiBiLOkeLagSJOkWLQgyJagSFwXUFO2oUdR+J0OgnEgAAhcB0FItCEP8VLM1AALj/ ////XV9eWVvDi0IQ/xUszUAAMcCI2F1fXllbwwAAAAAAAAAAAAAAAP8VJM1AAAXaAAAAw6Fc 5MgBjUAAw41AAOlXJQAAjUAA6acmAACNQABRVleJxonfidCJNVDkyAHoICoAAKNc5MgBhcB1 EYX2D4UFAgAAagEu/xVQsUAA6KUmAAAu/xWAsUAAMdKjAc1AAIkV9PHIAS7/FbSxQACJwonD ogfNQADB6BAl//8AAGgEAQAAZqMJzUAAMcCB4v//AABmoQnNQABoYOTIAaMLzUAAMcDB+giI 2IHi/wAAAKMPzUAAMcCIFQjNQACI0IsVD81AAGoAweIIu2DkyAEJwqMTzUAAiRUXzUAALv8V mLFAALpk5cgBiR3IzEAAMcC7CAIAALlk5cgB6GosAACJDdTMQAAu/xVosUAA6PgsAACJwooY o1TkyAGA+yJ1Hop4AUA433QOgDgAdAmKUAFAgPoidfKAOAB0HkDrG4oQ/sKB4v8AAAD2ghDP QAACdQiAOAB0A0Dr5YoQ/sKB4v8AAAD2ghDPQAACdANA6+qjxMxAAC7/FWyxQACFwA+EfQAA AOjSLAAAicJmixijWOTIAWaD+yJ1KmaLSAKDwAJmOdl0E2aDOAB0DWaLWAKDwAJmg/side1m gzgAdCeDwALrIrsCAAAAihD+woHi/wAAAISaEM9AAHULZoM4AHQFg8AC6+O7AgAAAIoQ/sKB 4v8AAACEmhDPQAB0CoPAAuvpuOTDQACj0MxAAIX2dDloBAEAAGhs58gBV75s58gBuwgCAAAu /xWYsUAAunDoyAGJNczMQACJ+L9w6MgB6CkrAACJPdjMQAC4AQAAAF9eWcOLwFNRVleJxonX agAu/xWgsUAAicOJ+jHA6LT9//+65MxAAP8VJM1AAOgMLAAAifDotTAAALghAAAA6IcQAAD/ FWTNQAC4/wAAAOh3EAAAX15ZW8OLwInDgz1Q5MgBAHQYgz1szUAAAHQmuv8AAAAxwP8VbM1A AOsX6LgwAAC6DwAAADHA6IwQAAD/FWDNQABTLv8VULFAAIvAU1FSVosVVOTIAYXSdA+J0DHb 6KkwAACJHVTkyAGLDVjkyAGFyXQPicgx9uiQMAAAiTVY5MgBXlpZW8MAAAAAAFNRUonC6IYx AACDehQAdSaKYg32xAJ0CcdCFIYAAADrFfbEBHQJx0IUAQAAAOsHx0IUABAAAItCFOjyHAAA i1oIiUMIi0IIg3gIAHUjikoNgOH4iEoNiM2LQgiAzQSNWhiIag2JWAjHQhQBAAAA6wSASgwI i0IIi0AIx0IEAAAAAIkCWllbwwAAAAAAAAAAAFZXBlWD7HCJ3onPMNu5ZAAAAIhcJGww/4kE JGaJXCQeMdtmiUwkHIlcJBCKOolUJGiE/w+EYQMAAItEJGiLbCRoiihFgP0ldBEx0ongiOqJ bCRo/9fpGAMAAInjiwaNVCRgiUQkYIno6DsDAACJxYtEJGBFiQaKRf+JbCRoiEQkFYTAD4QR AwAAPG4PhXABAACKXCQe9sMgdE/2w4B0H4sWg8IIiRbEUviLRCQQJokCi0QkaIA4AHWA6dwC AAD2w0APhOgAAACLBoPABIkGi1D8i0QkEIkCi0QkaIA4AA+FVf///+mxAgAA9sMQD4SJAAAA 9sOAdCuLDoPBCIkOxFH4i0QkEGYmiQKLRCRogDgAD4Uj////i0QkEIPEcF0HX17D9sNAdCqL HoPDBIkei1P8i0QkEGaJAotEJGiAOAAPhfT+//+LRCQQg8RwXQdfXsOLFoPCBIkWi1L8i0Qk EGaJAotEJGiAOAAPhcr+//+LRCQQg8RwXQdfXsP2w4B0KosGg8AIiQbEUPiLRCQQJokCi0Qk aIA4AA+Fm/7//4tEJBCDxHBdB19ew/bDQHQpiy6DxQSJLotV/ItEJBCJAotEJGiAOAAPhW3+ //+LRCQQg8RwXQdfXsOLDoPBBIkOi1H8i0QkEIkCi0QkaIA4AA+FRP7//4tEJBCDxHBdB19e w41MJGyJ44sGjVQkZIlEJGSNRCQ46OoFAACJxYtEJGSOwokGi0QkIItMJCSLVCQoAciLXCQs AdCLTCQwAdiLVCQ0AciLXCQEAdApw4pUJB6JXCQE9sIIdR2AfCQWIHUWg3wkBAB+D7ogAAAA ieD/1/9MJATr6otEJCCNXCQ4hcB+FjHSieCKE//Xi1QkIEpDiVQkIIXSf+qDfCQkAH4PujAA AACJ4P/X/0wkJOvqilwkFYD7c3Uz9kQkHiB0D4njifmMwono6NUEAADrToN8JCgAfkcx0ong JopVAP/Xi1QkKEpFiVQkKOvjgPtTdQ+J44n5jMKJ6OikBAAA6x2DfCQoAH4WMdKJ4CaKVQD/ 14tMJChJRYlMJCjr44N8JCwAfg+6MAAAAIng/9f/TCQs6+qDfCQwAH4WMdKJ4CaKVQD/14tU JDBKRYlUJDDr44N8JDQAfg+6MAAAAIng/9f/TCQ06+r2RCQeCHQHg3wkBAB/GYtEJGiAOAAP hbr8//+LRCQQg8RwXQdfXsO6IAAAAIng/9f/TCQE69GLRCQQg8RwXQdfXsOLwFFWV4nWidrG QxYg6EQBAADHQwQAAAAAgDgqdSSLFoPCBIkWi1L8iVMEhdJ9EInXimse99+AzQiJewSIax5A 6x+KEID6MHIYgPo5dxNrSwQKMdKKEIPqMAHRQIlLBOvhx0MI/////4A4LnVUx0MIAAAAAIpo AUCA/Sp1G4sWg8IEiRaLUvyJUwiF0n0Hx0MI/////0DrH4oQgPowchiA+jl3E2tLCAox0ooQ g+owAdFAiUsI6+GDewj/dATGQxYgihCNSAGA+k5yHw+GewAAAID6bHILdiuA+nd0Jl9eWcOA +mh0K19eWcOA+klyC3YmgPpMdD1fXlnDgPpGdENfXlnDiksegMkgQIhLHl9eWcOASx4Q6zmA eAE2dTWAeAI0dS+Kax+AzQGDwAOIax9fXlnDilMfgMoBiciIUx9fXlnDgEsegInIX15Zw4BL HkCJyF9eWcNTUWbHQh4AAIoYgPstdQaASh4I60KA+yN1BoBKHgHrN4D7K3UTimoegM0EiOuI ah6A4/2IWh7rH4D7IHUSinoe9scEdRKI+YDJAohKHusIgPswdRGIWhZAihiA+y11toBKHgjr 8llbw1FWVwaJ0YnficKOwTHAidYmih5ChNt0Bzn4dANA6+8HX15Zw1FWBoPsBInejsKJwzHJ g/7/dSFmJoszZoX2dEAx0ongZonyg8MC6I0rAACD+P904wHB699mJoM7AHQdOfF/GTHSieBm JosTg8MC6GkrAACD+P904AHB69w58X8EicjrAonwg8QEB15Zw41AAFFWV1WD7ASJ1YkcJLsQ AAAAie/owCsAAAaM2I7AKclJMcDyrvfRSQeLBCRIhcl0FYnujRwojRQpS4pK/0pIiEsBOfJ1 840UKIXAfAdIxgIwSuv1AywkxkUAAIPEBF1fXlnDi8BRVldVg+wEicGJ3okUJIXSfQv32o1I AcYALYkUJIN+CP91B8dGCAQAAAC7CgAAADHAicpmi0QkAug3KwAAiiGJy4TkdAiKUQFBhNJ1 +IN+CAB0MsYBLjHAi24IQYXtfiIx0maJVCQCizwka9cKiRQkilQkAoDCMIgRQItuCEE56Hze xgEA9kQkAYAPhGb///852XVBjUsBxgMxgDkwdQiKQQFBPDB0+IohgPwudRjGATBBiCGKUQFB gPowdQmKcQFBgP4wdPfGATBBxgEAg8QEXV9eWcOKUf9JgPoudQFJijGA/jl0Dojz/sOIGYPE BF1fXlnDxgEw65eNQAD/FXTNQADDkFNRUlZXVfZAHgh1MIB4FjB1KotQBItYIItIJCnai3Ao KcqLeCwp8otoMCn6i1g0Keop2oXSfgUB0YlIJF1fXlpZW8OQVlcGVYPsBI7CiceLUyiJzYXS fkUx0ongZiaLF4PHAuh8KQAAicGD+P90DTtDKH8iieZJg/n/dQiDeygAf9brGTHSidiKFv/V i0MoSEaJQyjr4MdDKAAAAACDxARdB19ew41AAFZXBlWD7BCJxonZjNiOwMdDIAAAAADHQyQA AAAAx0MoAAAAAMdDLAAAAADHQzAAAAAAifeKQxXHQzQAAAAAPGlyHg+GkgAAADx1cgt2JTx4 dCHpYAEAADxvdBjpVwEAADxYD4JPAQAAdgk8ZHRs6UQBAAD2QR8BdCCLGoPDBIkai0P8iQQk iyqDxQSJKotF/IlEJATpHgEAAPZBHiB0E4sCg8AEiQKLQPyJRCQI6QUBAACLKoPFBIkqi0X8 iUQkCPZBHhAPhO0AAAAxwGaLRCQIiUQkCOndAAAA9kEfAXQdiwKDwASJAotA/IkEJIsag8ME iRqLQ/yJRCQE6zP2QR4gdAyLKoPFBIkqi0X86x2LGoPDBIkai0P8inkeiUQkCPbHEHQLi0Qk BsH4EIlEJAiKWR8xwPbDAXQJ9kQkB4B0CesLg3wkCAB8BIXAdEKLQSCNWAGJWSDGBAYt9kEf AXQpixwki2wkBPfT99WJHCSNQwGJbCQEiQQkhcB1BY1FAesCieiJRCQE6y33XCQI6yeKQR6o BHQPi0EgjVgBiVkgxgQGK+sRqAJ0DYtBII1YAYlZIMYEBiCKQRW7CgAAADxkcmEPhgsCAAA8 b3I4D4bhAQAAPHNyIQ+G8AAAADx1D4IEBAAAD4bpAQAAPHgPhHwBAADp8QMAADxwD4SJAgAA 6eQDAAA8Zg+CnQAAAHZmPGcPhpMAAAA8aQ+EtQEAAOnFAwAAPEdyOA+GfAAAADxTciEPhpUA AAA8WA+CqQMAAA+GKQEAADxjD4TOAgAA6ZYDAAA8UA+ELgIAAOmJAwAAPEVyBHZE6w08Qw+E OAMAAOl0AwAA9kEeEHQvixqDwwSJGotD/IlEJAiJy4nCifDowPv//7v/////jNqJ8Oi++v// iUEo6VIDAACJy4nw6J38//+JyOie/P//jNiNfgGOwIzCifiDxBBdB19ew8YGAIpBHqiAdB2L KoPFCIkqi0X4ZotV/IXAdQVmhdJ0Lo7CicfrKKhAdBCLMoPGBIkyi0b8hcB0FusOixqDwwSJ GotD/IXAdAaM2onHjsKAeRVTdQj2QR4QdAjrFPZBHiB0DozCifiLWQjoSPr//+sMjMKJ+ItZ COgW+v//i1EIiUEohdIPjKQCAAA50A+OnAIAAIlRKIzCifiDxBBdB19ew/ZBHgF0OvZBHwF0 D4M8JAB1EIN8JAQAdCfrB4N8JAgAdB6LQSCNUAGJUSDGBAYwi0EgjVABiVEgjRQGikEViAK7 EAAAAIB5FW91GopRHrsIAAAA9sIBdA2LQSCNUAGJUSDGBAYwjNqLQSCOwgHwinEficf2xgF0 NoN5CAB1FYM8JAB1D4N8JAQAdQgmxgAAMcDrWYtRIIngAfLo3iUAAIB5FVh1OYnw6O0BAADr MIN5CAB1D4N8JAgAdQgmxgAAMcDrKYtRIItEJAgB8uicJgAAgHkVWHUHifDouwEAALv///// jMKJ+OgF+f//iUEoicKLQQg5wn0FKdCJQSSDeQj/D4WGAQAAicjo2/r//4zCifiDxBBdB19e w4N5BAB1FvZBHoB0CcdBBA0AAADrB8dBBAgAAACAYR75iwKDwASJAopZHoto/PbDgHQpg8AE iQK7BAAAAItA/InyJf//AADoHPn//7sIAAAAjVYFiejGRgQ66wm7CAAAAInyiejo/vj//4B5 FVB1B4nw6AkBAAC7/////4zCifjoU/j//4lBIIzCifiDxBBdB19ew4pZHsdBIAEAAAD2wyB0 ZYsag8MEiRpmi0P8MdJmicKNRCQM6PgjAACD+P8PhK8AAACKRCQMiy0g8sgBiAaF7Q+EmwAA ADHAikQkDIqAMfLIASQBJf8AAAAPhIIAAACKRCQNiEYB/0EgjMKJ+IPEEF0HX17DiwKDwASJ AopA/IgGjMKJ+IPEEF0HX17DiyqDxQSJKmaLVfyB4v//AACJ8Oh8IwAAicKD+P90D4lBIIzC ifiDxBBdB19ew8dBIAAAAACMwon4g8QQXQdfXsPHQQQAAAAAikEViAbHQSABAAAAjMKJ+IPE EF0HX17DU1KJwoA4AHQTMcCKAuhSJQAAiAKKWgFChNt17VpbwwAAAAAAU1FSVldVicGLQBD/ FSjNQACKYQ0x7fbEEA+EegAAAIjngOfvikEMiHkNqAIPhKIAAACLeQiLXwiF2w+ElAAAAItx BInfhfYPhIcAAACF7Q+FfwAAAInzifqLQRDo7yQAAInCg/j/dQ2KWQyAyyCJxYhZDOschcB1 GLgMAAAA6E0SAACKYQyAzCC9/////4hhDAHXKdZ1tus7i0EIg3gIAHQygGEM7/ZBDSB1KItB BIXAdBGJwrsBAAAA99qLQRDofCUAAIP4/3ULilkMgMsgicWIWQyLQQiLQAjHQQQAAAAAiQGF 7XUYi0EI9kAQAXQPi0EQ6MglAACD+P91AonFi0EQ/xUszUAAiehdX15aWVvDAAAAAAAAAAAA AAAAAAAABoM4AHQEHgf/EAfDU1FSVga+eNVAAIjGuELVQACJ84jyOcZ2GoA4AnQLimgBOOpy BInDiOqDwAY9eNVAAHLmgft41UAAdA2NQwLosf///8YDAuvEB15aWVvDi8BTUVYGvpzVQACI wYjWuHjVQACJ84jKOcZ2GoA4AnQLimgBOOp3BInDiOqDwAY9nNVAAHLmgfuc1UAAdBI6cwFy CI1DAuhb////xgMC678HXllbwwBTUonD/xVQzUAAixVwzUAAjQQao3DNQADo8yYAAP8VVM1A AInQWlvDAAAAAAAAAABRidlQiNbB4giI8sHiCIjy6CsnAABYWcMAAAAAAAAAAFJVieWLFejM QACDwgOA4vzouycAADnCcxZS6AEoAACh6MxAAIPAAyT8KcSJ4OsCMcCLFejMQAAB0KPszEAA 6B0oAACLFVTzyAGhUPPIAejN3f//6BwoAACJ7F1awwBVieVRVleD7ECJxonZMcCKQxWJRcCK QRWLWwgkXyX/AAAAg/hHdSOF23UFuwEAAADHRbwEAAAAi33AuAEAAACD7wKJRbiJfcDrH4P4 RXUNvwEAAACJfbyJfbjrDb8CAAAAMcCJfbyJRbj2QR4BdASATbwQiwKDwAiJAonCi0D4iUXs i0L8jVXgiUXwjUXs3QDbOoP7/3UFuwYAAACNReAx0oldtIlVxI1eAY1VtOjtKAAAi0XQiUEo i0XUiUEsi0XYiUEwi0XciUE0g33IAH0Pi0EgjVABiVEgxgQGLespimEe9sQEdA+LQSCNUAGJ USDGBAYr6xL2xAJ0DYtBII1QAYlRIMYEBiCM2onwjWX0X15ZXcNTidMx0ujLMwAA3Rtbw1K6 AQAAALjow0AA6OwLAABawwAAAAAAAAAAAAAAAAAAwwAAAAAAAAAAAAAAAAAAAGZQm9vj2ejZ 7t752cDZ4N7Z3+CwAp50ArADm9vj2SwkZocEJGZYwwAAAAAAAAAAAFNRUonD9sMDdSy5ZAAA ADHS9/GF0nQJuAEAAABaWVvDuZABAACJ2PfxhdJ1CbgBAAAAWllbwzHAWllbw1NRVldVg+wk icOLSCCFyQ+FwwAAAI2CbAcAAOik////hcB0E4tDEIs8RebJQACLBEXkyUAA6xGLQxCLPEXM yUAAiwRFyslAAMH/EMH4ECnHMe2+AQAAAIksJIlsJASJbCQIiXQkDItDEIlEJBCJVCQUieCJ bCQg6Hjl//+LRCQYi1MYKcKDwge5BwAAAInQwfof9/mLSwyD+QV1ICnxjTTNAAAAAI1CASnO AfA5+H4Ii0MMg+gC6wiJyOsEicgp8ItcJBwB04nCweADKdAB2IPEJF1fXllbw4P5AXUOi0Ac KciDxCRdX15ZW8OLQByDxCRdX15ZW8OQUVZXidGDeCAAdRuDeSAAdRWLcBCLeRA5/n4JuAEA AABfXlnDfB+J2ujb/v//icaJ2onI6ND+//85xn4JuAEAAABfXlnDMcBfXlnDkFNRUlZXVYPs HInBodjOQACKEDH2hNIPhJACAAC6rM1AALiIzUAAi1kU6Ib///+JRCQUhcB0DL+szUAAu4jN QADrCr+IzUAAu6zNQACLQRQFbAcAAItpEOgk/v//jRRtAAAAAIXAdA6LgubJQACLkuTJQADr DIuCzMlAAIuSyslAAMH4EMH6ECnQiUQkEDHSi0cgiVQkBIXAD4WxAAAAi1cQOdV+Cr4BAAAA 6dUAAAAPhc8AAACLURiLRxiDwgcpwsdEJBgHAAAAidDB+h/3fCQYi0EMKdCLURiJRCQIg8IG K1cYx0QkGAcAAACJ0MH6H/d8JBiLQQxIKdCLVwyJRCQMg/oFdSWLRCQQi1QkCIPoBznQD41r AAAAi1QkDL4BAAAAOdB8XulgAAAAjUL/jRTFAAAAACnCjUIBO0QkCH9Ei1QkDL4BAAAAOdB+ N+s8i1ccg/gBdR+LQRQFbAcAAOgc/f//hcB0DaHOyUAAwfgQOcJ+AUJKi0EcOcJ/B74BAAAA dAeDfCQEAHQWifqJyOgoAQAAhcAPlMCJxoHm/wAAAIX2dROLRCQUhcAPhPsAAACJxun0AAAA MdKLeyCJFCSF/w+FjQAAAIt7EDn9fgcx9um3AAAAD4WxAAAAi1EYi3sYg8IHKfq/BwAAAInQ wfof9/+LaQwp1YtRGItDGIPCBinCidDB+h/3/4tBDEgp0ItTDDH2g/oFdR2LVCQQKfo51X8K vgEAAADpXwAAADnQf1vpXAAAAI16/40U/QAAAAAp+kI51X0HvgEAAADrPjnQfTrrPotrIIt7 HIP9AXUgi0EUBWwHAADoE/z//4XAdA6hzslAAMH4EDnHfgIB70+LQRw5x38GMfY5x3QGgzwk AHQLidqJyOggAAAAicaLbCQUhe10BonoKfCJxonwiXEgg8QcXV9eWllbw5BTUVZVi0gIi3II Mds58XwUdReLSASLagQ56XwIdQuLADsCfQW7AQAAAInYXV5ZW8MAAAAAVldViceJ1onYgfrA qAAAcxyF234YjbKAUQEAKd4x0ruAUQEAifD38wHHT+sPu4BRAQApxjHSifD38wHHu4BRAQCJ 8DHS9/OJ074QDgAAidgx0vf2MdKJQQiJ2Pf2idO+PAAAAInYMdL39jHSiUEEidj39rtuAQAA ifiJETHS9/OJwsHgAwHQweADAdCJw8HgAgHYifspw4XSdgiNQv/B6AIpw42CbAcAAOji+v// BW0BAAA5w3IHKcNCOcNz+b7MyUAAiVEUjYJsBwAAiVkc6Lz6//+FwHQFvubJQAC9HwAAAInY MdL39YsURsH6EDnTcgFAiUEQD78ERinDMdJDjUcBiVkMuwcAAAD384nIiVEYXV9ew41AAFNR x0IgAAAAAInRiwAx24nCuN9jAADozf7//1lbw4vAU1KJw/8VJM1AAI1QFInY6Mz///9aW8MA UooV6M5AAKHozkAAgOL8g+ABiBXozkAAWsOLwFKKFejOQACAygGh6M5AAIgV6M5AAIjWgOb9 g+ABiDXozkAAWsONQABTUVJWgeysAAAAiiXozkAA9sQBdAn2xAIPhc0AAACKNejOQACJ4IDO AlCINejOQAAu/xWwsUAAg/gBD4KrAAAAdjWD+AIPhaAAAADHBeDOQAABAAAAi5QkqAAAAInQ weAEKdDB4AKJw6PkzkAA99uJHeTOQADrCDHJiQ3gzkAAi1QkVIs0JAHyidDB4AQp0LuAAAAA weACjVQkBKPczkAAuNDNQADooi0AAIP4/3UKMP+IPdDNQADrCDDbiB1QzkAAu4AAAACNVCRY uFHOQADoeC0AAIP4/3UKMO2ILVHOQADrCDDJiA3RzkAAgcSsAAAAXlpZW8OLwLgMxEAA6Aou AACFwA+E9v7//4vAU1FSVldVg+wEizXkzkAAMdK73M5AAIkV4M5AALrQzUAA6FYAAACJwYog hOQPhVACAACLNeTOQACIJVHOQACJNeTOQACDxARdX15aWVvDi8BTUYnRihgx0oD7MHIZgDg5 dxRr0gox24oYQAHaijiD6jCA/zBz54kRWVvDkFFWV1WD7BSJxYkUJIlcJBCAODp1AUWJ7opV AITSdByA+ix0F4D6LXQSgPordA2A+jByBYD6OXYDRevdiesp84H7gAAAAH4Fu4AAAACLPCSJ 2QaM2I7AV4nIwekC8qWKyIDhA/KkXwcB+41FATH2xgMAgPotdQe+AQAAAOsFgPordQKJxYpF ADwwD4J9AAAAPDkPh3UAAACNVCQMMf+J6Il8JASJfCQIiXwkDOgk////ihCJxYD6OnUfjVQk CEDoEf///4owicWA/jp1DI1UJARA6P7+//+JxYtcJAyJ2MHgBCnYi1QkCMHgAgHCidDB4AQp 0ItUJATB4AIBwotEJBCJEIX2dAaJ0/fbiRiJ6IPEFF1fXlnDi8BTUVaD7BCJw4nRiiC+//// /4D8SnUHvgEAAAAB84A7TXUDQzH2jVQkDInYiXEg6In+//+JwonDhfZ0CYtEJAyJQRzrRYtE JAxIiUEQgDoudTGNQgGNVCQM6F/+//+JxonDi0QkDIlBDIA+LnUVjVQkDI1GAehD/v//icOL RCQMiUEYx0EcAAAAALoCAAAAMfaJVCQIiTQkiXQkBIA7L3UyjVQkCI1DAegQ/v//ijCJw4D+ OnUdjVQkBEDo/f3//4oQicOA+jp1ConiQOjs/f//icOLBCSJAYtEJASJQQSLRCQIiUEIidiD xBBeWVvDuwEAAAC6Uc5AAKHczkAAiR3gzkAALRAOAACJ44kEJInI6NT9//+LDCSLNdzOQACJ winOihiJNeTOQACA+yx1DbqIzUAAQOjH/v//icKKOos15M5AAID/LHVnjUIBuqzNQADoq/7/ /4s15M5AALsQDgAAifKJ8MH6H/f7iz20zUAAuzwAAACJ8inHifDB+h/3+4nCwfof9/uLLbDN QAAp1YnyifDB+h/3+6GszUAAiT20zUAAKdCJLbDNQACjrM1AAIk15M5AAIPEBF1fXlpZW8MA AAAAAAAAAAAAAAAAU1FSVlcGD6APqFWD7ASJx4XAdAWD+NR2BzHA6b4AAACNaAuD5fiD/RBz Bb0QAAAA/xVAzUAAMOQx0ogkJDst9M5AAHYMiw3wzkAAhcl1EOsCMcmJDfTOQACLDezOQACF yXQ8i3EUiQ3wzkAAOf5yHLjszkAAjNqB4v//AACJy4n46HoqAACJwoXAdUI7NfTOQAB2Bok1 9M5AAItJCOvAgDwkAHULifjoaC0AAIXAdRWJ+OjJLQAAhcB0EzDJiAwk6Xj////GBCQB6W// //8w7YgtYPPIAf8VSM1AAInQg8QEXQ+pD6EHX15aWVvDAAAAU1FWV4PsBInHidYx0usBQonD igtAhMl19moAjUQkBFBSofjOQABXi1AIUi7/FRSyQACJ8Okj4f//jUAAU1GJw4nRjNroVy0A AIXAdQmJyonY6Kr///9ZW8MAAAAAAAAAU1FSVlf/FTjNQACLNUTkyAGF9nQZi04Ei3kMiwaB 5wNAAACjROTIAWaDzwPrTbmsykAAgfm0zEAAcyj2QQwDdRe4HQAAAOhl/v//icaFwHRYvwMA AADrI4PBGoH5tMxAAHLYuDcAAAC/A0AAAOg+/v//icaFwHQxjUgduxoAAACJyDHS6Lfx//+J eQyJTgShQOTIAYlxCIk1QOTIAYkG/xU8zUAAicjrErgFAAAA6F0CAAD/FTzNQAAxwF9eWllb w41AAFNRUonDukDkyAGLAoXAdCWLSAQ5y3QEicLr74pJDIDJA4hLDIsYiRqLFUTkyAGjROTI AYkQWllbw41AAFKDPUTkyAEAdBahROTIAYsQ6OoQAACJFUTkyAGF0nXqWsMAAAAAAAAAAAAA AAAAAFNSicP/FTjNQAChQOTIAYXAdQ7/FTzNQAC4/////1pbwztYBHQIiwCFwHTn6/P/FTzN QAC6AQAAAInY6AQAAABaW8OQU4nD6MgAAACJwonY6D////+J0FvDjUAAg8Awg/g5fgODwCfD U1FWV1WD7ASJxYkUJOjGKwAAicPB6xCJ7wnD6NgsAACJxleKBogHPAB0EIpGAYPGAohHAYPH AjwAdehfBozYjsApyUkxwPKu99FJB400KY1WBMYGdInxidiD4A9K6JL////B6wSIQgE5ynXr iwQkwfgEg+APxkYFX+h2////iEYGiwQkg+AP6Gj////GRgguxkYJdMZGCm3GRgtwxkYMAIhG B4PEBF1fXllbw1NRVldVgewUAQAAicGJ14N4DAB1Crj/////6ZcAAACKYA0x9vbEEHQJicjo zO3//4nGi0EQ/xUozUAAicjo2iwAAIP4/3QOicKLaRAx24no6OcTAACF/3QKi0EQ6BstAAAJ xvZBDAh0FYtBCItACOhIDwAAi0EIx0AIAAAAAPZBDQh0GotRCIpSFInggeL/AAAA6L3+//+J 4OhuLQAAi0EQ/xUszUAAhf90CYtBEP8VNM1AAInwgcQUAQAAXV9eWVvDAAAAAAAAAAAAAFKJ wv8VJM1AAIlQBFrDi8C4DQAAAFKJwv8VJM1AAIlQBFrDkLgOAAAAUonC/xUkzUAAiVAEWsOQ uAkAAADovv///7j/////w1KJwv8VJM1AAIlQCFrDAAAAAAAAAAAAAFNRUosN+M5AAKH8zkAA OwVs0kAAcwQxwOsvhcB+Josd/M5AAInKMcDB4wKDPAIAdQwxwIkN+M5AAFpZW8ODwAQ52Hzn uAEAAACJDfjOQABaWVvDkFNRUlaJxv8VWM1AAIsd/M5AADHShdt+LY0MnQAAAAAxwIsd+M5A AAHDgzsAdQ+JM/8VXM1AAInQXlpZW8ODwARCOch83IsV/M5AAEKh+M5AAMHiAujDLAAAixX8 zkAAo/jOQACNSgGJNJCJDfzOQAD/FVzNQACh/M5AAEheWllbw41AAFNRVldVg+wEiz34zkAA iQQkidaF0g+MqgAAAP8VWM1AAIP6AXIJdhOD+gJ0FushhdJ1HYsMJFFq9usOixwkU2r16waL FCRSavQu/xXksUAAiy38zkAAjQy1AAAAAIs9+M5AADnufQkB+YsEJIkB60mNUQSJ+OgeLAAA ixX8zkAAicc58n0YjQSVAAAAAIn7xwQDAAAAAIPABEI5yHzxjQS1AAAAAIsUJEYB+Ik9+M5A AIk1/M5AAIkQ/xVczUAAiz34zkAAiz34zkAAg8QEXV9eWVvDUonC/xVYzUAAhdJ+HDsV/M5A AH0UjQSVAAAAAIsV+M5AAMcEAgAAAAD/FVzNQABaw41AAFFSavYu/xWssUAAicKFwHQFg/j/ dQXoRQAAAOhU/v//avUu/xWssUAAicKFwHQFg/j/dQXoJwAAAOg2/v//avQu/xWssUAAicKF wHQFg/j/dQXoCQAAAOgY/v//WlnDkFFSagBqAGoAagAu/xU4sUAAicKFwHUNixUAz0AAQokV AM9AAInQWlnDjUAAU1KLFfjOQACF0nQPidAx2+gPDAAAiR34zkAAWlvDAAAAAAAAw41AAFNR UlaLFeDxyAGD+kB9Ho0ElQAAAAC7sOvIASnQjXIBweADiTXg8cgBAcPrZ7oYAAAAuAEAAADo ICsAAInDhcB1D7oBAAAAuBDEQADoh/n//4sV5PHIAUKh6PHIAcHiAuh3KgAAo+jxyAGFwHUP ugEAAAC4NMRAAOhb+f//oeTxyAGLFejxyAGNSAGJHIKJDeTxyAFTLv8VuLFAAInYXlpZW8OQ U1FSVosV4PHIATH2hdJ+G7uw68gBU0Yu/xVEsUAAiw3g8cgBg8MYOc586l5aWVvDU1FSVldV ixXk8cgBMfaF0n4tMduh6PHIAYsMA1Eu/xVEsUAAoejxyAGLBANG6PYKAACLPeTxyAGDwwQ5 /nzViy3o8cgBhe10B4no6NgKAABdX15aWVvDkMdACAAAAADHQAwAAAAAw5C4gOrIAemSAAAA i8C4gOrIAenqAAAAi8CD4A/B4AQFoOrIAel0AAAAg+APweAEBaDqyAHpyAAAAOlH/P//jUAA UonCg+APweAEBaDqyAHonf///4nQ6IL9//9aw7ig68gB6zmQuKDryAHpkgAAAIvAuJDqyAHr JZC4kOrIAel+AAAAi8C4wPHIAesRkLjA8cgB622QuNDxyAGNQABTUVJWVYnDLv8VeLFAAItT CInGOdB0O4N7BAB1KLiw8cgB6Nn///+DewQAdQ7o9v3//8dDBAEAAACJA7iw8cgB6B8AAACL K1Uu/xVMsUAAiXMI/0MMXV5aWVvDkLjQ8cgBjUAAU1FSVotQDIXSdheNWv+JWAyF23UNizBW iVgILv8VvLFAAF5aWVvDkFNRUlYu/xWQsUAAixUgzUAAUonGLv8V9LFAAInDhcB1B+jSDgAA 6wuAeFMAdAfoAQ8AAInDVi7/FeCxQACJ2F5aWVvDU1KJwoXAdSa4AQAAAIsVcM1AAOicKAAA icKFwHQQix1wzUAAxkBSAYmY8AAAAInQ6J8oAACJ0Fpbw4vAU1FSix0gzUAAg/v/dTsu/xXs sUAAZosVCc1AAInDZoH6AIByJIA9B81AAARzG4P7/3QWg/sCdxGJHSDNQAAu/xXssUAAicPr 5YP7/w+VwCX/AAAAiR0gzUAAWllbw4vAU1FSgz0gzUAA/3UGMcBaWVvD6E3///+Jw4XAdDKJ 2ouA2gAAAOj+DgAAhcB1DYnY6IMIAAAxwFpZW8NTix0gzUAAUy7/FfixQAC4AQAAAFpZW8OQ U1FSVleJw4sVIM1AAIP6/3Q9Ui7/FfSxQACFwHQxi7DeAAAAi4DaAAAA6AkPAABqAIs9IM1A AFcu/xX4sUAAhfZ0DIXbdAhWLv8VNLFAAF9eWllbw41AALgBAAAA6J7///+LwFFSixUgzUAA g/r/dBJSLv8V8LFAAMcFIM1AAP////9aWcOLwFNRUlZXVbroRkAAu/hGQAC5CEdAAL4QR0AA v9BGQAC93EZAALhUR0AAiRUozUAAiR0szUAAiQ0wzUAAiTU0zUAAiT04zUAAiS08zUAAo1DN QAC6XEdAALtsR0AAudBHQAC+wEZAAL8sR0AAvUBHQAC4NEdAAIkVVM1AAIkdBM9AAIkNCM9A AIk1DM9AAIk9QM1AAIktRM1AAKNIzUAAukhHQAC7AQAAALlkR0AAvshHQAC/hElAAIkVTM1A AOge+///ixVc5MgBo7DxyAGJHbTxyAGJDVjNQACJNVzNQACLgtoAAACJPWDNQADoXA0AAIst XOTIAVWhIM1AAFAu/xX4sUAAxwUkzUAA+EdAAF1fXlpZW8ONQABTUriA6sgBuqDqyAH/FQzP QACNmgABAACJ0IPCEP8VDM9AADnadfG40PHIAf8VDM9AAOhl+///6DgmAADoQyYAALig68gB /xUMz0AAuJDqyAH/FQzPQAC4wPHIAf8VDM9AALiw8cgB/xUMz0AA6Pr6///oTf7//1pbwwAA AAAAAAAAAABRVldVicWJ1onfLv8VtLFAAMHoECX//wAAZj0AgHMPU1ZVLv8VnLFAAF1fXlnD uAgCAADouPL//4nDhcB0UmgIAgAAUFUu/xWYsUAAhcB1DonY6OkFAAAxwF1fXlnDV1Zq/1Nq AWoBLv8VyLFAAInFidjoyQUAAIXtdQcxwF1fXlnDifBmx0R+/gAA6IAnAABdX15ZwwAAAAAA AAAAAAAAUVJWV4nGiccGjNiOwCnJSTHA8q730UkHQYnI6DDy//+JwoXAdBiJxwaM2I7AV4nI wekC8qWKyIDhA/KkXweJ0F9eWlnDAAAAAAAAAAAAAABTUVKJwugWJwAAQI0cRQAAAACJ2Ojn 8f//icGFwHQF6BwnAACJyFpZW8MAAAAAAABTUVaD7CCJw4nWahyNRCQEUI1EJCRQLv8VCLJA AIsUJANUJAxmiw0JzUAAi0QkBGaB+QCAcwcFADAAAOsXchCAPQfNQAAEcwcFACABAOsFBQAw AQCF23QCiQOF9nQCiRaDxCBeWVvDAAAAAAAAAAAAU1FSaFjEQAAu/xXAsUAAMduFwHQXaGTE QABQLv8VqLFAAInChcB0BP/SicOF2w+VwCX/AAAAWllbw4vAUVZXid+KCDHbhMl0CIpoAUCE 7XX4jXAJigqICITJdBKA+TB1CIB6AXh1AonzRkBC6+aF23QcifiF/3QWicKD4g9LipIc0EAA wegEiFMBhcB16l9eWcOLwFNXgewEAQAAi5wkEAEAAIsLi1sE6FP///+FwHUK6L4mAACD+P91 BzHA6YwBAAAw5IgkJIsBPZAAAMByTQ+GwgAAAD2TAADAcjMPhqsAAAA9lgAAwHIWD4bsAAAA Pf0AAMAPhPYAAADp+AAAAD2UAADAD4TfAAAA6egAAAA9kQAAwA+GaQAAAOsvPY0AAMByC3ZA PY4AAMB2Q+tLPQUAAMAPgr4AAAB2Zj0dAADAD4ScAAAA6awAAAD2QyECdAq6dMRAAOmvAAAA usjEQADppQAAALocxUAA6ZsAAAC6cMVAAOmRAAAAusTFQADphwAAALoYxkAA6X0AAAC6ZMZA AOlzAAAAurTGQADpaQAAALoMx0AAieCLWQzoi/7//7pAx0AAieCLWRjofP7//4N5FAB1C7po x0AAieAx2+s/unDHQACJ4DHb6zS6fMdAAOsourzHQADrIbr4x0AA6xq6PMhAAOsTunjIQACJ 4IsZ6Db+//+6rMhAAIngi1kM6Cf+//9qAI2EJAQBAABQjXwkCAaM2I7AKclJMcDyrvfRSQdR jUQkDFCh+M5AAItYCFMu/xUUskAAuAEAAACBxAQBAABfW8IEAI1AAKJPQABWUEAArE9AAMpP QAC2T0AAiE9AAMBPQABTVleD7AiLdCQYi3wkIPZGBAYPhaEBAACLBgVz//8/g/gGD4cfAQAA Lv8khTxPQAD2RyECdAq7igAAAOnKAAAAu4sAAADpwAAAALuCAAAA6bYAAAC7hgAAAOmsAAAA u4QAAADpogAAALuFAAAA6ZgAAACLRyhmixC7gQAAAGaB+tn6dQq7iAAAAOl8AAAAZoH62fF1 CruOAAAA6WsAAAB1CruPAAAA6V8AAACKMID+23QFgP7fdRCKUAGA4jCA+hB1BbuNAAAA9gAB dTmKQAEkMDwwdTCLRyAl//8AAMHgAjHJZsHoDYtXJGaJwYHi//8AAAHJ0+qD4gGD+gF1B7uD AAAA6wWD+/90PcYFBPLIAQHo8iMAAInY6L8lAACD+P8PhHQAAACAPQTyyAEAD4RnAAAAZotf IDDbgOd/McBmiV8g6XcAAACDPRjQQAAAdEq7AQAAAInZidiLFv8VFNBAAIXAdC+D+AF0MIP4 AnQrg/gDdCaJ2IgNBPLIAf8VGNBAAIA9BPLIAQB0CTHAg8QIX15bw0OD+wx+vYngUIl0JASJ fCQILv8V/LFAAIXAdAlq/y7/FVCxQAC4AQAAAIPECF9eW8ONQABRUonC/xUkzUAAiVBUMcBk iwCJwv8VJM1AAItAVIkQ/xUkzUAAi0BUMdLHQARYT0AA/xUkzUAAi0BUZIkCaHRNQAAu/xXo sUAAWlnDUv8VJM1AAItAVIXAdAeLADHSZIkC/xUkzUAAx0BUAAAAAFrDU1FSVonGhcAPhPMA AAD/FUDNQACLDRDyyAGFyXRAOfF3DIsBAcg58A+HjQAAAInKi0kEhcl0EDnxdwyLAQHIOfAP h3QAAACLSgiFyXQQOfF3DIsBAcg58A+HXQAAAIsN8M5AAIXJdDQ58XcIiwEByDnwd0eJyotJ BIXJdAw58XcIiwEByDnwdzKLSgiFyXQMOfF3CIsBAcg58Hcfiw3szkAAhcl0VTnxdwiLAQHI OfB3CYtJCIXJde3rQLjszkAAjNqB4v//AACJy4nw6I8XAACLFfDOQACJDRDyyAE50XMSix30 zkAAi0EUOdh2BaP0zkAAMOSIJWDzyAH/FUjNQABeWllbwwAAAAAAAAAAAABTUVKJwvZADSB1 IotCEOh9JQAAhcB0FopaDYDLIIhaDfbDB3UIiNmAyQKISg1aWVvDAAAAAAAAAAAAAAAAAAAA UYPsBIkUJIXAdCpqAGoAagJQagGNRCQUUGgAAgAAixU400AAUi7/FRCyQACFwHUFuP////+D xARZwwAAAAAAAFFWV1WD7CiJ1YnfidYw0o1MJAGIFCSNXCQkiXwkJDHS9zOJA4qCMNBAAIgB i0QkJEGFwHXhRopB/0mIRv+EwHX0ieiDxChdX15Zw5BRidGD+wp1CoXAfQb32MYCLULonf// /4nIWcMAAAAAAAAAAABRVlcGVYPsRInGiVQkPI18JDSNbCQBjNiJVCRAjsAx0jDkpaWJXCQk iVQkKIgkJI18JCyNdCQ0jVwkJI1EJDSLUASLAItLBIsb6HsmAACJVgSJBolPBIkfi0QkLIqA WNBAAIhFAItcJDRFhdt1wYN8JDgAdbqLXCRAikX/TY1zAYgDiXQkQITAdeuLRCQ8g8REXQdf XlnDkFFWVwaD7AiJxonRjNiJ547ApaWD+wp1LfZEJAeAdCbGAi2LFCSLdCQE99L31okUJIl0 JASLPCSNUQFHiTwkdQFGiXQkBIng6Bz///+JyIPECAdfXlnDAABRVldVg+woidWJ34nWMNKN TCQBiBQkjVwkJIl8JCQx0vcziROLXCQkQYqbgNBAAIhZ/4XAdeBGikH/SYhG/4TAdfSJ6IPE KF1fXlnDUYnRg/sKdQqFwH0G99jGAi1C6J3///+JyFnDAAAAAAAAAAAAioAx8sgBJAEl/wAA AMOLwDHA6ZkjAAAAAAAAAAAAAACD+GF8CIP4en8Dg+ggwwAAUVZXVYPsBInGideFwHwIOwVs 0kAAdhS4BAAAAOhf7f//uP/////puAAAAIst+M5AAItstQD/FSjNQACJ8OiOGgAAqIB0KGoC agBqAFUu/xXcsUAAg/j/dRWJ8P8VLM1AAOjeJQAAg8QEXV9eWcODPXzTQAAAdCiJ8P8VUNNA AIXAdByJ+v8VfNNAAInDifD/FSzNQACJ2IPEBF1fXlnDagCNRCQEUFNXVS7/FRSyQACFwHUV ifD/FSzNQADogyUAAIPEBF1fXlnDOxwkdAq4DAAAAOio7P//ifD/FSzNQACLBCSDxARdX15Z wwAAAAAAUVaJxonRhcB8CDsFbNJAAHYSuAQAAADodOz//7j/////XlnD/xUozUAAifDorxkA AIXJfhCogHUMgMyAicKJ8OjzGQAAU2oAixX4zkAAUYsEslAu/xXcsUAAicGJ8P8VLM1AAIP5 /3UF6OYkAACJyF5ZwwAAAAAAAAAAAAAAAABTUVJWicMx9oXAfAg7BWzSQAB2FLgEAAAA6PLr //+4/////15aWVvD/xUozUAAofjOQACLBJhQLv8VWLFAAIXAdQrojSQAAL7/////idj/FSzN QACJ8F5aWVvDAAAAAABTUVIxwOje8f//MduFwHQQixUgzUAAUi7/FfSxQACJw4XbdQ+6AQAA ALjYyEAA6Eno//+J2FpZW8ONQABTUVJWV1X/FVDNQAAu/xV4sUAAix1A88gBhdt0CztDBHQG ixuF23X1g3sMAHQlixVwzUAAi0MI6AgZAACJxYXAdV66AQAAALgAyUAA6O/n///rTbgBAAAA ixVwzUAA6GEZAACJxYXAdQ+6AQAAALgoyUAA6Mjn//+LcwiJ74uO8AAAAAaM2I7AV4nIwekC 8qWKyIDhA/KkXwfHQwwBAAAAiWsIoXDNQABVxkVSAYs1IM1AAMZFUwBWiYXwAAAALv8V+LFA AP8VVM1AAInoXV9eWllbw1NRVonGidP/FVDNQAC5AQAAALoQAAAAicjo0hgAAInChcB0L4nY 6GkjAACFwHQJidDoWvn//+sbiVoIiXIEikNSiUIMoUDzyAGJFUDzyAGJAusCMcn/FVTNQACJ yF5ZW8ONQABTUVKJw/8VUM1AAIsVQPPIAblA88gBhdJ0KDtaBHUbg3oMAHQIi0II6P/4//+L AokBidDo9Pj//+sIidGLEoXSddj/FVTNQABaWVvDUv8VUM1AAKFA88gBhcB0DYtQCMZCUwGL AIXAdfP/FVTNQABaw41AAFNRUosVQPPIAYXSdB6LSgyLGoXJdAiLQgjomvj//4nQ6JP4//+J 2oXbdeJaWVvDAAAAAAAAAAAACcl0LDgQqAN0CYgQQMHKCEl181HB6QLoGwAAAFmD4QN0DogQ SXQJiHABSXQDiFACw5CQkJCQkAnJdGeoH3QIiRCNQARJdfRRwekCdDpJdCmJEIlQBEmJUAiJ UAx0GDhQIIlQEIlQFEmJUBiJUByNQCB13Y1A8I1AEIkQiVAEiVAIiVAMjUAQWYPhA3QViRCN QARJdA2JEI1ABEl0BYkQjUAEwwAAAAAAAAAAAAAAAABS/xUkzUAAicKJ4CsCWsMAw4dEJATo BwAAAItEJATCBABQOeBzGing99hWUP8VJM1AAInGWDsGXnYG6BEAAADDWLio0EAAugEAAADo XOX//1BTi0QkDLv8////iRwcgesAEAAALQAQAAB/8FtYwgQAUOja////UI1EJAgrBCSUi0AE /+AAAAAAAAAAAAAAAADDAAAAAAAAAAAAAAAAAAAAw41AAFNSicP/FbzQQACAPVDkyAEAdQ+6 /wAAALgQAAAA6MbW//+J2OgDAAAAWlvDUonC/xW80EAA/xXA0EAAgz2Q00AAAHQG/xWQ00AA idDp18X//1WJ5VNRVleD7AyJx4nWgfoAIAAAfAW+ACAAALnE0EAA6y5m98YBAHQiZotBCGaJ ReyLQQSJReiNVeSLAYn7iUXkifjbKNsq3snbO9H+g8EKhfZ/zo1l8F9eWVtdw1WJ5VNRVleD 7AyJwYXSdOdqAJvZPCRYicaAzAMl//8AALv/PwAAUNksJFgx/2aJXey7AAAAgIl95Ild6IXS fRuNReT32uhZ////icuNVeSJyNso2yre+ds76xeNReToQP///4nLjVXkicjbKNsq3snbOzHA ZonwUNksJFjpdv///1NRicGJ04H6ABAAAH4SugAQAADoZ////4HrABAAAOsYgfoA8P//fRC6 APD//+hN////gcMAEAAAidqJyOg+////WVvDi8CAXEAAvFxAAI5cQACsXEAAgFxAAFWJ5VFW V4HsjAAAAInHidaJXeBqAJvZPCRYiUXsgMwDJf//AABQ2SwkWMdCFAAAAABmi0cIZolF1ItH BIlF0IsHiUXM9kXVgHQHx0IU/////4Bl1X/HRhwAAAAAx0YgAAAAAMdGJAAAAACNRczHRigA AAAAMf/HRhgAAAAA6GwfAACD+AQPh0kBAAAu/ySF2VtAAMdGFAAAAAAxyekzAQAAi0XgxgBu xkABYcZAAm7GQAMAx0YcAwAAAOlwAwAAi0XgxgBpxkABbsZAAmbr4DHJZotN1IHp/j8AAGnR l3UAALughgEAidDB+h/3+41I/IXJD4TcAAAAfQ/32YPBA4Dh/PfZ6b8AAACLVdRmgfoZQHKH dQ2BfdAAILy+D4J4////i0XUZj00QHIgD4WUAAAAi13Qgfu/yRuOcg8PhYMAAACBfcwAAAAE c3q7GUAAAI1VtI1FzGaJXby7ACC8vjHJiV24jV2oiU202yjbKt752zuNRaiNVaiNXbTbKFBQ m9k8JFhQgMwMUNksJFjbXCQE2SwkWFiJx1DbBCRY2zqNVbSNRajbKNsq3snbO41dzI1VtI1F zLkIAAAA2yjbKt7p2zvrD4Dh/InKjUXM99ro0v3///ZGCAJ0FosGAciDwAqJReiLRgSFwH4N AUXo6wiLBoPAB4lF6IpeCLgPAAAA9sMgdAW4FAAAAPZGCEB0AgHAi13og8AEOdh9A4lF6MaF aP///zAw5Iilaf///42Faf///4lF2DHAiUXci1XohdIPjugAAACNWviJXeiF/3VyZvdF1P9/ D4TSAAAAjUXM2yhQUJvZPCRYUIDMDFDZLCRY21wkBNksJFhYiceF235CjVXAjV3MUNsEJFjb Oo1VwI1FzNso2yre6ds7uBlAAAC7ACC8vo1VwGaJRciJXcQxwI1dzIlFwI1FzNso2yreyds7 i13YifiLVdxRUugCAAAA60W5ECcAACnSOciScgOS9/FS6AEAAABYuWQAAAAp0jnIknIEkmb3 8VLoAQAAAFixCjjIhsRyBIbE9vGAxDAEMIgDQ4gjQ8NaWbAAiAOJXdiDwggx/4lV3OkN//// i0XcjZVp////g8EHgDowdQVISULr9op+CIs+9scCdAoDTgSNWQEB3+sV9scBdBCLXgSF234D R+sCAd9BK04Ehf8PjGUAAAA5x34Cice7DwAAAPZGCCB0BbsUAAAA9kYIQHQCAds5334DjXsB xkXwMDn4fgqAPDo1cgTGRfA5iX3kjQQXi13kS0iJXeSKGIp98Dj7dQNP6+yA/zl1Bojf/seI OIN95AB9A0pHQYX/fxi/AQAAALAwMcmIhWj///+NlWj///+JThSKZgj2xAJ1FPbEBHQeg/n8 fAQ7DnwG9kYICHQPi13gU4nwifvoIgAAAOsNi0XgUIn7ifDoMQIAADHAZotF7FDZLCRYjWX0 X15ZXcNVieVWV4PsBFBSUYsQQYlV9IlN6IpgCDHS9sQEdCU7XfR9DIt18PZGCBB1A4ld9It1 6It99Cn3iX30hf99BTHJiU30i3XohfYPj6QAAACLdfD2RggIdSaLfQiJ1gH+i0X0QsYGMIXA fwmLdfD2RggQdAuLTQiJ1wHPQsYHLot18IlWHIt16It99PfeOf5+Con4iX3o99iJReiLffCL deiJdxiJ8Yl3IPfZiU8gi330AfeJffQ5+34CifuLfQiLdeyJ2QHXBozYjsBXicjB6QLypYrI gOED8qRfB4t18IleJIt19AHaKd6LXfCJcyjpDQEAADnzfXCLdeyLfQiJ2QaM2I7AV4nIwekC 8qWKyIDhA/KkXweLdfCJXhyLdegB2inei13wiXMgi3Xwi13oik4IiV4Y9sEIdSCDffQAfwX2 wRB0FYtNCInTAcvGAy6LXfBCx0MkAQAAAIt18Itd9IleKOmZAAAAi30IifGLdewGjNiOwFeJ yMHpAvKlisiA4QPypF8Hi33wi3Xoim8IAfIp84l3GPbFCHUYg330AH8F9sUQdByLRQiJ1wHH QsYHLusPi3UIgD4wdQfHRxgAAAAAi030Oct+AonLi3XsA3Xoi30IidkB1waM2I7AV4nIwekC 8qWKyIDhA/KkXweLdfAB2olWHIt19Cnei13wiXMgA1UIxgIAjWX4X15dwgQAUmRAAGJkQABt ZEAAeGRAAFWJ5VZXg+wMUFJTUYsAiUXsi0Xoi1AEhdJ/BQFV7OsOi13sKdOJXeyNSwGJTeyL Regx/4pQCIl99PbCBHQbi0XgO0XsfQOJReyLXexLiV3shdt9BTH2iXXsi0Xoi3gEhf9/FotF 9ItdCI1QAQHYiVX0xgAw6XwAAACLVeCJffA5134DiVXwi30Ii130i03wi3XkAd+LVegGjNiO wFeJyMHpAvKlisiA4QPypF8Hi0Xwi3Xki33gjQwDAcYpx4lN9Il15ItaBIl94DnYfSiLTfCJ 2It19CnIujAAAACJRfCLRQiLXfAB8OiXzv//i0XwjTwGiX30i1Xoi0X0iUIY9kIICHUgg33s AH8Ji0Xo9kAIEHQRi0X0i30IjXABAfiJdfTGAC6LReiLUASF0n0li0UIi030idOJVfC6MAAA APfbAciJXfDoOc7//4tF8I00AYl19It97IX/fkk7feB9A4l94Itd4IXbdCaLfQiLTfSLdeQB z4nZBozYjsBXicjB6QLypYrIgOED8qRfBwFd9ItV6ItF9It94IlCHItF7Cn4iUIgi0Xog3gM AHQWi0X0i00IjVgBi1XoAciKUgyJXfSIEItF9It13ECF9nwPi1X0iUX0i0UIxgQCK+sUi1X0 ifeJRfT334tFCIl93MYEAi2LXeiLWxCD+wN3PonYLv8khT1iQACBfdzoAwAAfSa7AwAAAOsk g33cCnwFuwIAAACDfdxkfAW7AwAAAIF93OgDAAB8BbsEAAAAi0XoiVgQg/sEfEsx/4tF3Il9 8D3oAwAAfCaJwr/oAwAAwfof9/+JRfCJwsHgBSnQweACAdCLVdzB4AMpwolV3ItF9It1CI1I AYpV8AHwgMIwiU30iBCD+wN8STH/i0XciX3wg/hkfCaJwr9kAAAAwfof9/+JRfCJwsHgAinQ weADAdCLVdzB4AIpwolV3ItF9It1CI1IAYpV8AHwgMIwiU30iBCD+wJ8QzH/i0XciX3wg/gK fCCJwrsKAAAAwfof9/uJRfCJwsHgAgHQi1XcAcApwolV3ItF9ItNCI1YAYpV8AHIgMIwiV30 iBCLRfSLfQiNcAGKVdwB+IDCMIgQi0XoifKLQBwpwotF6IlQJI0EN4l19MYAAOmR/P//VYnl UVZXg+w0ideJXeiJReCKEID6IHQKgPoJcgiA+g13A0Dr7I1YATDJgPordAeA+i11BLEBidgx 27YwiV3kihBAgPoudQr2wQh1LYDJCOvugPowciOA+jl3HvbBCHQD/0XkCNaA/jB0CoP7E30E iFQrwEOAyQTrxjH29sEED4RlAAAAgPpldAWA+kV1V41Q/4ooiVXwjVABgP0rdAiA/S11BYDJ AonQgOH7ihCA+jByIoD6OXcdgf7oAwAAfQ9r9gqJdewPtvIDdeyD7jCAyQRA69f2wQJ0Avfe 9sEEdQaLRfDrAUiJReCDfegAdAiLVeiLReCJAit15IP7E34Kg+sTAd67EwAAAIXbfguAfCu/ MHUERkvr8YXbdRFmx0cIAACJXwSJ+IkfMfjrXzDSjUXAiFQrwI1V1OhLFQAAhfZ0Co1F1Iny 6KT0///2wQF0BIBN3YCLRdxmiUcIi0XYiUcEi0XUiQeNRB7/PTQBAAB+B7gDAAAA6xM9zP7/ /30HuAIAAADrBbgBAAAAjWX0X15ZXcNVieVTUYPsJInTjVXU6Fr+//+Jw4XAdQuJReiJRezp rQAAAItF3IDkfyX//wAAPf9DAAB8KuhT2///9kXdgHQQ3QUAykAA2eDdXejpgAAAAKEAykAA iUXooQTKQADrvD3NOwAAfUKD+Mx9DzHJ6Bvb//+JTeiJTezrVo1V6I1F1Nso3RqLReiJRfCL ReyLXeiJRfSp////f3UEhdt0LWb3Rfbwf3Uq6yONVeiNRdTbKN0ag/sDdBSLReiJReCLReyJ ReRm90Xm8H91Bei92v//3UXojWX4WVtdwwAAAAAAAAAAAAAAAAAAAFFWV1WD7AyJRCQEidWJ HCQx24XAD4RxAAAAgzwkAA+GiQAAAGaLTQBmhcl0IDHSjUQkCGaJyuiF6v//icKD+P8PhGkA AAA7BCR3YusJi0QkBMYAAOtXjXQkCIt8JASJwQaM2I7AV4nIwekC8qWKyIDhA/KkXweDxQIB 04sEJAHXKdCJfCQEiQQk649mg30AAHQbMdKNRCQIZotVAOgf6v//g/j/dAmDxQIBw+veidiD xAxdX15ZwwAAAAAAAAAAAFNRUlZXVYnFizX08cgBhfZ0P4XAdDuJxwaM2I7AKclJMcDyrvfR SQeJz+sfifuJ6onI6LkTAACFwHUNgDw5PXUHjUcBAcjrC4PGBIsOhcl12zHAXV9eWllbwwAA AAAAAFFWV1Ux7YXAD4SaAAAAjUgLOcEPgo8AAACJ34Dh+ItDFIP5EHMFuRAAAAA5wQ+HdQAA AItfEItHDDnZdwWLRygx2413IIsQOdF2EjnadgKJ04tACDnwde2JXxTrS4lfEItfGEMpyolf GIP6EHIejRwIiV8MiROJCItIBIlLBItQCIlTCIlZCIlaBOsS/08ci1AEiVcMi1gIiVoIiVME ihCAygGNaASIEInoXV9eWcOQUVZXVYnZhcAPhB0BAACNWPz2AwEPhBEBAACLE4Di/o0EE4nO 9gABdSKLOAH6iRM7QQx1A4lZDItQBItACIlCCIlQBP9OHOmUAAAAiROLQQw5w3MWO1gED4eC AAAAi0EoOcMPgncAAADrGYtACDnDD4JqAAAAjUEgi1AEOdMPh1wAAACLfhyLRhgx0o1PAffx icE5+HMoi0YYKfgByTn4dwW5/////4sDAdiLEPbCAXQug/r/dAiA4v4B0El17ItGDDnDcwOL Rig5w3ITi0AIOcNyDItACDnDcgWLQAjr6YtQBIs6AdeLCznfdRKLKgHpiQo7Xgx1A4lWDInT 6w//RhyJQwiJUwSJWgiJWASLVhhKi34MiVYYOftzCDtOEHYDiU4QO04UdgOJThRdX15Zw1NS icKh7M5AADHbhcB0DTnCcgmJw4tACIXAdfOJWgSJQgiF23QFiVMI6waJFezOQACFwHQDiVAE jVogg8Isx0L0AAAAAMdC5AAAAADHQuwAAAAAx0LwAAAAAIla+ItC1Ila/IPoLIla4IkCxwQC /////4nQWlvDkFNRUlZXg+wEiQQkgz2g00AAAHUHMcDpmAAAAIM9wMxAAP51CzHAg8QEX15a WVvDieDomgAAAIXAD4R1AAAAakBoABAAAItMJAhRagAu/xUAskAAicKFwHRaiwQkizQkg+gE OfB2CzHAg8QEX15aWVvDiQQkg/g4cwsxwIPEBF9eWllbw4kCidDo+/7//4nDiwCJBCQMAYkD i3oYx0IUAAAAAEeNQwSJehjoSuX//7gBAAAAg8QEX15aWVvDUonC6PQEAACJ0Og1////WsON QABRUonCiwCDwAck+IXAdD2JAoPAPDsCcwUxwFpZw4sNpNNAADnIcwSJyCT+iQIF/w8AADsC cwUxwFpZwzDAgOTwiQKFwA+VwCX/AAAAWlnDAAAAAAAAAAAAMcDDAAAAAAAAAAAAAAAAAIA9 UNFAAAB0GoHi//8AAFJQzOsGV1ZJREVPuAEAAACDxAjDMcDDAAAAAAAAAAAAAFFSLv8VdLFA AFpZwwAAAABTUVJWVYHsKAIAAP8VJM1AAL1wyUAAi3AEMcmNhCQUAQAAicro79P//7oCAAAA jYQkFAEAAEHoNRAAAIXAdNyNhCQUAQAAieroyxIAAIXAdRvo2hMAAIM4C3QK6NATAACDOAZ1 tjHA6ZgAAAAx2+g90///ih20zEAAieCJ2uiW0///ieKNhCQUAQAA6MATAACFwHVRieCJ6uh7 EgAAicKFwHQqimANgMwIiGINi0IIiFgUifCIHbTMQADo0dT//4nQgcQoAgAAXV5aWVvD6F4T AACDOAt1DjHAgcQoAgAAXV5aWVvDugIAAACNhCQUAQAAQ+h5DwAAhcAPhRz////pdf///4HE KAIAAF1eWllbw1NRUlZXgD1o0UAAAA+FqwAAALtU0UAA6znoovr//4nChcB0K4nHBozYjsAp yUkxwPKu99FJB4H5AwEAAHcRuwMBAAC4aNFAAOgTEwAA6wqDwwSLA4A4AHXAgD1o0UAAAHUq MdIxwL9o0UAA6KATAACJxleKBogHPAB0EIpGAYPGAohHAYPHAjwAdehfv2jRQAAGjNiOwCnJ STHA8q730UkHjUH/BWjRQACKGID7XHQNgPsvdAhAxgBcQMYAALho0UAAX15aWVvDAAAAAAAA AABTUVJWicL2QAyAdA32Qg0QdAeJ0OjHEwAAi0IQ6M8TAACJw4nBg/j/dCqLQhD/FSjNQACL cgSF9nQP9kINEHQFjQwe6wSJ2Snxi0IQ/xUszUAAicheWllbwwAAAAAAAABTUVJWV1WJw4XA fAg7BWzSQAB2EbgEAAAA6DLT//+4/////+thixX4zkAAiy1c00AAMckx9os8moXtdB7/FVDT QACJwoXAdBKJ2P8VVNNAAInQ/xVc00AA6yGFyXUdVy7/FTSxQACFwHURuAQAAAC+/////+jX 0v//6wmJ2DHS6HQAAACJ8F1fXlpZW8MAAADpWxMAAAAAAAAAAAAAAAAAU1FSicI7BWzSQABy BjHAWllbw4P4A30zjRyFAAAAAKHA0kAAAdiKSAH2wUB1HYjNgM1AiGgBidDoAwgAAIXAdAqh wNJAAIBMAwEgocDSQACLBJBaWVvDkFPB4AKF0nQOix3A0kAAgM5AiRQDW8OLHcDSQACJFANb wwAAAAAAAABTUVZXVYnDidGFwHUJidDovM3//+tihdJ1DegB4f//McBdX15ZW8PoxBIAAInG idjoyxIAAInFhcB1O4nI6I7N//+JxYXAdCWJ8YnHid4GjNiOwFeJyMHpAvKlisiA4QPypF8H idjoteD//+sJifKJ2OiKEgAAiehdX15ZW8MAAFMPr8KJw+hFzf//hcB0BzHS6MrA//9bwwAA AAAAAAAAU1FSicOFwHQbMdLHQAwBAAAA6Enb//8u/xV4sUAAiYPaAAAAWllbwwAAAAAAAAAA 6ZPn//8AAAAAAAAAAAAAAFNRUlb/FUDNQACh7M5AAIXAdByLGItIJIPrLIsxi1AIOfN1Behz AAAAidCF0nXk/xVIzUAAMcBeWllbw1NRUlZXicNoAIAAAGoAUItwCC7/FQSyQACFwHUHuP// ///rNjsd8M5AAHUchfZ0CIk18M5AAOsQoezOQACJNfTOQACj8M5AADsdEPLIAXUIMf+JPRDy yAExwF9eWllbw1NSi1gEi1AI6JP///+FwHUWhdt1CIkV7M5AAOsDiVMIhdJ0A4laBFpbwwAA AAAAAAAAAADpG8z//41AAFNRUlZXVYs99PHIAYX/D4XBAAAAiy0BzUAAMcmKVQCJ6ITSdBaK MI1YAYT2dASJ2OvzQYozidiE9nXqiT308cgBKeh1BbgBAAAA6MnL//+JwonDhcAPhHUAAACj cPPIAY0EjQAAAACDwAQByOiny///hcB0VInHilUAMfYxyYnohNJ0Gon6iRwRQ4oQQIhT/4TS dfWDwQSKEEaE0nXmifjHBAEAAAAAg8EEifONBA8x0qPw8cgBiT308cgB6Oe+//+LPfTxyAHr B4nQ6Jje///ocxIAAIs99PHIAV1fXlpZW8OLwFNRUlZX6HoTAACLFfTxyAGF0nQPidAx2+hn 3v//iR308cgBiw1w88gBhcl0D4nIMfboTt7//4k1cPPIAYs9Ac1AAIX/dAhXLv8VXLFAAF9e Wllbw1FSicJmgzgAdAxmi0gCg8ACZoXJdfQp0NH4WlnDAAAAUVZXidmJ1onHBozYjsBXicjB 6QLypYrIgOED8qRfB4n4X15ZwwAAAAAAAAAAAAAAZoM4AXUcg3gEAHQWZotACmY9EAByBmY9 EgB2BrgBAAAAwzHAw41AAFFSMcD/FSjNQACDPcTSQAD/dSNqAGiAAAAAagNqAGoBaAAAAIBo dMlAAC7/FTyxQACjxNJAAIM9yNJAAP91I2oAaIAAAABqA2oAagJoAAAAQGh8yUAALv8VPLFA AKPI0kAAMcD/FSzNQABaWcONQADoi////6HE0kAAw5Dof////6HI0kAAwwDb4sMAAAAAAAAA AAAAAAAAU1GJw4P4B3QFg/gEdRiNBN0AAAAAi5jM0kAAiZDM0kAAidhZW8ONDN0AAAAA/xUk zUAAi1wBWP8VJM1AAIlUAViJ2Flbw4vAUonCg/gHdAWD+AR1CYsE1czSQABaw/8VJM1AAItE 0Fhaw4vAUonCg/gHdAWD+AR1CYsE1dDSQABaw/8VJM1AAItE0Fxaw4vAU4nD6NT///85wnUJ idjopf///1vDMcBbw41AAItEJASFwHYKg/gBdCQxwMIEALgEAAAA6ID///+JwoXAdQPCBAC4 BAAAAOjdAQAA6x24BwAAAOhh////icKFwHUDwgQAuAcAAADovgEAAIP6AnQFg/oDdQUxwMIE ALgBAAAAwgQAjUAAUrgEAAAA6Cn///+JwrgHAAAA6B3///+D+gJ0BYP6A3UKg/gCdAyD+AN0 B7gBAAAAWsMxwFrDi8BRUoA9NNNAAAB1GWoBaBx1QAAu/xXMsUAAhcB0B8YFNNNAAAExwKA0 00AAWlnDi8BRUoA9NNNAAAB0GmoAaBx1QAAu/xXMsUAAhcB0CDDSiBU000AAMcCgNNNAAIXA D5TAJf8AAABaWcONQAC4AQAAAOn2AAAAi8BTUVKJw7gCAAAA6HX+//+JwYP4AXQlg/gCdCCD +AN0G7oCAAAAidDoEP7//7gCAAAAidr/0THAWllbw7j/////Wllbw1NRVonDidGD+AF8BYP4 DH4TuAkAAADoDcz//7gDAAAAXllbw8cFqNNAACh2QACD+QJ0H4P5A3Qa6Cf+//+FwHQRg/sC dQy6nwAAADHA6IYQAACJ2Ojn/f//icqJxonY6JT9///oo/7//4XAdAfo0v7//+sF6Pv+//+J 8F5ZW8OQUHdAAER3QABad0AAWndAAFp3QABad0AAWndAAFp3QABad0AAWndAAFp3QABad0AA U1FSicPohv3//41T/4nBg/oLd00u/ySV+HZAALiMAAAA6Ob+///rQoP4AnUF6N4PAACD+QF0 GoP5AnQVg/kDdBC6AgAAAInY6Pv8//+J2P/R6Ab+//+FwHUQ6GX+///rCbj/////WllbwzHA Wllbw1NSVlcGuggAAAD/FSTNQACNNBCM2I7AjX5YjbLM0kAAg8IIpaWD+mh14LoAdUAAuyh3 QACJFRTQQACJHRjQQAAHX15aW8OLwFLonv3//4XAdCPo/f3//7oCAAAAuAQAAADocvz//7oC AAAAuAcAAADoY/z//1rDkFNSuph3QAC74HdAAIkVZM1AAIkdaM1AAFpbwwAAAAAAU1FSicP/ FSjNQACDPVDTQAAAdB2J2P8VUNNAAIXAdBGJ2P8VLM1AALgBAAAAWllbw6H4zkAAiwSYUC7/ FYixQACD+AJ1EYnY/xUszUAAuAEAAABaWVvDidj/FSzNQAAxwFpZW8MAAAAAAAAAAAAAAFNR UlZVg+wUicaD+P91Di7/FWCxQACJxumYAAAAg/j+dQ4u/xWksUAAicbphQAAAIP4/XUmuwEB AAC4MPLIATHS6Oq4//8x0jHAiRUg8sgBiRU400AA6f8AAACD+Px1VbsBAQAAuDDyyAEx0ui/ uP//uIEAAACyAUCIkDDyyAE9nwAAAH7yuOAAAAC2AUCIsDDyyAE9/AAAAH7yvQEAAAC4pAMA AIktIPLIAaM400AA6aMAAACF9nUFvgEAAACJ4FBWLv8VZLFAAIXAdQ64AQAAAIPEFF1eWllb w7sBAQAAuDDyyAEx0uhEuP//McmKZCQGiQ0g8sgBhOR0CscFIPLIAQEAAAAx27EB6xwxwIpE HAbrB0CIiDDyyAEx0opUHAc50H7vg8MCgHwcBgB13YB8HAcAddaD/gF1Fy7/FaSxQACjONNA ADHAg8QUXV5aWVvDiTU400AAMcCDxBRdXlpZW8MACdJ4HgnJeAboSAAAAMP32ffbg9kA6DsA AAD32vfYg9oAw/fa99iD2gAJyXkU99n324PZAOgcAAAA99n324PZAMPoDwAAAPfZ99uD2QD3 2vfYg9oAwwnJdRpLdBZDOdN3CYnBidAp0vfzkffzidOJyinJwznRchx1EjnDdw4p2InDKckp 0rgBAAAAwynJKduTh8rDVVZXKfaJ94n1AdsRyXITRTnRcvV3BDnDdu/4EfYR/014ItHZ0dsp 2BnK9XLuAfYR/014DNHp0dsB2BHKc+/r2wHYEcqJw4nRifCJ+l9eXcMAAAAAAAAAAACF0nUE McDDkIXSdPpSicLo5Mf//4P6e3UHuAEAAADrRYH6zgAAAHURuAkAAADofsf//7j/////WsOB +rcAAAB1EbgHAAAA6GXH//+4/////1rDg/oTdgW6EwAAAIuCOdNAAMH4GOhGx///uP////9a w41AAFFSLv8VkLFAAOiK////WlnDAAAAAAAAAAAAAAAxwMOQ/xWU00AAw5D/FZjTQADDkP8V nNNAAMOQUVaFwHQQiwiLNZTTQACJDZTTQACJMIXSdA+LAosNmNNAAKOY00AAiQqF23QPiwOL FZzTQACjnNNAAIkTXlnDUmaLUAiA5n9mgfr/f3UcgXgEAAAAgHUMgzgAdQe4AwAAAFrDuAIA AABaw2b3QAj/f3UWg3gEAHUJgzgAdQQxwFrDuAQAAABaw7gBAAAAWsMAVVZXUVNSicYp0inJ Ke0pwIoGPAB0L4nXicuJ6AHtEckR0gHtEckR0gHFEdkR+gHtEckR0inAigYkDwHFg9EAg9IA RuvLici/XkAAAOgRAAAAXYlVBIlFAGaJdQhbWV9eXcMp9gnGCdYJ7nQ2CdJ1CYnCiegp7YPv IAnSdQmJwonoKe2D7yAJ0ngJTwHtEcAR0uvzAe2D0ACD0gBzA9HaR4n+wwBRVonBidYx0oXb djyJyOhtDAAAhcB1MYnw6GIMAACFwHUmifKJyOiVDAAAicKFwHU9icjo6AwAAInBifBL6N4M AACJxoXbd8SF23YgicjoLQwAAIXAdQuJ8OgiDAAAhcB0CjHSMcCKEYoGKcKJ0F5ZwwAAAAAA AAAAAAAAAAAAAFNSicP/FUDNQACLFaDTQACJHaDTQAD/FUjNQACJ0FpbwwAAAAAAAAAAAAAA AAAAAFNRidNQLv8VhLFAAIP4/3UI6M79//9ZW8P2wwJ0EagBdA24BQAAAOhQ/f//WVvDMcBZ W8MAAAAAAAAAAAAAAABTUVZXVYPsBInBida6AQAAADHbMf8x7YkcJIX2dA47FazTQAB1BIkW 6wKJHooBPHJyCHYMPHd0DesVPGF0DOsPgMsB6xuAywLrFoDLgusRuAkAAADoh8T//zHS6c8A AACKYQFBhOQPhK4AAACF0g+EpgAAAIoBPGNyHQ+GZQAAADxuD4KGAAAAD4ZsAAAAPHR0Mul3 AAAAPCsPgm8AAAB2CTxidC7pZAAAAIM8JAB0BzHS6VcAAAC4AQAAAIDLA4kEJOtKhf90BDHS 60K/AQAAAOs7hf90BDHS6zO/AQAAAIDLQOsphe10BDHS6yGKBgwBvQEAAACIBusUhe10BDHS 6wyKJoDk/r0BAAAAiCaKQQFBhMAPhVL///+F/3UPgT24zEAAAAIAAHUDgMtAidqJ0IPEBF1f Xllbw4vAVldVicaI0InKi0wkFIBhDPwJWQwl/wAAAOgWCwAAPHJ1IjHA9sMCdAW4AgAAAPbD QHQFgMwC6wOAzAFqAIt8JBRX6zD2wwEPlcAl/wAAAIPAIfbDgHQEDBDrAgxA9sNAdAWAzALr A4DMAWiAAQAAi2wkFFVQVujtCgAAg8QQiUEQg3kQ/3ULicjoysD//zHA607HQQQAAAAAi0EI x0EUAAAAAMdADAAAAACLQQiJUBCLQQjHQBUAAAAAi0EIx0AIAAAAAPbDgHQOuwIAAACJyDHS 6EUNAACJyOjG0v//ichdX17CCACLwFFWV4PsBInGidGJ34niicjox/3//4nDhcB0IjHA6Iq/ //+JwoXAdBNQMcJXifCKEYtMJAjo3/7//4nCidCDxARfXlnDUzHb6LT///9bw4vAU1FSicP/ FTjNQAChQOTIAYXAdCuLUAQ503Ue9kIMA3QMugEAAACJ2Ohzwf///xU8zUAAidhaWVvDiwCF wHXVukTkyAGLAoXAdCY7WAR1HYsIiQqLFUDkyAGjQOTIAYkQ/xU8zUAAidhaWVvDicLr1LgE AAAA6PLB////FTzNQAAxwFpZW8OLwFFWV1WD7ASJx4nRieKJyOjw/P//icWFwHRVi3MQifD/ FSjNQACDPVTTQAAAdAiJ8P8VVNNAAInY6Dv///+Jw4XAdCJQi1gMMdKB4wBAAABqAIlYDInr ifiKEYtMJAjo3f3//4nDifD/FSzNQACJ2IPEBF1fXlnDAAAAAAAAAAAAAAAAAP8VJM1AAIPA BMOLwP8VJM1AAIPACMMAAAAAAAAAAAAAUVJQLv8VxLFAAIXAdQfo8fn//1nDMcBZwwAAAAAA AABRVleD7ASJxonRhcB1JbsEAQAAidjoprz//4nHhcB1EbgFAAAA6PbA//8xwOl0AAAAicaF yXQFgDkAdRCJ2onw6GsAAACDxARfXlnDuoTJQACJyOi4DAAAhcB1KIP7BHMaifjoqM///7gO AAAA6K7A//8xwIPEBF9eWcOhhMlAAIkG6x+J4FBWU1Eu/xWMsUAAhcB1DuhK+f//McCDxARf XlnDifCDxARfXlnDAAAAAFNRVleB7AQBAACJw4nWieBQaAQBAAAu/xVwsUAAicGFwHUJ6Az5 //8xwOtlhdt1KUA5xnYCifDo1Lv//4nDhcB1MrgFAAAA6CTA//8xwIHEBAEAAF9eWVvDOfB2 F7gOAAAA6AnA//8xwIHEBAEAAF9eWVvDieZBid8GjNiOwFeJyMHpAvKlisiA4QPypF8HidiB xAQBAABfXllbwwAAAACFwA+F+Kz//+jzCwAAMcDDU1FSicOFwHwIOwVs0kAAdhO4BAAAAOil v///uP////9aWVvDixX4zkAAixSa/xUozUAAagFqAGoAUi7/FdyxQACJwonY/xUszUAAg/r/ dQXoL/j//4nQWllbwwAAAAAAUVJQLv8VSLFAAIXAdQjoEfj//1pZwzHAWlnDAAAAAACD6ASL ACT+g+gEwwAAAAAAU1FWg+wEicaJ0/8VQM1AALjszkAAjNoxwInhZonQifLoFwAAAIXAD4W7 AQAA/xVIzUAAifCDxAReWVvDVldVg+wQiUQkDInXic2NQwsk+DnYcwe4/////+sKg/gQcwW4 EAAAAI1X/IkUJIsSgOL+OdAPhgsBAACLDCQp0AHRiUUAixmD+/91CrsCAAAA6U0BAAD2wwEP hdkAAACLQQiJRCQIi0EEjNqJRCQEuOzOQABmO1QkDHUhizXszkAAg34IAHQVOf53CIsGAfA5 +HcJi3YIg34IAHXri1YMOdF1BotCBIlGDDtdAHIMidiLVQAp0IP4EHM0i0QkBItUJAiJUAiJ 0ItUJASJUASLBCQBGItGHEgw0olGHIgVYfPIAYtVADnTckTpswAAAAHRiQGLRCQEiUEEi0Qk CIlBCItEJASJSAiLRCQIiUgEixQki0UAiwox2wHBMOSJCoglYfPIAYnYg8QQXV9ew4nQKdgB 2ekK////uwEAAACJ2IPEEF1fXsMpwoP6EHJUicOLDCSAywGAygGJGY0cAbjszkAAiRwkiROM 2otcJAxmOdp1IYs17M5AAIN+CAB0FTn+dwiLBgHwOfh3CYt2CIN+CAB16/9GGIsEJIPABOhD zP//MduJ2IPEEF1fXsP/FUjNQAAxwIPEBF5ZW8MAAAAAAAAAAABTUVJWV1WD7BC4iMlAAOhd 4///icGFwA+E+QAAAIA5AA+E5gAAALI6ic6KBjrCdBI8AHQMRooGOsJ0B0Y8AHXqK/aJ4In1 Kc6Jyonz6F8JAAC7EAAAADDSieCIFDQx0ujcCgAAjXUBsjqJx4nxigY6wnQSPAB0DEaKBjrC dAdGPAB16iv2ieCJ9SnOicqJ8+gcCQAAuxAAAAAw9ongiDQ0MdLomQoAAI11AbIqiUQkDInx igY6wnQSPAB0DEaKBjrCdAdGPAB16iv2ieCJ9SnOicqJ8+jXCAAAieAw2zHSiBw0uxAAAADo VAoAAInBifqLRCQM6G+9//+Jyon4jXUB6NPp//+J8ekR////uJTJQADomgoAAIPEEF1fXlpZ W8NTUVJWixX08cgBhdIPhIAAAADrMIM98PHIAQB0JIs19PHIAYnRKfGLHfDxyAHB+QKAPBkA dAXoxcr//8cCAAAAAIPCBIsChcB1yoM98PHIAQB1DLgFAAAA6FO3///rD7oFAAAAofTxyAHo cun//4XAdQq4/////15aWVvDo/TxyAHHAAAAAACDwASj8PHIAcYAADHAXlpZW8MAAAAAAAAA /xWo00AAi8BSuv8AAAC4pMlAAOgkuP//WsMAAAAAAABTUVZXg+wEicMxyYolqMpAAIkMJITk dCaJ5zbZP5uF0nQcidCLNCT30CHaIfAJ0InniQQkNtkvm4nnNtk/mzHAZosEJIPEBF9eWVvD AAAAAFNWVwYPoA+oVYnlg+wMi3UgixVQ5MgBiwaLfgSJRfyF0nUxoXDNQACDwAMk/CnEieGL HXDNQACJyOjyqf//oXDNQACJgfAAAACJyOjowP//hcB0Mf8VJM1AAIteDAXeAAAAU4lGEC7/ FdixQACNRfToC8n///8VZM1AAIn4/1X86Gfz//+J7F0PqQ+hB19eW8ONQABRVlcGVYPsOInB idWDPSDNQAD/dRLoKsD//4XAD4S+AAAA6FnB//+J54lMJCCJXCQkLv8VfLFAAIlEJCiM2DHb jsC+vMlAAIlcJDC7EAAAAKWlpC7/FXixQACNVCQI6NXK//+J4FBqAGoAagAu/xU4sUAAiUQk LI1EJDRQagCNRCQoUIHF/w8AAGigh0AAgeUA8P//VWoALv8VQLFAAInDicWFwHUKx0QkNP// ///rImr/i0wkMFEu/xUMskAAi3QkMIX2dASJHusIUy7/FTSxQACLRCQsUC7/FTSxQACJ6IPE OF0HX15Zw41AAFFS/xVozUAA6D/I//+DPVDkyAEAdQq4AQAAAOjsv///agAu/xVUsUAAWlnD AAAAAFKAOAB1B7gBAAAAWsODPSDyyAEAdCIx0ooQipIx8sgBgOIBgeL/AAAAdA2AeAEAdQe4 AgAAAFrDMcBawwAAAABTUYPsCInDidGJwong6O4LAACJ2DDS6CUMAACIFASNRCQEicro1wsA AInIMPboDgwAAIh0BASJ4OgzDAAAjUQkBI1UJAToJgwAAIng6G8MAACDxAhZW8MAAAAAAAAA AABSgz0g8sgBAHQgMdKKEIqSMfLIAYDiAYHi/wAAAHQLgHgBAHQFg8ACWsNAWsMAAADHBazT QAABAAAAwwAAAAAAg/hBfAiD+Fp/A4PAIMMAAFNSjUQkFIPABItA/FBqAItUJBhSi1wkGFPo BgAAAIPEEFpbw1NRUoPsBI1EJCCJ4YtcJByLVCQYiQQki0QkFOgIAAAAg8QEWllbw5BWV1WD 7CCJx4lUJByJ3ehhuP//hcB0FLgLAAAA6PO3//+4/////+n+AQAAidaNXCQUg+YHjVQkEInw 6BUMAACNVCQMiegx2wnwvYAAAADoNQwAALoMAAAAimQkHIlcJASJFCT2xIAPlMAl/wAAAIlE JAiDPVjTQAAAdDS6yMlAAIn46H0DAACFwHUk6CS6////FTDNQABq/1BTugAgAACJwf8VWNNA AIPEDOkqAQAAilQkHPbCIA+EcgAAAIsRg8IEiRGLQvyJRCQUxwEAAAAAoTzUQACLXCQU99Ah w4lcJBT2RCQVAXQM9kQkFIB1Bb0BAAAA9kQkHQR0DboBAAAAiVQkGInQ6zf2RCQcQHQPx0Qk GAIAAAC4AQAAAOshuQQAAAC4AwAAAIlMJBjrEfbCQHQHuAUAAADrBbgDAAAAagBVUI1EJAxQ i1wkHFOLTCQkUVcu/xU8sUAAicOD+P91NvZEJBwgdB5qAFWLbCQgVWoAi0QkHFCLVCQkUlcu /xU8sUAAicOD+/91DOhI7///g8QgXV9ew4nY/xUwzUAAMdKLPWzSQACJwTn4ch5TLv8VNLFA ALgFAAAA6FW2//+4/////4PEIF1fXsPo1Ov//4XAdAW6ACAAAIP+AnUFgMoD6xGF9nUFgMoB 6wiD/gF1A4DKAvZEJBwQdAOAyoCJ0IpcJB0MQPbDA3QH9sMCdBDrDIE9uMxAAAACAAB1AonC icjoleP//4nIg8QgXV9ewwAAAABTUVZXO0IEfyyLWgiLCotbCCnLOdh8HopaDIsyi3oEgOPv AcaIWgwpx4kyMcCJegRfXllbw7gBAAAAX15ZW8OQUoBgDO+LUAiLUgjHQAQAAAAAiRBaw4vA UVZXicGJ1otAEP8VKM1AAIphDPbEBg+EdQAAAPZBDRB0L4nI6I+i//+FwHQ7hdt1DoX2fQq4 CQAAAOhJtf//i0EQ/xUszUAAuP////9fXlnDg/sBdQMrcQSLQQiLQAjHQQQAAAAAiQGAYQzr ifKLQRDogsj//4P4/w+F5AAAAItBEP8VLM1AALj/////X15Zw4P7AXINdhGD+wIPhIoAAADr koXbdEXrjInKifCLeQTo8/7//4XAD4SmAAAAifKLQRAp+ugvyP//g/j/dRKLQRD/FSzNQAC4 /////19eWcOJyOgB////6XcAAACLQRDo1PT//4tRBIn3KdApx4nKifjoov7//4XAdFmJ8otB EOjkx///g/j/dceLQRD/FSzNQAC4/////19eWcOI4oDi74tBCIhRDItACIkBifKLQRDHQQQA AAAA6KzH//+D+P91EotBEP8VLM1AALj/////X15Zw4tBEP8VLM1AADHAX15ZwwAAAAAAAFOJ w4oDiiI8QXIGPFp3AgQggPxBcgiA/Fp3A4DEIDjgdQiE5HQEQ0Lr2THSiMKI4CX/AAAAKcKJ 0FvDAAAAAAC4/////41AAFNRUonB/xU4zUAAixVA5MgBMduF0nQai0IEhUgMdAxD9kANEHQF 6M2g//+LEoXSdeb/FTzNQACJ2FpZW8MAAAAAAAAAAAAAAFFWicaF23QPgDoAdApAigpCiEj/ S3XxMdKF23QHQEuIUP/r9YnwXlnDAAAAAAAAAFZXVYPsEIlEJAyJVCQEid6JTCQIhdJ0AokC i1QkDIoC/sAl/wAAAPaAEM9AAAJ0A0Lr64oqgP0rdAWA/S11AUKF9nUsgDowdRSKSgGA+Xh0 BYD5WHUHvhAAAADrM4A6MHUHvggAAADrPL4KAAAA6zWD/gJ8BYP+JH4RuA0AAADozLL//zHA 6cEAAACD/hB1FYA6MHUQinoBgP94dAWA/1h1A4PCAokUJI0stQAAAAAwyTHbMcCKAui0AAAA icc58H0aO52o00AAdgKxAYnYD6/eAfs5w3MCsQFC69c7FCR1BItUJAyLfCQEhf90AokXg3wk CAF1D4H7AAAAgHIHdQmA/S11BITJdDq4DgAAAOg9sv//g3wkCAB1DLj/////g8QQXV9ew4D9 LXUMuAAAAICDxBBdX17DuP///3+DxBBdX17DgP0tdQL324nYg8QQXV9ew5BRMcnopP7//1nD i8BRuQEAAADolf7//1nDjUAAUjwwcg48OXcKJf8AAACD6DBawyX/AAAA6FP5//+Jwjxhcg08 aXcJMcCI0IPoV1rDPGpyDjxydwol/wAAAIPoV1rDPHNyDjx6dwol/wAAAIPoV1rDuCUAAABa wwAAAAAAU1FSVldVg+wYicW6AgAAALsBAAAAiRQkuj0AAACJXCQE6CoGAACJw4XAdQq4//// /+l8AQAAOeh1D7j/////g8QYXV9eWllbwynoiUQkEEDoy6z//4nCiUQkCIXAdQ+4/////4PE GF1fXlpZW8OLTCQQie6JxwaM2I7AV4nIwekC8qWKyIDhA/KkXweLRCQQjXMBxgQCAInw6BIG AACJRCQUhcB0S0DodKz//4nDiUQkDIXAdRaJ0Oizv///uP////+DxBhdX15aWVvDi0wkFInH BozYjsBXicjB6QLypYrIgOED8qRfBwNcJBTGAwDrBIlEJAyLRCQMUItUJAxSLv8V0LFAAInC i0QkCOhfv///i0QkDOhWv///hdJ1D7j/////g8QYXV9eWllbw4no6IgAAACFwHQPuP////+D xBhdX15aWVvDgz348cgBAHRiiejoWQUAAI1YAYsEJA+vw+i7q///icGFwHUZuAUAAADoZ+j/ /7j/////g8QYXV9eWllbww+vXCQEieroUAUAAIP4/3UWicjo1L7//7j/////g8QYXV9eWllb w4nI6EYHAACDxBhdX15aWVvDU1FSVldVg+wMiUQkBIXAdQq4/////+mJAQAAgDgAdBWNUAGA OgB0DYA6PXQIikoBQoTJdfOAOgB1D7j/////g8QMXV9eWllbw4B6AQAPlMAPtuih9PHIAYlE JAiFwHVBhe0PhTwBAAC4CQAAAOjyqv//iUQkCIXAdQ+4/////4PEDF1fXlpZW8Oj9PHIAYPA CIlo+Ilo/KPw8cgB6eUAAACLRCQEieroBAEAAIXtD4XuAAAAhcAPj8cAAACJxffdjV0BjQyt AAAAAI1BCIs18PHIAYkEJAHYhfZ1T+iDqv//icKJRCQIhcB1D7j/////g8QMXV9eWllbw4s1 9PHIAYnHBozYjsBXicjB6QLypYrIgOED8qRfB4sEJAHQMdKj8PHIAejLnf//6zyJwotEJAjo Xtz//4lEJAiFwHUPuP////+DxAxdX15aWVvDiwwkixXw8cgBAcGJ64nI6HMIAACJDfDxyAGL fCQIjQStAAAAAAH4iT308cgBx0AEAAAAAOsDjWj/i1QkCI0ErQAAAAAB0ItUJASJEKHw8cgB xgQoADHAg8QMXV9eWllbw4vAU1FWV1WJx4nVizX08cgB6csAAACKJ4n7hOQPhLwAAAAxwIoC 6FvA//+JwTHAigPoUMD//znBD4WgAAAAgDo9D4WKAAAAixX08cgBifcp18H/AoXtD4RsAAAA ixaJ8YXSdA+LQQSJAYtZBIPBBIXbdfGLNfDxyAGF9nRDifCAPAcAdAeJ0Oh6vP//iy308cgB ic4p7osV8PHIAcH+AonIifPofgcAAIkN8PHIATn3fQ6NBA9AihBHiFD/Ofd89THAXV9eWVvD jUcBXV9eWVvDQopLAUOEyQ+FRP///4PGBIsWhdIPhSv///+h9PHIASnwwfgCXV9eWVvDAAAA AAAAAAAAAABTgz0g8sgBAHQhMduKGoqbMfLIAYDjAYHj/wAAAHQMihqIGIpSAYhQAVvDihKI EFvDAAAAAAAAAAAAAAAAAAAAgz0g8sgBAHQcigAl/wAAAIqAMfLIASQBJf8AAAB0BrgCAAAA w7gBAAAAwwAAAAAAU1FSg+wEicGJw4nY6D/z//+FwHUxidiJ4ujiBgAA6B0HAADoiAcAAIng MNLon////4gUBIniidjoU////4nY6Kzz///rwonIg8QEWllbwwBTUYoYOhp1QYM9IPLIAQB0 HzHbihiKmzHyyAGA4wGB4/8AAAB0CopYAYpKATjLdQUxwFlbw4jYJf8AAACIyoHi/wAAACnQ WVvDMduKGDHAigIpw4nYWVvDAAAAAAAAAACoAXQOxwIAAACAxwMBAAAA6wzHAgAAAMDHA4AA AACoAnQDgAsCqAR0A4ALBMONQACD+AJ1DccCAAAAwMcDgAAAAMOD+AF1DccCAAAAQMcDgAAA AMPHAgAAAIDHAwEAAADDjUAAU4nDg+Bwg+MHg/ggchB2OIP4MHJBdimD+EB0NFvDhcB2B4P4 EHQRW8PHAgEAAACF23UjgAoCW8PHAgAAAABbw8cCAgAAAFvDxwIBAAAAW8PHAgMAAABbwwAA AABTUYPsBInDidGJ4onI6C4GAACJ4DDS6EX+//+IFASJ2Oi78f//hcB1GIniidjorv7//4XA dAuJ2OhD8v//icPr3YnY6Jjx//+FwHQEhcl1BInY6wIxwIPEBFlbwwAAAABTUonCMduJ0Ohz 8f//hcB1DInQ6Ajy//9DicLr6YnYWlvDAAAAAAAAAAAAAAAAAABRVldVicaJ0YndMf+FwHQ4 he12WoA5AHQYuwIAAACJyonw6KkFAACD+P91DF1fXlnDZscGAADrNonITeiw8f//g8YCR4nB 68i+AgAAAIA5AHQcifOJyjHA6HMFAACD+P90DonI6Ifx//9HicHr34n4XV9eWcMAAAAAAAAA AAAAAFNRUlZXVYPsEInBugIAAACJVCQEiRQkuj0AAADozwUAAInHhcB1Crj/////6VoBAAA5 yHUPuP////+DxBBdX15aWVvDKcjR+I0sRQAAAACNRQLoaaX//4nGiUQkCIXAdQ+4/////4PE EF1fXlpZW8OJ64nK6Ifa//+NVwKJ0GbHBC4AAOhX2v//hcB0QgHAiUQkDIPAAuglpf//iceJ xYXAdRaJ8OhmuP//uP////+DxBBdX15aWVvDi3QkDItcJAwB9+g42v//ZscHAADrAjHti0Qk CInq6EQFAACJwotEJAjoKbj//4no6CK4//+F0nUPuP////+DxBBdX15aWVvDgz348cgBAHUF 6BEGAACJyOiCAAAAhcB0D7j/////g8QQXV9eWllbw4nI6LDZ//+NWAGLRCQED6/D6IGk//+J xoXAdRm4BQAAAOgt4f//uP////+DxBBdX15aWVvDD68cJInK6DfO//+D+P91Fonw6Ju3//+4 /////4PEEF1fXlpZW8OJ8OjR+P//g8QQXV9eWllbw41AAFNRUlZXVYPsCInHhcB1Crj///// 6WMBAABmgzgAdBuNUAJmgzoAdBJmgzo9dAxmi3ICg8ICZoX2de5mgzoAdQ+4/////4PECF1f XlpZW8Nmg3oCAA+UwDHbiw348cgBiMOFyXVBhdsPhQ0BAAC4CQAAAOixo///icGFwHUPuP// //+DxAhdX15aWVvDo/jxyAGDwAiJWPgx9olY/KPw8cgB6cgAAACJ2on46NkAAACF2w+FwwAA AIXAD4+sAAAA99iLLfDxyAGJxkCNHLUAAAAAiQQkjUMIixQkiUQkBAHQhe11Pug/o///icWJ wYXAdQ+4/////4PECF1fXlpZW8OLFfjxyAHoXdj//4tEJASLHCQB6DHSo/DxyAHomJb//+s5 icKJyOgt1f//icGFwHUPuP////+DxAhdX15aWVvDi2wkBIsV8PHIAQHFifOJ6OhDAQAAiS3w 8cgBiQ348cgBx0SxBAAAAADrA41w/6Hw8cgBiTyxxgQGADHAg8QIXV9eWllbw4vAU1FWV1WJ x4nVizX48cgB6dcAAABmiw+J+2aFyQ+ExgAAADHAZosC6DQEAACJwTHAZosD6CgEAABmOcEP hacAAABmgzo9D4WKAAAAixX48cgBifcp18H/AoXtD4RsAAAAixaJ8YXSdA+LQQSJAYtZBIPB BIXbdfGLNfDxyAGF9nRDifCAPAcAdAeJ0Ohgtf//iy348cgBic4p7osV8PHIAcH+AonIifPo ZAAAAIkN8PHIATn3fQ6NBA9AihBHiFD/Ofd89THAXV9eWVvDjUcBXV9eWVvDg8ICZotDAoPD AmaFwA+FOv///4PGBIsWhdIPhR////+h+PHIASnwwfgCXV9eWVvDAAAAAAAAAAAAAABRVleJ 2TnCdEBzKI08GjnHdiGNd/+NPBiM2k8GjsL9Tk/R6WbzpRHJRkdm86QH/F9eWcOJ1onHjNoG jsJRwekC86VZg+ED86QHX15ZwwAAAFKDPSDyyAEAdCgx0ooQipIx8sgBgOIBgeL/AAAAdBMx 0ooQweIIikABJf8AAAAJ0FrDigAl/wAAAFrDAAAAAABTUVKD7ASJw4ni6GEAAACJ4DDS6Hj4 //+IFASBPTjTQACkAwAAdSaDPSDyyAEAdB0xwIoEJIqAMfLIASQBJf8AAAB0CYnY6IUCAADr FWoBjUQkBFAu/xUssUAAieDoXv///4PEBFpZW8MAAAAAAAAAU4nDidD2x/90DInaiFgBweoI iBBbw4gaW8MAAAAAAABRVldVieWD7ASJxoXSdQcxwOmDAAAAhdsPhnYAAACAOgB1EoX2dAVm xwYAADHAiexdX15Zw4M9IPLIAQB0GTHAigKKgDHyyAEkASX/AAAAdAaAegEAdD2J0Oii9/// agGNTfyJx1E5w3MCidhQUmoIoTjTQABQLv8VyLFAAIXAdBOF9nQGi0X8ZokGifiJ7F1fXlnD uP////+J7F1fXlnDU1FXZosIidNmOcp0FonCZos6g8ACZoX/dAdmOxh0BOvsMcBfWVvDAAAA AAAAAAAAU1FWV1WD7ASJBCSJ12aBPQnNQAAAgHMOV1Au/xXUsUAA6cMAAADoldT//wHAjUgB icjoaZ///4nCicWFwA+EpgAAAIX/dQQx9usrifjobtT//wHAjUgBicjoQp///4nGhcB1EonQ 6IWy//8xwIPEBF1fXllbw4sUJInLiejo/sj//4P4/3UdiejoYrL//4X2dAeJ8OhXsv//McCD xARdX15ZW8OF9nQiicuJ+onw6M3I//+D+P91Eonw6DGy//8xwIPEBF1fXllbw1ZVLv8V0LFA AInCiejoFLL//4X2dAeJ8OgJsv//idCDxARdX15ZW8MAAAAAAAAAAAAAAAAAAFNRUlaLNfTx yAGF9nRBixaDxgSF0nQ4idDoEvj//41YAY0EXQAAAADoc57//4nBhcB02+go+P//g/j/dAmJ yOg0+v//68iJyOijsf//679eWllbwwAAAAAAAAAAAAAAAGY9YQByCWY9egB3A4PoIMNSicLo GAAAAIXAdAOD6iGJ0FrDAAAAAAAAAAAAAAAAAFKJwuhIAAAAhcB0H4H6gYIAAHIRgfqaggAA dwm6AQAAAInQWsMx0onQWsOI0P7AJf8AAACKgBDPQAAx0iSAiMKJ0FrDAAAAAAAAAAAAAAAA UoM9IPLIAQB0MYnCweoIgeL/AAAAipIx8sgBgOIBgeL/AAAAdBUl/wAAAOgQAAAAhcB0B7gB AAAAWsMxwFrDAIM9IPLIAQB0KYE9ONNAAKQDAAB0CoXAdBu4AQAAAMMl/wAAAIqAQdRAACQI Jf8AAADDMcDDAAAAAAAAAAAAAAD/JaCxQAD/JWyxQAD/JWixQAD/JZixQAD/JbSxQAD/JYCx QAD/JVCxQAD/JZSxQAD/JdSxQAD/JSyxQAD/JdCxQAD/JVSxQAD/JQyyQAD/JUCxQAD/JXyx QAD/JdixQAD/JUixQAD/JXCxQAD/JYyxQAD/JcSxQAD/JYSxQAD/JWSxQAD/JaSxQAD/JWCx QAD/JYixQAD/JcyxQAD/JTyxQAD/JVyxQAD/JQSyQAD/JXSxQAD/JQCyQAD/JVixQAD/Jdyx QAD/JRCyQAD/JeixQAD/JfyxQAD/JaixQAD/JcCxQAD/JQiyQAD/JcixQAD/JZyxQAD/JfCx QAD/JTSxQAD/JfixQAD/JeyxQAD/JeCxQAD/JfSxQAD/JZCxQAD/JbyxQAD/JUyxQAD/JXix QAD/JUSxQAD/JbixQAD/JTixQAD/JayxQAD/JeSxQAD/JRSyQAD/JbCxQAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA8sAAAAAAAAAAAAAAcsgAALLEAAESw AAAAAAAAAAAAACeyAAA0sQAAAAAAAAAAAAAAAAAAAAAAAAAAAAA0sgAAAAAAAEayAABUsgAA ZLIAAHKyAACCsgAAmrIAAKiyAADAsgAAzrIAANyyAADwsgAACrMAABSzAAAgswAAMrMAAESz AABcswAAcrMAAIizAACcswAAtLMAAMqzAADYswAA7LMAAPyzAAAMtAAAIrQAADi0AABMtAAA WLQAAGq0AAB6tAAAlLQAAKK0AAC+tAAA1rQAAOa0AADytAAACLUAACC1AAA6tQAAVLUAAGC1 AABytQAAgrUAAJK1AACwtQAAvLUAAMa1AADUtQAA4rUAAP61AAAOtgAAHLYAACy2AABCtgAA WLYAAAAAAAA0sgAAAAAAAEayAABUsgAAZLIAAHKyAACCsgAAmrIAAKiyAADAsgAAzrIAANyy AADwsgAACrMAABSzAAAgswAAMrMAAESzAABcswAAcrMAAIizAACcswAAtLMAAMqzAADYswAA 7LMAAPyzAAAMtAAAIrQAADi0AABMtAAAWLQAAGq0AAB6tAAAlLQAAKK0AAC+tAAA1rQAAOa0 AADytAAACLUAACC1AAA6tQAAVLUAAGC1AABytQAAgrUAAJK1AACwtQAAvLUAAMa1AADUtQAA 4rUAAP61AAAOtgAAHLYAACy2AABCtgAAWLYAAAAAAABVU0VSMzIuZGxsAEtFUk5FTDMyLmRs bAABAENoYXJVcHBlckJ1ZmZBAAABAENsb3NlSGFuZGxlAAIAQ3JlYXRlRXZlbnRBAAADAENy ZWF0ZUZpbGVBAAQAQ3JlYXRlVGhyZWFkAAAFAERlbGV0ZUNyaXRpY2FsU2VjdGlvbgAGAERl bGV0ZUZpbGVBAAcARW50ZXJDcml0aWNhbFNlY3Rpb24AAAgARXhpdFByb2Nlc3MACQBFeGl0 VGhyZWFkAAAKAEZsdXNoRmlsZUJ1ZmZlcnMAAAsARnJlZUVudmlyb25tZW50U3RyaW5nc0EA DABHZXRBQ1AAAA0AR2V0Q1BJbmZvAA4AR2V0Q29tbWFuZExpbmVBAA8AR2V0Q29tbWFuZExp bmVXABAAR2V0Q3VycmVudERpcmVjdG9yeUEAABEAR2V0Q3VycmVudFByb2Nlc3NJZAASAEdl dEN1cnJlbnRUaHJlYWRJZAAAEwBHZXRDdXJyZW50VGhyZWFkAAAUAEdldEVudmlyb25tZW50 U3RyaW5ncwAVAEdldEZpbGVBdHRyaWJ1dGVzQQAAFgBHZXRGaWxlVHlwZQAXAEdldEZ1bGxQ YXRoTmFtZUEAABgAR2V0TGFzdEVycm9yAAAZAEdldExvY2FsVGltZQAAGgBHZXRNb2R1bGVG aWxlTmFtZUEAABsAR2V0TW9kdWxlRmlsZU5hbWVXAAAcAEdldE1vZHVsZUhhbmRsZUEAAB0A R2V0T0VNQ1AAAB4AR2V0UHJvY0FkZHJlc3MAAB8AR2V0U3RkSGFuZGxlAAAgAEdldFRpbWVa b25lSW5mb3JtYXRpb24AACEAR2V0VmVyc2lvbgAAIgBJbml0aWFsaXplQ3JpdGljYWxTZWN0 aW9uACMATGVhdmVDcml0aWNhbFNlY3Rpb24AACQATG9hZExpYnJhcnlBAAAlAE1vdmVGaWxl QQAmAE11bHRpQnl0ZVRvV2lkZUNoYXIAJwBTZXRDb25zb2xlQ3RybEhhbmRsZXIAKABTZXRF bnZpcm9ubWVudFZhcmlhYmxlQQApAFNldEVudmlyb25tZW50VmFyaWFibGVXACoAU2V0RXZl bnQAACsAU2V0RmlsZVBvaW50ZXIAACwAU2V0TGFzdEVycm9yAAAtAFNldFN0ZEhhbmRsZQAA LgBTZXRVbmhhbmRsZWRFeGNlcHRpb25GaWx0ZXIALwBUbHNBbGxvYwAAMABUbHNGcmVlADEA VGxzR2V0VmFsdWUAMgBUbHNTZXRWYWx1ZQAzAFVuaGFuZGxlZEV4Y2VwdGlvbkZpbHRlcgAA NABWaXJ0dWFsQWxsb2MAADUAVmlydHVhbEZyZWUANgBWaXJ0dWFsUXVlcnkAADcAV2FpdEZv clNpbmdsZU9iamVjdAA4AFdpZGVDaGFyVG9NdWx0aUJ5dGUAOQBXcml0ZUZpbGUAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEBAQBDb3B5OiAgICAgIABTY2FsZTogICAg IABBZGQ6ICAgICAgIABUcmlhZDogICAgIAAtLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tCgAAVGhpcyBzeXN0ZW0gdXNlcyAl ZCBieXRlcyBwZXIgRE9VQkxFIFBSRUNJU0lPTiB3b3JkLgoAAABBcnJheSBzaXplID0gJWQs IE9mZnNldCA9ICVkCgAAAFRvdGFsIG1lbW9yeSByZXF1aXJlZCA9ICUuMWYgTUIuCgAAAEVh Y2ggdGVzdCBpcyBydW4gJWQgdGltZXMsIGJ1dCBvbmx5CgAAAAB0aGUgKmJlc3QqIHRpbWUg Zm9yIGVhY2ggaXMgdXNlZC4KAABZb3VyIGNsb2NrIGdyYW51bGFyaXR5L3ByZWNpc2lvbiBh cHBlYXJzIHRvIGJlICVkIG1pY3Jvc2Vjb25kcy4KAAAAAFlvdXIgY2xvY2sgZ3JhbnVsYXJp dHkgYXBwZWFycyB0byBiZSBsZXNzIHRoYW4gb25lIG1pY3Jvc2Vjb25kLgoAAAAARWFjaCB0 ZXN0IGJlbG93IHdpbGwgdGFrZSBvbiB0aGUgb3JkZXIgb2YgJWQgbWljcm9zZWNvbmRzLgoA ICAgKD0gJWQgY2xvY2sgdGlja3MpCgAASW5jcmVhc2UgdGhlIHNpemUgb2YgdGhlIGFycmF5 cyBpZiB0aGlzIHNob3dzIHRoYXQKAHlvdSBhcmUgbm90IGdldHRpbmcgYXQgbGVhc3QgMjAg Y2xvY2sgdGlja3MgcGVyIHRlc3QuCgAAV0FSTklORyAtLSBUaGUgYWJvdmUgaXMgb25seSBh IHJvdWdoIGd1aWRlbGluZS4KAAAAAEZvciBiZXN0IHJlc3VsdHMsIHBsZWFzZSBiZSBzdXJl IHlvdSBrbm93IHRoZQoAAHByZWNpc2lvbiBvZiB5b3VyIHN5c3RlbSB0aW1lci4KAAAAAEZ1 bmN0aW9uICAgICAgUmF0ZSAoTUIvcykgICBSTVMgdGltZSAgICAgTWluIHRpbWUgICAgIE1h eCB0aW1lCgAlcyUxMS40ZiAgJTExLjRmICAlMTEuNGYgICUxMS40ZgoAAAAAAAAAAAAAAAAA AEAAAAAAgIQuQQAAAAAAAAhAmpmZmZmZuT+N7bWg98awPgAAAACAhC5Bje21oPfGsD4AAAAA AAAQQE5vdCBlbm91Z2ggbWVtb3J5IHRvIGFsbG9jYXRlIGZpbGUgc3RydWN0dXJlcw0KAAAA AABGbG9hdGluZy1wb2ludCBzdXBwb3J0IG5vdCBsb2FkZWQNCgBUWgAAVW5hYmxlIHRvIGFs bG9jYXRlIHNlbWFwaG9yZSBkYXRhDQoAVW5hYmxlIHRvIGFsbG9jYXRlIHNlbWFwaG9yZSBk YXRhDQoAVVNFUjMyLkRMTAAAR2V0QWN0aXZlV2luZG93AFRoZSBpbnN0cnVjdGlvbiBhdCAw eDAwMDAwMDAwIGNhdXNlZCBhIHN0YWNrIG92ZXJmbG93IGZsb2F0aW5nIHBvaW50CmV4Y2Vw dGlvbi4KAAAAAFRoZSBpbnN0cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNhdXNlZCBhIHN0YWNr IHVuZGVyZmxvdyBmbG9hdGluZyBwb2ludApleGNlcHRpb24uCgAAAFRoZSBpbnN0cnVjdGlv biBhdCAweDAwMDAwMDAwIGNhdXNlZCBhIGRlbm9ybWFsIG9wZXJhbmQgZmxvYXRpbmcgcG9p bnQKZXhjZXB0aW9uLgoAAFRoZSBpbnN0cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNhdXNlZCBh IGRpdmlzaW9uIGJ5IHplcm8gZmxvYXRpbmcgcG9pbnQKZXhjZXB0aW9uLgoAAFRoZSBpbnN0 cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNhdXNlZCBhbiBpbmV4YWN0IHZhbHVlIGZsb2F0aW5n IHBvaW50CmV4Y2VwdGlvbi4KAAAAAFRoZSBpbnN0cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNh dXNlZCBhbiBvdmVyZmxvdyBmbG9hdGluZyBwb2ludCBleGNlcHRpb24uCgBUaGUgaW5zdHJ1 Y3Rpb24gYXQgMHgwMDAwMDAwMCBjYXVzZWQgYW4gdW5kZXJmbG93IGZsb2F0aW5nIHBvaW50 IGV4Y2VwdGlvbi4KAAAAAFRoZSBpbnN0cnVjdGlvbiBhdCAweDAwMDAwMDAwIGNhdXNlZCBh biBpbnZhbGlkIG9wZXJhdGlvbiBmbG9hdGluZyBwb2ludApleGNlcHRpb24uCgAAAABUaGUg aW5zdHJ1Y3Rpb24gYXQgMHgwMDAwMDAwMCByZWZlcmVuY2VkIG1lbW9yeSAAAAAAYXQgMHgw MDAwMDAwMC4KVGhlIG1lbW9yeSBjb3VsZCBub3QgYmUgAHJlYWQuCgAAd3JpdHRlbi4KAAAA QSBwcml2aWxlZ2VkIGluc3RydWN0aW9uIHdhcyBleGVjdXRlZCBhdCBhZGRyZXNzIDB4MDAw MDAwMDAuCgAAAEFuIGlsbGVnYWwgaW5zdHJ1Y3Rpb24gd2FzIGV4ZWN1dGVkIGF0IGFkZHJl c3MgMHgwMDAwMDAwMC4KAEFuIGludGVnZXIgZGl2aWRlIGJ5IHplcm8gd2FzIGVuY291bnRl cmVkIGF0IGFkZHJlc3MgMHgwMDAwMDAwMC4KAAAAQSBzdGFjayBvdmVyZmxvdyB3YXMgZW5j b3VudGVyZWQgYXQgYWRkcmVzcyAweDAwMDAwMDAwLgoAAAAAVGhlIHByb2dyYW0gZW5jb3Vu dGVyZWQgZXhjZXB0aW9uIDB4MDAwMDAwMDAgYXQgAAAAAGFkZHJlc3MgMHgwMDAwMDAwMCBh bmQKY2Fubm90IGNvbnRpbnVlLgoAAAAAVGhyZWFkIGhhcyBubyB0aHJlYWQtc3BlY2lmaWMg ZGF0YQ0KAAAAAFVuYWJsZSB0byByZXNpemUgdGhyZWFkLXNwZWNpZmljIGRhdGENCgBVbmFi bGUgdG8gcmVzaXplIHRocmVhZC1zcGVjaWZpYyBkYXRhDQoAVE1QAFRFTVAAAAAAVE1QRElS AABURU1QRElSAAAAAAB3YisAY29uaW4kAABjb25vdXQkAGNvbgBDX0ZJTEVfSU5GTwBDX0ZJ TEVfSU5GTz0AAAAAQUJOT1JNQUwgVEVSTUlOQVRJT04NCgAAX19iZ250aGQAAAAAY29uAAAA HwA7AFoAeACXALUA1ADzABEBMAFOAW0BAAAfADwAWwB5AJgAtgDVAPQAEgExAU8BbgEAAAAA AADwfwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAADVw8N///+9HNXDw3///70c1cPDf///vRzVw8N///+9H BMBAABDAQAAcwEAAKMBAAAAAAACAhG5BAAAAAICEbkEAAAAAYON2QQAAAABg43ZBAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACAAAAAQAAAAAA AAAAAAAAAAAAAAAAAAAAAAIAAAACAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAGEAAAAAAQAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACAGUAAAP/////sHUAA9B1AAPQd QAD4HUAAAB5AAPQdQAD0HUAA9B1AAPQdQAD0HUAA9B1AAPQdQAD0HUAA9B1AAPQdQAD0HUAA 9B1AAPQdQAAAAAAA9AAAAHAzQABwM0AAkDNAAJAzQAB/EgAAAAAAAAAAAAACAAAAAQAAAAMA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABAAAABQAAAAkAAAAAAAAAAAAAAAAAAAAAAAAA RVNUAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAARURUAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAADQzUAAUc5AAFBG AAABAAAAEA4AAAEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAICQRUAAkEVAAJBFQAAAAQEB AQEBAQEBAwMDAwMBAQEBAQEBAQEBAQEBAQEBAQEKDAwMDAwMDAwMDAwMDAwMODg4ODg4ODg4 OAwMDAwMDAxYWFhYWFhISEhISEhISEhISEhISEhISEhISAwMDAwMDJiYmJiYmIiIiIiIiIiI iIiIiIiIiIiIiIiIDAwMDAEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAMDEyMzQ1 Njc4OWFiY2RlZgAAAAAwMTIzNDU2Nzg5YWJjZGVmZ2hpamtsbW5vcHFyc3R1dnd4eXoAAAAA MDEyMzQ1Njc4OWFiY2RlZmdoaWprbG1ub3BxcnN0dXZ3eHl6AAAAADAxMjM0NTY3ODlhYmNk ZWZnaGlqa2xtbm9wcXJzdHV2d3h5egAAAABTdGFjayBPdmVyZmxvdyENCgAAAGBaQABgWkAA AAAAAAAAAKACQAAAAAAAAADIBUAAAAAAAABAnAxAAAAAAAAgvL4ZQAAAAAS/yRuONECetXAr qK3FnWlA1abP/0kfeMLTQOCM6YDJR7qTqEGO3vmd++t+qlFDx5EOpq6gGeOjRhcMdYGGdXbJ SE3lXT3FXTuLnpJam5cgigJSYMQldQAAAAAAAACA/38AAAAAUMlAAFTJQABcyUAAZMlAAGzJ QAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABQAAAABAAAA AgAAAAIAAAADAAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAHDSQAD//////////wEAAAD/////AgAAAP////8BAAAA /////wIAAAAdAADAAgAAADoBAMACAAAABQAAwAIAAAA6AQDAAgAAADoBAMABAAAA/////wEA AAD/////AQAAAP////8CAAAAlAAAwAEAAACVAADAAAAAAAEAAAAACQEBCwYEBQUFAgP//wcJ BggB/wAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAANIhAACCJQACAe0AAAQAAAAAAAQA4h0AAAAAAAP///39VVVVV ////PzMzMzOqqqoqJEmSJP///x8cx3EcmZmZGXTRRRdVVVUVEzuxE5IkSRIRERER////Dw8P Dw+O4zgOXkN5DczMzAwMwzAMuuiiCyxkIQuqqqoKo3A9Comd2AleQnsJSZIkCbDc0wiIiIgI hBBCCP///wd88MEHh4eHBwd1UAfHcRwHAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACAgICAgICAgICAgICAgI CAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAAgMDAwM DAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMCAoKCgoKCQkJCQkJCQkJCQkJCQkJCQkJCQkJ CQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQwMDAwMDAwMDAwMDAwMDAwM DAwMDAwMDAwMDAwMAAAAAAAg0BhAAAAB8BZAAAACPBlAAAAgfBdAAAAgsBtAAAAgKHJAAAAg AFVAAAAg0FlAAAAgEHhAAAAB8RZAAAAgUBxAAAAg0CBAAAAfbEVAAAAKzEpAAAAgDHNAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEAAAeAEAAB0w LDA5MFAwZDBzMIUwnzClMKswsTC3ML0wyjDnMAMxCTEUMTAxUDGGMZMxoDGtMboxxzHUMeEx ADIGMgwyEjI2MjwyRzJyMngygTKwMrYyvjLHMgMzCTMPMyEzLjNCM08zVTNnM3QzjjOeM7gz yjPQM9gz3jPkM+oz8DP3M/4zBTQMNBM0IDQmNE40vTT8NFc1XTV1NXs1gDWGNYw1qzWxNbY1 vDXCNc411zXhNfQ1/zUFNgo2EDYWNhw2QzZ1NjU3QTd+N4M3rzfHNxg4WTiEOJU4oTjTONg4 3jjkOAA5CDkNORM5GTkhOTI5PzlJOWs5dDl6OZ45HDpyOho7NjtiO7Y7wTvKO9A71TsBPA88 NTxHPHU8fjyjPMw85DwSPXM9sj3GPeI97T0TPh0+Mj4+PkU+Sz5SPls+bj58PoE+hj6YPqA+ qD6yPrk+vz7GPss+0T7dPug+7z79PjA/Sz9WP10/dD+5P9s/5z/sP/o/ACAAADgAAAAAMAww ETAXMB4wKTBFMFUwWzB3MJAwmTCpMMIwyjDWMOkw7zACMfY49z4NP90/AAAAMAAA3AAAANQw AjEJMSkxMTFRMVoxejGCMaYxrDG0Mb8x9jEQMiMyKjI1MjoydzM2ND00STRQNFw1bTVyNYc1 jDWTNZg1uTW/Ncc1zTW5NsI3/zgZOX45kzmYOaQ5rzm3Ob05yzngOfQ5ADoHOiE6PTpFOk86 cDp1Ooc6kTqfOrE6uzrNOus68jr4Ov06FDsaOyA7aD1tPXM9jT2ZPaM9sz3APcs93z38PQw+ Ej4aPh8+JT5yPn8+hz6VPps+qD6xPs4+1j4SPxg/Uj9eP5c/nT+0P78/xT/tPwAAAEAAAMgB AAAgMCkwMTBFMFowfjCDMJMwmzCoMMYwyzDVMPAwCzJ3MoQypTK6Ms4y7TIFMwozEDMgMzUz SzNcM2IzdzOGM5szoTOvM7QzwDPGM8sz4jP1MyQ0KjQ3NFQ0hTSLNJM0mTSfNLE0uzTKNNc0 5zQFNSM1TTVZNWA1cDWDNZo1qzW5Ndw15zXtNfo1CDYSNhg2JDYsNj42STZSNlg2cDZ9Nog2 jTacNqk20TbdNu82/zYaNy03NTdBN0k3VTddN2U3djeKN6g3tzfJN+43/zcFOA84MThLOFw4 fTiJOJA4nziyOLk4zjjdOBY5Hjk1OUI5XzlnOXc5lDmhOac5uznAOcU5yjnPOdQ52TnfOeU5 6znxOfc5/TkCOgc6DDoROhY6GzogOiU6KzoxOjc6PTpDOkk6TjpTOl06YjpnOm06eDp9OoM6 iTqPOps6pjqsOrQ6ujq+Os861DraOus69Dr6Og47FDsZOx87JDsqOy87NTtdO3U7mDu6O4k8 lzyxPOQ86zz2PP48YD0zPj0+Rz5RPls+ZT5vPnk+gz6SPqc+sj69PsQ+yz7SPtk+5z4aPyU/ PD9AP0Q/SD9MP1A/VD+EPwAAAFAAANAAAABkMIAwoTC1MNAw1jDcMAAxDTEmMTYxQTFNMVMx XjFlMW8xgzGgMaYx8DEuMkwyYzJpMnMyfzKHMo0yADMIM0Yz7jOrNPI0MTVNNVc1cDV9NZA1 mzWnNbE1zDXYNfw1HDY2Nlo2ZjZwNp42uja/Nso24DYCNwo3Gjc0Nzs3QTdcN3Q3hjebN9A3 2zftN/M3CThFOEs4VzhrOHE4djioOLM4uDjPON04wznuOQI6ajpwOpU6mzqhOqo60TrZO907 4TvlO+k7fDwAAABgAABYAAAAPTJBMkUySTJONHk3iDeQN+o4JTtNO6U7tTvnO4c80jwFPR09 Ij2FPcY9Jz4zPmQ+ez6HPq0+xT7ePh8/Pj9eP3c/fT+OP5w/pD+yP/c/AAAAcAAAEAEAABAw MjA8MFIwYDAqMVYxWzGBMaExsjG+McUxyzHQMdYx4DEDMjAyPjJlMoQy2jLgMusy/zIYMysz MTNEM0ozVjPeM+Qz/jMFNAo0EDQqNDE0NjQ+NE40WjSHNI00nzSpNMg00DTsNPQ0wDXKNdE1 2zXjNfA1+jUBNg02FDaeNqI2+Db8NgA3BDcINww3EDcUNxg3HDcgNyQ3QDekN7Q3wzfIN843 1DcTOBg4HjgkODc4PThIOFQ4YjhtOHo4iziyOMU42zjsOPI4BjkbOTA5RzlMOWU5gTmUOZ45 tzneOeM59DlOO2k7hjuOO5Y7pjusO7o7vzvNO9I7Vj1cPWI9aD2IPeA94D4AAACAAADIAAAA OzBAMGMwdDCJMI4wljCwMN0w4zDuMCUxQjFOMWYxzTH8MQ0ySTLrMgYzDzMdMyczRjN8M4Ez nzMcNCk0ijTMNAo1IDVaNXo1fTaWNqY2rza5Ntw29TYNNxs3Mjc/N103szfEN9M33zf2Nwk4 FzhCOGg4dziKOKM4vjjOOO44AjkOOSQ5LzlFOV85bDnzOQA6IjoAOwg7HjsvO1074jsMPCc8 Lzw9PKA8JD1cPZw96z02Pm4+gD7fPuU+Cz95PwAAAJAAAHwAAAAUMCIyaTIwM2YzdDOrM9gz /DMuND80UjR2NJc00zT/NBg1IjU0NXY1kzWgNdI14jVaNmc2gzlzOqU6tTrbOho7LzteO287 dTuHO6s77TsZPDI8PDxOPJY8Az0QPV09aT13PZc9Bz4UPkE+ST6wPr0+bz+mPwCgAACIAAAA ZjCDMJcwwjDLMOYwWDJSMkwyRjJAMjoyNDIuMigyIjIcMhYyEDIKMgQy/jH4MfIx7DHmMeAx 2jHUMc4xyDHCMbwxtjGwMaoxpDGeMZgxkjGMMYYxgDF6MXQxbjFoMWIxXDFWMVAxSjFEMT4x ODEyMSwxJjEgMRoxFDEOMQgxAjEAwAAASAAAAHA6dDp4Onw6Gz0kPSg9LD0wPTQ9OD08PUA9 RD1IPUw9UD1UPVg9XD1gPWQ9aD10PXg9fD2APdQ+2D4EPwg/DD8A0AAAQAAAAEQ1SjV6NVA1 VjWANVw1hjWMNZI1mDViNWg1bjW8MMAwVDFYMVwxYDFkMcAydDWUM5gznDOoMwAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAA --Message-Boundary-10580 Content-type: text/plain; charset=US-ASCII Content-disposition: inline Content-description: Attachment information. The following section of this message contains a file attachment prepared for transmission using the Internet MIME message format. If you are using Pegasus Mail, or any another MIME-compliant system, you should be able to save it or view it from within your mailer. If you cannot, please ask your system administrator for assistance. ---- File information ----------- File: stream_d.c Date: 17 Dec 1999, 4:31 Size: 6763 bytes. Type: Program-source --Message-Boundary-10580 Content-type: Application/Octet-stream; name="stream_d.c"; type=Program-source Content-disposition: attachment; filename="stream_d.c" Content-transfer-encoding: BASE64 IyBpbmNsdWRlIDxzdGRpby5oPg0KIyBpbmNsdWRlIDxtYXRoLmg+DQojIGluY2x1ZGUgPGZs b2F0Lmg+DQojIGluY2x1ZGUgPGxpbWl0cy5oPg0KIyBpbmNsdWRlIDx0aW1lLmg+DQoNCi8q DQogKiBQcm9ncmFtOiBTdHJlYW0NCiAqIFByb2dyYW1tZXI6IEpvZSBSLiBaYWdhcg0KICog UmV2aXNpb246IDQuMC1CRVRBLCBPY3RvYmVyIDI0LCAxOTk1DQogKiBPcmlnaW5hbCBjb2Rl IGRldmVsb3BlZCBieSBKb2huIEQuIE1jQ2FscGluDQogKg0KICogVGhpcyBwcm9ncmFtIG1l YXN1cmVzIG1lbW9yeSB0cmFuc2ZlciByYXRlcyBpbiBNQi9zIGZvciBzaW1wbGUgDQogKiBj b21wdXRhdGlvbmFsIGtlcm5lbHMgY29kZWQgaW4gQy4gIFRoZXNlIG51bWJlcnMgcmV2ZWFs IHRoZSBxdWFsaXR5DQogKiBvZiBjb2RlIGdlbmVyYXRpb24gZm9yIHNpbXBsZSB1bmNhY2hl YWJsZSBrZXJuZWxzIGFzIHdlbGwgYXMgc2hvd2luZw0KICogdGhlIGNvc3Qgb2YgZmxvYXRp bmctcG9pbnQgb3BlcmF0aW9ucyByZWxhdGl2ZSB0byBtZW1vcnkgYWNjZXNzZXMuDQogKg0K ICogSU5TVFJVQ1RJT05TOg0KICoNCiAqCTEpIFN0cmVhbSByZXF1aXJlcyBhIGdvb2QgYml0 IG9mIG1lbW9yeSB0byBydW4uICBBZGp1c3QgdGhlDQogKiAgICAgICAgICB2YWx1ZSBvZiAn TicgKGJlbG93KSB0byBnaXZlIGEgJ3RpbWluZyBjYWxpYnJhdGlvbicgb2YgDQogKiAgICAg ICAgICBhdCBsZWFzdCAyMCBjbG9jay10aWNrcy4gIFRoaXMgd2lsbCBwcm92aWRlIHJhdGUg ZXN0aW1hdGVzDQogKiAgICAgICAgICB0aGF0IHNob3VsZCBiZSBnb29kIHRvIGFib3V0IDUl IHByZWNpc2lvbi4NCiAqLw0KDQojIGRlZmluZSBOTgkoMTAwMDAwMCBeIDQwOTYgXiAxMDI0 IF4gNTEyIF4gMjU2IF4gMTI4IF4gNjQpDQojIGRlZmluZSBOCSgxMDAwMDAwKQ0KIyBkZWZp bmUgTlRJTUVTCTEwDQojIGRlZmluZSBPRkZTRVQJKE5OLU4rNjU1MzYpDQoNCi8qDQogKgkz KSBDb21waWxlIHRoZSBjb2RlIHdpdGggZnVsbCBvcHRpbWl6YXRpb24uICBNYW55IGNvbXBp bGVycw0KICoJICAgZ2VuZXJhdGUgdW5yZWFzb25hYmx5IGJhZCBjb2RlIGJlZm9yZSB0aGUg b3B0aW1pemVyIHRpZ2h0ZW5zDQogKgkgICB0aGluZ3MgdXAuICBJZiB0aGUgcmVzdWx0cyBh cmUgdW5yZWFzb25hYmx5IGdvb2QsIG9uIHRoZQ0KICoJICAgb3RoZXIgaGFuZCwgdGhlIG9w dGltaXplciBtaWdodCBiZSB0b28gc21hcnQgZm9yIG1lIQ0KICoNCiAqICAgICAgICAgVHJ5 IGNvbXBpbGluZyB3aXRoOg0KICogICAgICAgICAgICAgICBjYyAtTyBzdHJlYW1fZC5jIHNl Y29uZC5jIC1vIHN0cmVhbV9kIC1sbQ0KICoNCiAqICAgICAgICAgVGhpcyBpcyBrbm93biB0 byB3b3JrIG9uIENyYXksIFNHSSwgSUJNLCBhbmQgU3VuIG1hY2hpbmVzLg0KICoNCiAqDQog Kgk0KSBNYWlsIHRoZSByZXN1bHRzIHRvIG1jY2FscGluQGNzLnZpcmdpbmlhLmVkdQ0KICoJ ICAgQmUgc3VyZSB0byBpbmNsdWRlOg0KICoJCWEpIGNvbXB1dGVyIGhhcmR3YXJlIG1vZGVs IG51bWJlciBhbmQgc29mdHdhcmUgcmV2aXNpb24NCiAqCQliKSB0aGUgY29tcGlsZXIgZmxh Z3MNCiAqCQljKSBhbGwgb2YgdGhlIG91dHB1dCBmcm9tIHRoZSB0ZXN0IGNhc2UuDQogKiBU aGFua3MhDQogKg0KICovDQoNCiMgZGVmaW5lIEhMSU5FICItLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tXG4iDQoNCiMgaWZu ZGVmIE1JTg0KIyBkZWZpbmUgTUlOKHgseSkgKCh4KTwoeSk/KHgpOih5KSkNCiMgZW5kaWYN CiMgaWZuZGVmIE1BWA0KIyBkZWZpbmUgTUFYKHgseSkgKCh4KT4oeSk/KHgpOih5KSkNCiMg ZW5kaWYNCg0KLypzdGF0aWMqLyBkb3VibGUJYVtOK09GRlNFVF0sDQoJCWJbTitPRkZTRVRd LA0KCQljW04rT0ZGU0VUXTsNCg0Kc3RhdGljIGRvdWJsZQlybXN0aW1lWzRdID0gezB9LCBt YXh0aW1lWzRdID0gezB9LA0KCQltaW50aW1lWzRdID0ge0ZMVF9NQVgsRkxUX01BWCxGTFRf TUFYLEZMVF9NQVh9Ow0KDQpzdGF0aWMgY2hhcgkqbGFiZWxbNF0gPSB7IkNvcHk6ICAgICAg IiwgIlNjYWxlOiAgICAgIiwNCiAgICAiQWRkOiAgICAgICAiLCAiVHJpYWQ6ICAgICAifTsN Cg0Kc3RhdGljIGRvdWJsZQlieXRlc1s0XSA9IHsNCiAgICAyICogc2l6ZW9mKGRvdWJsZSkg KiBOLA0KICAgIDIgKiBzaXplb2YoZG91YmxlKSAqIE4sDQogICAgMyAqIHNpemVvZihkb3Vi bGUpICogTiwNCiAgICAzICogc2l6ZW9mKGRvdWJsZSkgKiBODQogICAgfTsNCg0KZXh0ZXJu IGRvdWJsZSBzZWNvbmQoKTsNCg0KaW50DQptYWluKCkNCiAgICB7DQogICAgaW50CQkJcXVh bnR1bSwgY2hlY2t0aWNrKCk7DQogICAgaW50CQkJQnl0ZXNQZXJXb3JkOw0KICAgIHJlZ2lz dGVyIGludAlqLCBrOw0KICAgIGRvdWJsZQkJc2NhbGFyLCB0Ow0KICAgIHZvbGF0aWxlIGRv dWJsZQl0aW1lc1s0XVtOVElNRVNdOw0KDQovLyAgICBwcmludGYoImEgPSAlMDh4XG5iID0g JTA4eFxuYyA9ICUwOHhcbiIsIGEsIGIsIGMgKTsNCg0KICAgIC8qIC0tLSBTRVRVUCAtLS0g ZGV0ZXJtaW5lIHByZWNpc2lvbiBhbmQgY2hlY2sgdGltaW5nIC0tLSAqLw0KDQogICAgcHJp bnRmKEhMSU5FKTsNCiAgICBCeXRlc1BlcldvcmQgPSBzaXplb2YoZG91YmxlKTsNCiAgICBw cmludGYoIlRoaXMgc3lzdGVtIHVzZXMgJWQgYnl0ZXMgcGVyIERPVUJMRSBQUkVDSVNJT04g d29yZC5cbiIsDQoJQnl0ZXNQZXJXb3JkKTsNCg0KICAgIHByaW50ZihITElORSk7DQogICAg cHJpbnRmKCJBcnJheSBzaXplID0gJWQsIE9mZnNldCA9ICVkXG4iICwgTiwgT0ZGU0VUKTsN CiAgICBwcmludGYoIlRvdGFsIG1lbW9yeSByZXF1aXJlZCA9ICUuMWYgTUIuXG4iLA0KCSgz ICogTiAqIEJ5dGVzUGVyV29yZCkgLyAxMDQ4NTc2LjApOw0KICAgIHByaW50ZigiRWFjaCB0 ZXN0IGlzIHJ1biAlZCB0aW1lcywgYnV0IG9ubHlcbiIsIE5USU1FUyk7DQogICAgcHJpbnRm KCJ0aGUgKmJlc3QqIHRpbWUgZm9yIGVhY2ggaXMgdXNlZC5cbiIpOw0KDQogICAgLyogR2V0 IGluaXRpYWwgdmFsdWUgZm9yIHN5c3RlbSBjbG9jay4gKi8NCg0KICAgIGZvciAoaj0wOyBq PE47IGorKykgew0KCWFbal0gPSAxLjA7DQoJYltqXSA9IDIuMDsNCgljW2pdID0gMC4wOw0K CX0NCg0KICAgIHByaW50ZihITElORSk7DQoNCiAgICBpZiAgKCAocXVhbnR1bSA9IGNoZWNr dGljaygpKSA+PSAxKSANCglwcmludGYoIllvdXIgY2xvY2sgZ3JhbnVsYXJpdHkvcHJlY2lz aW9uIGFwcGVhcnMgdG8gYmUgIg0KCSAgICAiJWQgbWljcm9zZWNvbmRzLlxuIiwgcXVhbnR1 bSk7DQogICAgZWxzZQ0KCXByaW50ZigiWW91ciBjbG9jayBncmFudWxhcml0eSBhcHBlYXJz IHRvIGJlICINCgkgICAgImxlc3MgdGhhbiBvbmUgbWljcm9zZWNvbmQuXG4iKTsNCg0KICAg IHQgPSBzZWNvbmQoKTsNCiAgICBmb3IgKGogPSAwOyBqIDwgTjsgaisrKQ0KCWFbal0gPSAy LjBFMCAqIGFbal07DQogICAgdCA9IDEuMEU2ICogKHNlY29uZCgpIC0gdCk7DQoNCiAgICBw cmludGYoIkVhY2ggdGVzdCBiZWxvdyB3aWxsIHRha2Ugb24gdGhlIG9yZGVyIg0KCSIgb2Yg JWQgbWljcm9zZWNvbmRzLlxuIiwgKGludCkgdCAgKTsNCiAgICBwcmludGYoIiAgICg9ICVk IGNsb2NrIHRpY2tzKVxuIiwgKGludCkgKHQvcXVhbnR1bSkgKTsNCiAgICBwcmludGYoIklu Y3JlYXNlIHRoZSBzaXplIG9mIHRoZSBhcnJheXMgaWYgdGhpcyBzaG93cyB0aGF0XG4iKTsN CiAgICBwcmludGYoInlvdSBhcmUgbm90IGdldHRpbmcgYXQgbGVhc3QgMjAgY2xvY2sgdGlj a3MgcGVyIHRlc3QuXG4iKTsNCg0KICAgIHByaW50ZihITElORSk7DQoNCiAgICBwcmludGYo IldBUk5JTkcgLS0gVGhlIGFib3ZlIGlzIG9ubHkgYSByb3VnaCBndWlkZWxpbmUuXG4iKTsN CiAgICBwcmludGYoIkZvciBiZXN0IHJlc3VsdHMsIHBsZWFzZSBiZSBzdXJlIHlvdSBrbm93 IHRoZVxuIik7DQogICAgcHJpbnRmKCJwcmVjaXNpb24gb2YgeW91ciBzeXN0ZW0gdGltZXIu XG4iKTsNCiAgICBwcmludGYoSExJTkUpOw0KICAgIA0KICAgIC8qCS0tLSBNQUlOIExPT1Ag LS0tIHJlcGVhdCB0ZXN0IGNhc2VzIE5USU1FUyB0aW1lcyAtLS0gKi8NCg0KICAgIC8qIFRo aXMgbWFpbiBsb29wIGlzIGEgcGllY2Ugb2Ygc2hpdCAtLSB0aGUgV0FUQ09NIEMvQysrIGNv bXBpbGVyIGhhcyBhDQogICAgICAgZmllbGQgZGF5IHdpdGggaXQuICAqLw0KDQogICAgc2Nh bGFyID0gMy4wOw0KICAgIGZvciAoaz0wOyBrPE5USU1FUzsgaysrKQ0KCXsNCg0KICAgICAg ICAvKiBEZXBlbmRpbmcgb24gdGhlIG9wdGltaXphdGlvbiBzZXR0aW5ncywgV0FUQ09NIEMv QysrIHdpbGwgamFtDQogICAgICAgICAgIGVpdGhlciB0aGF0IGZpcnN0IHRoZSBsb29wIHdp dGggdGhlIHNlY29uZCBvciB2aWNlIHZlcnNhLiAgVGhlDQogICAgICAgICAgIHJlc3VsdCBp cyB0aGF0IHRoZSAic2Vjb25kKCkiIGNhbGwgaXMgc2ltcGx5IGluc3RhbnRseSBzdWNjZXNz aXZlDQogICAgICAgICAgIGNhdXNpbmcgYSBkaXZpZGUgYnkgMC4gIEFsbCBiZWNhdXNlIGEs YixjIHdlcmUgZGVjbGFyZWQgc3RhdGljLiAgDQogICAgICAgICAgIFJlYWwgZnVja2luZyB1 c2VmdWwgKi8NCg0KCXRpbWVzWzBdW2tdID0gc2Vjb25kKCk7DQoNCiNpZmRlZiBBU01IQUNL DQogICAgICAgIF9hc20gew0KICAgICAgICAgICAgIHhvciAgICAgICBlY3gsZWN4DQogICAg ICAgICAgTDE6DQogICAgICAgICAgICAgYWRkICAgICAgIGVjeCw4DQogICAgICAgICAgICAg ZmxkICAgICAgIHF3b3JkIHB0ciBhW2VjeC04XQ0KICAgICAgICAgICAgIGZzdHAgICAgICBx d29yZCBwdHIgY1tlY3gtOF0NCiAgICAgICAgICAgICBjbXAgICAgICAgZWN4LDB4MDA3YTEy MDANCiAgICAgICAgICAgICBqbmUgICAgICAgTDENCiAgICAgICAgfTsNCiNlbHNlDQoJZm9y IChqPTA7IGo8TjsgaisrKQ0KCSAgICBjW2pdID0gYVtqXTsNCiNlbmRpZg0KCXRpbWVzWzBd W2tdID0gc2Vjb25kKCkgLSB0aW1lc1swXVtrXTsNCgkNCgl0aW1lc1sxXVtrXSA9IHNlY29u ZCgpOw0KCWZvciAoaj0wOyBqPE47IGorKykNCgkgICAgYltqXSA9IHNjYWxhcipjW2pdOw0K CXRpbWVzWzFdW2tdID0gc2Vjb25kKCkgLSB0aW1lc1sxXVtrXTsNCgkNCgl0aW1lc1syXVtr XSA9IHNlY29uZCgpOw0KCWZvciAoaj0wOyBqPE47IGorKykNCgkgICAgY1tqXSA9IGFbal0r YltqXTsNCgl0aW1lc1syXVtrXSA9IHNlY29uZCgpIC0gdGltZXNbMl1ba107DQoJDQoJdGlt ZXNbM11ba10gPSBzZWNvbmQoKTsNCglmb3IgKGo9MDsgajxOOyBqKyspDQoJICAgIGFbal0g PSBiW2pdK3NjYWxhcipjW2pdOw0KCXRpbWVzWzNdW2tdID0gc2Vjb25kKCkgLSB0aW1lc1sz XVtrXTsNCgl9DQogICAgDQogICAgLyoJLS0tIFNVTU1BUlkgLS0tICovDQoNCiAgICBmb3Ig KGs9MDsgazxOVElNRVM7IGsrKykNCgl7DQoJZm9yIChqPTA7IGo8NDsgaisrKQ0KCSAgICB7 DQoJICAgIHJtc3RpbWVbal0gPSBybXN0aW1lW2pdICsgKHRpbWVzW2pdW2tdICogdGltZXNb al1ba10pOw0KCSAgICBtaW50aW1lW2pdID0gTUlOKG1pbnRpbWVbal0sIHRpbWVzW2pdW2td KTsNCgkgICAgbWF4dGltZVtqXSA9IE1BWChtYXh0aW1lW2pdLCB0aW1lc1tqXVtrXSk7DQoJ ICAgIH0NCgl9DQoNCi8vICAgIHQgPSAwOw0KDQogICAgcHJpbnRmKCJGdW5jdGlvbiAgICAg IFJhdGUgKE1CL3MpICAgUk1TIHRpbWUgICAgIE1pbiB0aW1lICAgICBNYXggdGltZVxuIik7 DQogICAgZm9yIChqPTA7IGo8NDsgaisrKSB7DQoJcm1zdGltZVtqXSA9IHNxcnQocm1zdGlt ZVtqXS8oZG91YmxlKU5USU1FUyk7DQoNCi8vICAgICAgICB0ICs9IDEuMEUtMDYgKiBieXRl c1tqXS9taW50aW1lW2pdOw0KDQoJcHJpbnRmKCIlcyUxMS40ZiAgJTExLjRmICAlMTEuNGYg ICUxMS40ZlxuIiwgbGFiZWxbal0sDQoJICAgICAgIDEuMEUtMDYgKiBieXRlc1tqXS9taW50 aW1lW2pdLA0KCSAgICAgICBybXN0aW1lW2pdLA0KCSAgICAgICBtaW50aW1lW2pdLA0KCSAg ICAgICBtYXh0aW1lW2pdKTsNCiAgICB9DQoNCi8vICAgIHByaW50ZigiQXZlcmFnZSBvZiBT dHJlYW0gc2NvcmVzOiAlMTEuNGZcbiIsdC80LjApOw0KICAgIHJldHVybiAwOw0KfQ0KDQoj IGRlZmluZQlNCTIwDQoNCmludA0KY2hlY2t0aWNrKCkNCiAgICB7DQogICAgaW50CQlpLCBt aW5EZWx0YSwgRGVsdGE7DQogICAgZG91YmxlCXQxLCB0MiwgdGltZXNmb3VuZFtNXTsNCg0K LyogIENvbGxlY3QgYSBzZXF1ZW5jZSBvZiBNIHVuaXF1ZSB0aW1lIHZhbHVlcyBmcm9tIHRo ZSBzeXN0ZW0uICovDQoNCiAgICBmb3IgKGkgPSAwOyBpIDwgTTsgaSsrKSB7DQoJdDEgPSBz ZWNvbmQoKTsNCgl3aGlsZSggKCh0Mj1zZWNvbmQoKSkgLSB0MSkgPCAxLjBFLTYgKQ0KCSAg ICA7DQoJdGltZXNmb3VuZFtpXSA9IHQxID0gdDI7DQoJfQ0KDQovKg0KICogRGV0ZXJtaW5l IHRoZSBtaW5pbXVtIGRpZmZlcmVuY2UgYmV0d2VlbiB0aGVzZSBNIHZhbHVlcy4NCiAqIFRo aXMgcmVzdWx0IHdpbGwgYmUgb3VyIGVzdGltYXRlIChpbiBtaWNyb3NlY29uZHMpIGZvciB0 aGUNCiAqIGNsb2NrIGdyYW51bGFyaXR5Lg0KICovDQoNCiAgICBtaW5EZWx0YSA9IDEwMDAw MDA7DQogICAgZm9yIChpID0gMTsgaSA8IE07IGkrKykgew0KCURlbHRhID0gKGludCkoIDEu MEU2ICogKHRpbWVzZm91bmRbaV0tdGltZXNmb3VuZFtpLTFdKSk7DQoJbWluRGVsdGEgPSBN SU4obWluRGVsdGEsIE1BWChEZWx0YSwwKSk7DQoJfQ0KDQogICAgcmV0dXJuKG1pbkRlbHRh KTsNCiAgICB9DQoNCg== --Message-Boundary-10580-- From - Sun Nov 21 09:30:22 1999 Received: from aux153.plano.net ([209.217.36.11]) by mail.austin.rr.com with Microsoft SMTPSVC(5.5.1877.197.19); Wed, 17 Nov 1999 10:49:08 -0600 Received: (qmail 24374 invoked by uid 10000); 17 Nov 1999 16:54:43 -0000 Delivered-To: mccalpin.com-john@mccalpin.com Received: (qmail 24370 invoked from network); 17 Nov 1999 16:54:42 -0000 Received: from ares.cs.virginia.edu (128.143.136.41) by aux153.plano.net with SMTP; 17 Nov 1999 16:54:42 -0000 Received: from mailgw2a.lmco.com (mailgw2a.lmco.com [192.91.147.7]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id LAA05993 for ; Wed, 17 Nov 1999 11:48:36 -0500 (EST) Received: from emss03g01.ems.lmco.com (emss03g01.ems.lmco.com [141.240.4.144]) by mailgw2a.lmco.com (8.8.8/8.8.8) with ESMTP id LAA22922 for ; Wed, 17 Nov 1999 11:48:31 -0500 (EST) Return-receipt-to: larry.albertelli@lmco.com Received: from CONVERSION-DAEMON by lmco.com (PMDF V5.2-32 #38888) id <0FLC00001MLR0A@lmco.com> for mccalpin@cs.virginia.edu; Wed, 17 Nov 1999 11:47:15 -0500 (EST) Received: from emss35m01.ems.lmco.com ([158.187.107.140]) by lmco.com (PMDF V5.2-32 #38888) with ESMTP id <0FLC00J20ND0MI@lmco.com> for mccalpin@cs.virginia.edu; Wed, 17 Nov 1999 11:05:34 -0500 (EST) Received: by emss35m01.owg.fs.lmco.com with Internet Mail Service (5.5.2580.0) id ; Wed, 17 Nov 1999 11:06:29 -0500 Content-return: allowed Date: Wed, 17 Nov 1999 11:05:18 -0500 From: "Albertelli, Larry" Subject: STREAM Results To: "'mccalpin@cs.virginia.edu'" Message-id: <99AA2270B1E6D111BCE10000F805F17F058F33C5@emss35m02.owg.fs.lmco.com> MIME-version: 1.0 X-Mailer: Internet Mail Service (5.5.2580.0) Content-type: MULTIPART/MIXED; BOUNDARY="Boundary_(ID_QOFUfqR6ADb64esGL987iw)" Return-Path: larry.albertelli@lmco.com X-Mozilla-Status: 8001 X-Mozilla-Status2: 00400000 X-UIDL: 35000000.eml Status: RO This message is in MIME format. Since your mail reader does not understand this format, some or all of this message may not be legible. --Boundary_(ID_QOFUfqR6ADb64esGL987iw) Content-type: text/plain Content-transfer-encoding: 7BIT The text file contains results for the new NT version of streamd run under DOS with 1000 reps. Intel Outrigger motherboard with dual Rambus, 2x600MHz EB CPUs, 256MB 133MHz RAM. <> Larry Albertelli Lockheed-Martin Federal Systems Division M/S 0315 1801 Rte.17C Owego, NY 13827 Phone (607)751-2027 FAX (607)751-3621 --Boundary_(ID_QOFUfqR6ADb64esGL987iw) Content-type: text/plain; name=STREAM.TXT Content-disposition: attachment; filename=STREAM.TXT Content-transfer-encoding: 7BIT STREAM for DOS v2 by Dennis Lee =============================== 1 MB = 1000000 Bytes in the following measurements. For accurate results, this benchmark should be executed in a true DOS session, and not a DOS shell under another OS. Time Operation Mem Speed Error ---- --------- --------- ----- 14.12 sec COPY32 453.26 MB/s 0.4% 15.05 sec COPY64 425.25 MB/s 0.4% 15.54 sec SCALE 411.84 MB/s 0.4% 16.86 sec ADD 569.40 MB/s 0.4% 22.52 sec TRIAD 426.29 MB/s 0.3% These results are comparable with those on the STREAM website. See for info on STREAM. Type 'streamd ?' for help --Boundary_(ID_QOFUfqR6ADb64esGL987iw)-- From - Sat Dec 11 07:32:41 1999 Received: from aux153.plano.net ([209.217.36.11]) by mail.austin.rr.com with Microsoft SMTPSVC(5.5.1877.197.19); Wed, 24 Nov 1999 18:54:20 -0600 Received: (qmail 15668 invoked by uid 10000); 25 Nov 1999 00:59:50 -0000 Delivered-To: mccalpin.com-john@mccalpin.com Received: (qmail 15658 invoked from network); 25 Nov 1999 00:59:50 -0000 Received: from ares.cs.virginia.edu (128.143.136.41) by aux153.plano.net with SMTP; 25 Nov 1999 00:59:50 -0000 Received: from gatekeeper.collins.rockwell.com (firewall-user@gatekeeper.collins.rockwell.com [205.175.225.1]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id TAA17176 for ; Wed, 24 Nov 1999 19:54:04 -0500 (EST) From: rgvande5@collins.rockwell.com Received: by gatekeeper.collins.rockwell.com; id SAA24098; Wed, 24 Nov 1999 18:54:04 -0600 (CST) Received: from mnpcl1.collins.rockwell.com(131.198.67.150) by gatekeeper.collins.rockwell.com via smap (V4.2) id xma023963; Wed, 24 Nov 99 18:53:10 -0600 Received: from psnotes.collins.rockwell.com (psnotes [131.198.17.8]) by mnpcl1.collins.rockwell.com (8.9.1/8.9.1) with SMTP id SAA18525 for ; Wed, 24 Nov 1999 18:53:08 -0600 (CST) Received: by psnotes.collins.rockwell.com(Lotus SMTP MTA Internal build v4.6.2 (651.2 6-10-1998)) id 88256834.00056A22 ; Wed, 24 Nov 1999 16:59:08 -0800 X-Lotus-FromDomain: ROCKWELL To: mccalpin@cs.virginia.edu Message-ID: <88256834.00056951.00@psnotes.collins.rockwell.com> Date: Wed, 24 Nov 1999 16:49:38 -0800 Subject: Embedded PowerPC Benchmark Mime-Version: 1.0 Content-type: multipart/mixed; Boundary="0__=4za5ycH8FfCO9J1jGWZqghcs6d5ZNrzOGDH7RCM9dSm8K3W3fppxEuJa" Content-Disposition: inline Return-Path: rgvande5@collins.rockwell.com X-Mozilla-Status: 8001 X-Mozilla-Status2: 00000000 X-UIDL: 57000000.eml Status: RO --0__=4za5ycH8FfCO9J1jGWZqghcs6d5ZNrzOGDH7RCM9dSm8K3W3fppxEuJa Content-type: text/plain; charset=us-ascii Content-Disposition: inline The raw data is for a single Motorola MPC860 (a PowerPC with embedded peripherals) at 40MHz, believed to be running with no wait states and no non-clock events under OS-9000. The benchmark was ported to an N of 10,000 to fit into the embedded system's small memory. The port may help characterize CPU cores of embedded systems. ------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 10000, Offset = 0 Total memory required = 0.2 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 29999 microseconds. Each test below will take on the order of 5739999 microseconds. (= 191 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING -- The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Copy: 12.9032 1.2581 1.2400 1.2800 Scale: 2.3599 7.0577 6.7800 7.8000 Add: 4.6784 5.4139 5.1300 5.7000 Triad: 2.6002 9.6532 9.2300 10.8200 (See attached file: wstream.c)(See attached file: second_e.c)(See attached file: wstream.mak) --0__=4za5ycH8FfCO9J1jGWZqghcs6d5ZNrzOGDH7RCM9dSm8K3W3fppxEuJa Content-type: application/octet-stream; name="wstream.c" Content-Disposition: attachment; filename="wstream.c" Content-transfer-encoding: x-uuencode begin 644 wstream.c M(VEN8VQU9&4@/'1I;64N:#X-"B,@:6YC;'5D92`\0T*("H@;V8@8V]D92!G96YE2!T;R!R M=6XN("!!9&IUF%T:6]N+B`@36%N>2!C;VUP:6QE MF5R('1I9VAT96YS#0H@*B`@("`@("`@('1H M:6YG2!C;VUP:6QI;F<@=VET:#H-"B`J("`@("`@("`@("`@("`@8V,@+4\@ M"QY*2`H*'@I/"AY*3\H>"DZ*'DI*0T*(R!E M;F1I9@T*(R!I9FYD968@34%8#0HC(&1E9FEN92!-05@H>"QY*2`H*'@I/BAY M*3\H>"DZ*'DI*0T*(R!E;F1I9@T*#0IS=&%T:6,@9&]U8FQE("`@85M.*T]& M1E-%5%TL#0H@("`@("`@("`@("`@("`@8EM.*T]&1E-%5%TL#0H@("`@("`@ M("`@("`@("`@8UM.*T]&1E-%5%T[#0H-"G-T871I8R!D;W5B;&4@("!R;7-T M:6UE6S1=(#T@>S!]+"!M87AT:6UE6S1=(#T@>S!]+`T*("`@("`@("`@("`@ M("`@(&UI;G1I;65;-%T@/2![1DQ47TU!6"Q&3%1?34%8+$9,5%]-05@L1DQ4 M7TU!6'T[#0H-"G-T871I8R!C:&%R("`@("`J;&%B96Q;-%T@/2![(D-O<'DZ M("`@("`@(BP@(E-C86QE.B`@("`@(BP-"B`@("`B061D.B`@("`@("`B+"`B M5')I860Z("`@("`B?3L-"@T*PT*("`@(#(@*B!S:7IE;V8H9&]U8FQE*2`J($XL#0H@("`@,B`J('-I>F5O M9BAD;W5B;&4I("H@3BP-"B`@("`S("H@71EF4@/2`E9"P@3V9FPT*("`@("`@("!A6VI=(#T@,2XP.PT*("`@("`@("!B M6VI=(#T@,BXP.PT*("`@("`@("!C6VI=(#T@,"XP.PT*("`@("`@("!]#0H- M"B`@("!P2]PPT*("`@("`@("!F;W(@ M*&H],#L@:CPT.R!J*RLI#0H@("`@("`@("`@("![#0H@("`@("`@("`@("!R M;7-T:6UE6VI=(#T@"!T M:6UE7&XB*3L-"B`@("!F;W(@*&H],#L@:CPT.R!J*RLI('L-"B`@("`@("`@ MPT*("`@(&EN="`@("`@("`@(&DL(&UI M;D1E;'1A+"!$96QT83L-"B`@("!D;W5B;&4@("`@("!T,2P@=#(L('1I;65S M9F]U;F1;35T[#0H-"B\J("!#;VQL96-T(&$@PT*("`@("`@("!T,2`]('-E8V]N9"@I M.PT*("`@("`@("!W:&EL92@@*"AT,CUS96-O;F0H*2D@+2!T,2D@/"`Q+C!% M+38@*0T*("`@("`@("`@("`@.PT*("`@("`@("!T:6UEPT*("`@("`@("!$96QT82`]("AI;G0I*"`Q+C!%-B`J("AT M:6UE2X-"B`@(%1H:7,@:7,@82!V86QU86)L92!P;W)T(&9O7-T96US(&)E8V%UF5D('-I;F-E(`T*("`@2!C;VYV96YT:6]N M+"!I9B!C;&]C:R@I(')E='5R;G,@+3$L(&ET(&ES(&YO="!I;7!L96UE;G1E M9"X-"B`@(%!R;V=R86UM97(Z(%)A>2!686X@1&4@5V%L:V5R("@W,30I-#0T M+38Y,#,-"B`J+PT*#0HC:6YC;'5D92`\=&EM92YH/@T*#0ID;W5B;&4@#L-"GT-"@T*+RH@96YD(&]F(&9I;&4@*B\-"@T* ` end --0__=4za5ycH8FfCO9J1jGWZqghcs6d5ZNrzOGDH7RCM9dSm8K3W3fppxEuJa Content-type: application/octet-stream; name="wstream.mak" Content-Disposition: attachment; filename="wstream.mak" Content-transfer-encoding: x-uuencode begin 644 wstream.mak M(R`-"B,@1F%S5')A:R!'96YE96%S/2X@+69D/7=S=')E86TN; Wed, 17 Nov 1999 13:15:37 -0500 (EST) From: rgvande5@collins.rockwell.com Received: by gatekeeper.collins.rockwell.com; id MAA19458; Wed, 17 Nov 1999 12:15:33 -0600 (CST) Received: from mnpcl1.collins.rockwell.com(131.198.67.150) by gatekeeper.collins.rockwell.com via smap (V4.2) id xma018742; Wed, 17 Nov 99 12:14:48 -0600 Received: from psnotes.collins.rockwell.com (psnotes [131.198.17.8]) by mnpcl1.collins.rockwell.com (8.9.1/8.9.1) with SMTP id MAA03043 for ; Wed, 17 Nov 1999 12:14:45 -0600 (CST) Received: by psnotes.collins.rockwell.com(Lotus SMTP MTA Internal build v4.6.2 (651.2 6-10-1998)) id 8825682C.0064BFF5 ; Wed, 17 Nov 1999 10:20:27 -0800 X-Lotus-FromDomain: ROCKWELL To: mccalpin@cs.virginia.edu Message-ID: <8825682C.0064BF62.00@psnotes.collins.rockwell.com> Date: Wed, 17 Nov 1999 10:11:25 -0800 Subject: wstream code Mime-Version: 1.0 Content-type: multipart/mixed; Boundary="0__=wUNbNeufhEQb8CZ74eftjSFnPsk3Te9ipgUm5wynCpB3W6Q8OTcm83rU" Content-Disposition: inline Return-Path: rgvande5@collins.rockwell.com X-Mozilla-Status: 8001 X-Mozilla-Status2: 00000000 X-UIDL: 45000000.eml Status: RO --0__=wUNbNeufhEQb8CZ74eftjSFnPsk3Te9ipgUm5wynCpB3W6Q8OTcm83rU Content-type: text/plain; charset=us-ascii Content-Disposition: inline I ported it to OS-9000 (A real-time, POSIX-compliant OS). My benchmarks are for crippled embedded systems with slow memory and low power requirements, so I don't submit them. The second routine seems worthy of preservation because it uses a more restricted set of the C library. It should port easily to most embedded systems. (See attached file: second_e.c) --0__=wUNbNeufhEQb8CZ74eftjSFnPsk3Te9ipgUm5wynCpB3W6Q8OTcm83rU Content-type: application/octet-stream; name="second_e.c" Content-Disposition: attachment; filename="second_e.c" Content-transfer-encoding: x-uuencode begin 644 second_e.c M+RH@2X-"B`@(%1H:7,@:7,@82!V86QU86)L92!P;W)T(&9O7-T96US(&)E8V%UF5D('-I;F-E(`T*("`@2!C;VYV96YT:6]N M+"!I9B!C;&]C:R@I(')E='5R;G,@+3$L(&ET(&ES(&YO="!I;7!L96UE;G1E M9"X-"B`@(%!R;V=R86UM97(Z(%)A>2!686X@1&4@5V%L:V5R("@W,30I-#0T M+38Y,#,-"B`J+PT*#0HC:6YC;'5D92`\=&EM92YH/@T*#0ID;W5B;&4@#L-"GT-"@T*+RH@96YD(&]F(&9I;&4@*B\-"@T* ` end --0__=wUNbNeufhEQb8CZ74eftjSFnPsk3Te9ipgUm5wynCpB3W6Q8OTcm83rU-- From - Sun Nov 21 09:30:05 1999 Received: from aux153.plano.net ([209.217.36.11]) by Mail.austin.rr.com with Microsoft SMTPSVC(5.5.1875.185.18); Thu, 28 Oct 1999 17:02:31 -0500 Received: (qmail 4736 invoked by uid 10000); 28 Oct 1999 22:15:12 -0000 Delivered-To: mccalpin.com-john@mccalpin.com Received: (qmail 4682 invoked from network); 28 Oct 1999 22:15:07 -0000 Received: from ftp.cs.virginia.edu (HELO ares.cs.Virginia.EDU) (128.143.137.19) by aux153.plano.net with SMTP; 28 Oct 1999 22:15:07 -0000 Received: from mail.span.ch (mail.span.ch [144.85.10.50]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id SAA27117 for ; Thu, 28 Oct 1999 18:09:06 -0400 (EDT) Received: from super-7 (mp2-c82-p239.span.ch [195.15.82.239]) by mail.span.ch (8.9.3/8.9.3) with SMTP id AAA27583 for ; Fri, 29 Oct 1999 00:08:59 +0200 (MET DST) Message-ID: <001301bf219a$03776620$ef520fc3@super-7> From: "Armand Hirt" To: Subject: wstream Date: Fri, 29 Oct 1999 00:13:14 +0100 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_0010_01BF21A2.64693BE0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 4.72.3110.1 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Return-Path: armand.hirt@span.ch X-Mozilla-Status: 8001 X-Mozilla-Status2: 00000000 X-UIDL: 70000000.eml Status: RO This is a multi-part message in MIME format. ------=_NextPart_000_0010_01BF21A2.64693BE0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi, I got considerably lower results with wstream (your binary) than with = streamd2 on my computer (a K6-2 400, on a VIA-MVP 3 motherboard using = 100 MHZ SDRAM). Here are the figures: Johan, Here are the results with write allocation on: streamd (DOS) / wstream (W98) copy 145 / 138 scale 145 / 138 add 163 / 96 triad 163 / 78 And with write allocation off: Here are the results with WA off: streamd / wstream copy 104 / 101 scale 155 / 101 add 194 / 131 triad 194 / 130 In both cases, wstream figures are significantly lower. Why do you think = it is so? Regards, -- "S'arreter sur l'autoroute, c'est courrir un danger mortel. L'apparente = securite donne par la bande d'arret d'urgence est illusoire". Armand Hirt / 021 624 32 41 ------=_NextPart_000_0010_01BF21A2.64693BE0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Hi,
 
I got considerably lower results with wstream (your = binary)=20 than with streamd2 on my computer (a K6-2 400, on a VIA-MVP 3 = motherboard using=20 100 MHZ SDRAM).
 
Here are the figures:
 
Johan,
 
Here are the results with write allocation = on:
 
streamd (DOS) / wstream = (W98)
 
copy  145 / 138
scale 145 /=20 138
add   163 / 96
triad 163 / 78
 
And with write allocation off:
 
Here are the results with WA = off:
 
streamd / wstream
 
copy  104 / 101
scale 155 /=20 101
add   194 / 131
triad 194 / 130
 
In both cases, wstream figures are significantly = lower. Why do=20 you think it is so?
 
Regards,
 
--
"S'arreter sur = l'autoroute, c'est=20 courrir un danger mortel. L'apparente securite donne par la bande = d'arret=20 d'urgence est illusoire".
 
Armand Hirt / 021 624 32=20 41
------=_NextPart_000_0010_01BF21A2.64693BE0--