From fostertaz at access.ee.ntu.edu.tw Thu Oct 2 08:04:29 2008 From: fostertaz at access.ee.ntu.edu.tw (=?BIG5?B?pP2vRaZ0Zm9zdGVydGF6?=) Date: Thu, 2 Oct 2008 23:04:29 +0800 Subject: [Hotspot] thermal model on single function unit Message-ID: Hi, I'm triying to model the temperature of single function unit. The configuration is set as default. Floorplan and power trace are like below: ----Floorplan--- FU 2.03e-06 2.03e-06 0.00 0.00 ----ptrace--- FU 0.0041 ------------ The temperature turns out unreasonably high,5443K. Why does the result become like this? Then I found a thread similar to my question.([Hotspot] just want temperature for a single block) Suggestions of Prof. Skadron are very useful. I've put some silicon surrounding my function unit,forming a 1cm^2 chip. And the steady temperature became more reasonable,318K. But I'm still wondering about the reason of this floorplan. Moreover, I added some data to the ptrace file and got a ttrace file as follows. -----ptrace-------- silicon1 silicon2 silicon3 silicon4 FU 0.00 0.00 0.00 0.00 4.10e-03 0.00 0.00 0.00 0.00 4.10e-03 0.00 0.00 0.00 0.00 4.10e-03 0.00 0.00 0.00 0.00 4.10e-03 0.00 0.00 0.00 0.00 4.10e-03 -----ttrace--------- silicon1 silicon2 silicon3 silicon4 FU 279.83 400.94 385.36 279.83 272.11 278.6 400.26 386.03 278.6 271.56 277.43 399.72 386.58 277.43 270.93 276.32 399.27 387.02 276.32 270.27 275.27 398.89 387.38 275.27 269.58 (Are units of these temperatures centigrade?) ------------------- The temperature trace and the steady temperature are quit different. How can I solve this problem? Thanks for your patience to my four problems. Best Regards, Foster Wang From ks4kk at cs.virginia.edu Thu Oct 2 08:33:06 2008 From: ks4kk at cs.virginia.edu (Karthik Sankaranarayanan) Date: Thu, 2 Oct 2008 11:33:06 -0400 (EDT) Subject: [Hotspot] thermal model on single function unit In-Reply-To: References: Message-ID: The FU size (2 microns x 2 microns) and the power density (1000 W / mm^2) appear to be too "out-of-range". Were these the numbers you intended to use? If so, these are beyond what HotSpot is designed to model accurately. If there has been an error in the units, here are the units to be used: floorplan file: all dimensions are in meters power trace file: all values are in watts Hope this helps. -karthik On Thu, 2 Oct 2008, ???E?tfostertaz wrote: > Hi, > > I'm triying to model the temperature of single function unit. > The configuration is set as default. > Floorplan and power trace are like below: > ----Floorplan--- > FU 2.03e-06 2.03e-06 0.00 0.00 > ----ptrace--- > FU > 0.0041 > ------------ > The temperature turns out unreasonably high,5443K. > Why does the result become like this? > > Then I found a thread similar to my question.([Hotspot] just want > temperature for a single block) > Suggestions of Prof. Skadron are very useful. > I've put some silicon surrounding my function unit,forming a 1cm^2 chip. > And the steady temperature became more reasonable,318K. > But I'm still wondering about the reason of this floorplan. > > Moreover, I added some data to the ptrace file and got a ttrace file as follows. > -----ptrace-------- > silicon1 silicon2 silicon3 silicon4 FU > 0.00 0.00 0.00 0.00 4.10e-03 > 0.00 0.00 0.00 0.00 4.10e-03 > 0.00 0.00 0.00 0.00 4.10e-03 > 0.00 0.00 0.00 0.00 4.10e-03 > 0.00 0.00 0.00 0.00 4.10e-03 > -----ttrace--------- > silicon1 silicon2 silicon3 silicon4 FU > 279.83 400.94 385.36 279.83 272.11 > 278.6 400.26 386.03 278.6 271.56 > 277.43 399.72 386.58 277.43 270.93 > 276.32 399.27 387.02 276.32 270.27 > 275.27 398.89 387.38 275.27 269.58 > (Are units of these temperatures centigrade?) > ------------------- > The temperature trace and the steady temperature are quit different. > How can I solve this problem? > > Thanks for your patience to my four problems. > > Best Regards, > Foster Wang > _______________________________________________ > HotSpot mailing list > HotSpot at mail.cs.virginia.edu > http://www.cs.virginia.edu/mailman/listinfo/hotspot > From wh6p at cms.mail.virginia.edu Thu Oct 2 08:41:25 2008 From: wh6p at cms.mail.virginia.edu (Wei Huang) Date: Thu, 02 Oct 2008 11:41:25 -0400 Subject: [Hotspot] thermal model on single function unit In-Reply-To: References: Message-ID: <289F67739768313F651E8AE3@mstu2> Hi please responses below... --On Thursday, October 02, 2008 11:04 PM +0800 "=?BIG5?B?pP2vRaZ0Zm9zdGVydGF6?=" wrote: > Hi, > > I'm triying to model the temperature of single function > unit. The configuration is set as default. > Floorplan and power trace are like below: > ----Floorplan--- > FU 2.03e-06 2.03e-06 0.00 0.00 > ----ptrace--- > FU > 0.0041 > ------------ > The temperature turns out unreasonably high,5443K. > Why does the result become like this? > > Then I found a thread similar to my question.([Hotspot] > just want temperature for a single block) > Suggestions of Prof. Skadron are very useful. > I've put some silicon surrounding my function > unit,forming a 1cm^2 chip. And the steady temperature > became more reasonable,318K. But I'm still wondering > about the reason of this floorplan. Basically, if your chip is only a tiny block (2um x 2um in your case, if it is possible at all), there is no heat spreading at all inside the chip, therefore the high power density of that tiny block (1.0e+9 W/m^2) causes unrealistic temperatures. In a real chip, hot blocks usually have neighbor blocks that help spread heat laterally. The spreading is more dominant for a small block. > > Moreover, I added some data to the ptrace file and got a > ttrace file as follows. -----ptrace-------- > silicon1 silicon2 silicon3 silicon4 FU > 0.00 0.00 0.00 0.00 4.10e-03 > 0.00 0.00 0.00 0.00 4.10e-03 > 0.00 0.00 0.00 0.00 4.10e-03 > 0.00 0.00 0.00 0.00 4.10e-03 > 0.00 0.00 0.00 0.00 4.10e-03 > -----ttrace--------- > silicon1 silicon2 silicon3 silicon4 FU > 279.83 400.94 385.36 279.83 272.11 > 278.6 400.26 386.03 278.6 271.56 > 277.43 399.72 386.58 277.43 270.93 > 276.32 399.27 387.02 276.32 270.27 > 275.27 398.89 387.38 275.27 269.58 > (Are units of these temperatures centigrade?) > ------------------- > The temperature trace and the steady temperature are quit > different. How can I solve this problem? When running a transient simulation, it takes a long time (seconds to minutes simulated time) to reach steady-state. For the time duration in your simulation (~20us), it is far from steady-state temperatures yet. Additionally, for a floorplan like yours, I would recommend using the grid model, because some of the blocks (silicon1-4) have extreme aspect ratio, which renders the block model less accurate. You can still stick to the block model if you want, but you'll have to divide the high-aspect-ratio blocks into subblocks with aspect ratio closer to 1. Hope this helps, -Wei > > Thanks for your patience to my four problems. > > Best Regards, > Foster Wang > _______________________________________________ > HotSpot mailing list > HotSpot at mail.cs.virginia.edu > http://www.cs.virginia.edu/mailman/listinfo/hotspot From bernauer at informatik.uni-tuebingen.de Thu Oct 2 08:57:37 2008 From: bernauer at informatik.uni-tuebingen.de (Andreas Bernauer) Date: Thu, 02 Oct 2008 17:57:37 +0200 Subject: [Hotspot] thermal model on single function unit In-Reply-To: References: Message-ID: <48E4EF71.2000503@informatik.uni-tuebingen.de> ???fostertaz wrote: > -----ttrace--------- > silicon1 silicon2 silicon3 silicon4 FU > 279.83 400.94 385.36 279.83 272.11 > 278.6 400.26 386.03 278.6 271.56 > 277.43 399.72 386.58 277.43 270.93 > 276.32 399.27 387.02 276.32 270.27 > 275.27 398.89 387.38 275.27 269.58 > (Are units of these temperatures centigrade?) > ------------------- If I read README-2.0 of the source files correctly, the output is in Celsius, except for the steady temperature, which is in Kelvin. Cheers, Andreas. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 252 bytes Desc: OpenPGP digital signature Url : http://www.cs.virginia.edu/pipermail/hotspot/attachments/20081002/9000934f/attachment.bin From fostertaz at access.ee.ntu.edu.tw Thu Oct 2 09:40:22 2008 From: fostertaz at access.ee.ntu.edu.tw (=?BIG5?B?pP2vRaZ0Zm9zdGVydGF6?=) Date: Fri, 3 Oct 2008 00:40:22 +0800 Subject: [Hotspot] thermal model on single function unit In-Reply-To: References: Message-ID: Hi I think I inputed wrong data. The FU size would be 203um x 203um. Anyway, your answers still help a lot. Thanks. Best regards, Foster Wang 2008/10/2 Karthik Sankaranarayanan : > > The FU size (2 microns x 2 microns) and the power density (1000 W / mm^2) > appear to be too "out-of-range". Were these the numbers you intended to use? > If so, these are beyond what HotSpot is designed to model accurately. If > there has been an error in the units, here are the units to be used: > > floorplan file: all dimensions are in meters > power trace file: all values are in watts > > Hope this helps. > -karthik > > From ss_majzoub at yahoo.com Wed Oct 8 11:04:04 2008 From: ss_majzoub at yahoo.com (Sohaib Majzoub) Date: Wed, 8 Oct 2008 11:04:04 -0700 (PDT) Subject: [Hotspot] compiling HotSpot on 64bit Debian Message-ID: <911325.43377.qm@web50203.mail.re2.yahoo.com> Hi Everyone, I am trying to compile Hotspot on Debian 64bit and i am getting the error below. It seems something incompatible with the Intel Library? Appreciate any help, Thanks Sohaib gcc -O3? -I/intel/mkl/9.1.023/include -L/intel/mkl/9.1.023/lib/32 -DVERBOSE=1 -DMATHACCEL=1 -o hotspot hotspot.o temperature.o RCutil.o temperature_block.o? temperature_grid.o flp.o flp_desc.o npe.o shape.o? util.o wire.o -lmkl_lapack -lmkl_ia32 -lguide -lpthread -lm /usr/bin/ld: skipping incompatible /intel/mkl/9.1.023/lib/32/libmkl_lapack.so when searching for -lmkl_lapack /usr/bin/ld: skipping incompatible /intel/mkl/9.1.023/lib/32/libmkl_lapack.a when searching for -lmkl_lapack /usr/bin/ld: cannot find -lmkl_lapack collect2: ld returned 1 exit status make: *** [hotspot] Error 1 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20081008/7acba8a4/attachment.html From wh6p at cms.mail.virginia.edu Wed Oct 8 12:21:45 2008 From: wh6p at cms.mail.virginia.edu (Wei Huang) Date: Wed, 08 Oct 2008 15:21:45 -0400 Subject: [Hotspot] compiling HotSpot on 64bit Debian In-Reply-To: <911325.43377.qm@web50203.mail.re2.yahoo.com> References: <911325.43377.qm@web50203.mail.re2.yahoo.com> Message-ID: Hi, I am forwarding an earlier HotSpot mailing list email containing suggestions on solving this issue (thanks to my colleague Karthik). See below... Regards, -Wei ---------- Forwarded Message ---------- Date: Monday, September 29, 2008 4:16 PM -0400 From: Karthik Sankaranarayanan To: Wei Huang Subject: Re: Auotmatic reject (fwd) Hi Wei, Not sure what version of MKL he is using but this is an MKL problem. The LIBS flags need to be set from the MKL installation manual. The default flags in HotSpot are 32-bit flags from MKL 8.0.2. He needs the 64-bit flags. Following pages give information on what flags to use: http://www.intel.com/support/performancetools/libraries/mkl /linux/sb/CS-028423.htm http://www.intel.com/support/performancetools/sb/CS-028699. htm http://www.intel.com/software/products/mkl/docs/linux/WebHe lp/mkl_ug_linking_your_application/selecting_libraries_to_l ink.htm It appears that in the layered default model, following might work although I haven't tried it myself: -lmkl_lapack -lmkl -lguide -lm -lpthread -karthik On Mon, 29 Sep 2008, Wei Huang wrote: > Karthik, > > I've got this user subscribed to the mailing list. > > Would you mind have a look at his question during > compiling HotSpot? It seems there are some compatibility > issues between his OS and the library he used. > > Thanks, > -Wei > > ---------- Forwarded Message ---------- > Date: Monday, September 29, 2008 1:44 PM -0400 > From: Yan Zhang > To: hotspot-owner at mail.cs.virginia.edu > Subject: Auotmatic reject > > Hi there, > > I have subscribe the list but still being rejected. Just > want to post my question during installation. > > --------------------------------------------------------- > -- > --------------------------------------------------------- > -- ------------------------------------------------ > > After install INTEL's MKL, I have set up the configuration > like this: > --------------------------------------------------------- > -- -- > MATHACCEL = intel > INCDIR = /Program/mkl_intel/include > LIBDIR = /Program/mkl_intel/lib/32 > LIBS = -lm -lmkl_lapack -lmkl_ia32 -lguide > -lpthread EXTRAFLAGS = > --------------------------------------------------------- > -- --- > My Linux distribution is Ubuntu 7.04, running on Intel > Core2 Duo. After makefile, I got error: > --------------------------------------------------------- > -- --------------------------------- > make > gcc -O3 -I/Program/mkl_intel/include > -L/Program/mkl_intel/lib/32 -DVERBOSE=1 -DMATHACCEL=1 -o > hotspot hotspot.o temperature.o RCutil.o > temperature_block.o temperature_grid.o flp.o flp_desc.o > npe.o shape.o util.o wire.o -lmkl_lapack -lmkl_ia32 > -lguide -lpthread -lm > /Program/mkl_intel/lib/32/libmkl_lapack.so: undefined > reference to `mkl_serv_load_fun' > /Program/mkl_intel/lib/32/libmkl_lapack.so: undefined > reference to `mkl_serv_load_dll' > collect2: ld returned 1 exit status > make: *** [hotspot] Error 1 > > --------------------------------------------------------- > -- -------------------------------- > I think the default configuration is for 32 bit. How > should I change the configuration for 64-bit? Many thanks. > Yan Zhang > young.chang at hotmail.com > ------------------------------------- > --------------------------------------------------------- > -- > --------------------------------------------------------- > -- ----------------------------------------------------- > > > > > __________________________________________________ > Want to do more with Windows Live? Learn ?0 hidden > secrets?from Jamie. Learn Now > > ---------- End Forwarded Message ---------- > > > > > > ---------- End Forwarded Message ---------- --On Wednesday, October 08, 2008 11:04 AM -0700 Sohaib Majzoub wrote: > Hi Everyone, > I am trying to compile Hotspot on Debian 64bit and i am > getting the error below. > It seems something incompatible with the Intel Library? > Appreciate any help, > Thanks > Sohaib > > gcc -O3 -I/intel/mkl/9.1.023/include > -L/intel/mkl/9.1.023/lib/32 -DVERBOSE=1 -DMATHACCEL=1 -o > hotspot hotspot.o temperature.o RCutil.o > temperature_block.o temperature_grid.o flp.o flp_desc.o > npe.o shape.o util.o wire.o -lmkl_lapack -lmkl_ia32 > -lguide -lpthread -lm > /usr/bin/ld: skipping incompatible > /intel/mkl/9.1.023/lib/32/libmkl_lapack.so when searching > for -lmkl_lapack > /usr/bin/ld: skipping incompatible > /intel/mkl/9.1.023/lib/32/libmkl_lapack.a when searching > for -lmkl_lapack > /usr/bin/ld: cannot find -lmkl_lapack > collect2: ld returned 1 exit status > make: *** [hotspot] Error 1 > > > From sohaib.majzoub at ieee.org Thu Oct 9 11:33:28 2008 From: sohaib.majzoub at ieee.org (Sohaib Majzoub) Date: Thu, 9 Oct 2008 11:33:28 -0700 (PDT) Subject: [Hotspot] compiling HotSpot on 64bit Debian In-Reply-To: Message-ID: <83916.2472.qm@web50211.mail.re2.yahoo.com> thanks a lot for the help, Sohaib --- On Wed, 10/8/08, Wei Huang wrote: From: Wei Huang Subject: Re: [Hotspot] compiling HotSpot on 64bit Debian To: sohaib.majzoub at ieee.org, hotspot at mail.cs.virginia.edu Date: Wednesday, October 8, 2008, 12:21 PM Hi, I am forwarding an earlier HotSpot mailing list email containing suggestions on solving this issue (thanks to my colleague Karthik). See below... Regards, -Wei ---------- Forwarded Message ---------- Date: Monday, September 29, 2008 4:16 PM -0400 From: Karthik Sankaranarayanan To: Wei Huang Subject: Re: Auotmatic reject (fwd) Hi Wei, Not sure what version of MKL he is using but this is an MKL problem. The LIBS flags need to be set from the MKL installation manual. The default flags in HotSpot are 32-bit flags from MKL 8.0.2. He needs the 64-bit flags. Following pages give information on what flags to use: http://www.intel.com/support/performancetools/libraries/mkl /linux/sb/CS-028423.htm http://www.intel.com/support/performancetools/sb/CS-028699. htm http://www.intel.com/software/products/mkl/docs/linux/WebHe lp/mkl_ug_linking_your_application/selecting_libraries_to_l ink.htm It appears that in the layered default model, following might work although I haven't tried it myself: -lmkl_lapack -lmkl -lguide -lm -lpthread -karthik On Mon, 29 Sep 2008, Wei Huang wrote: > Karthik, > > I've got this user subscribed to the mailing list. > > Would you mind have a look at his question during > compiling HotSpot? It seems there are some compatibility > issues between his OS and the library he used. > > Thanks, > -Wei > > ---------- Forwarded Message ---------- > Date: Monday, September 29, 2008 1:44 PM -0400 > From: Yan Zhang > To: hotspot-owner at mail.cs.virginia.edu > Subject: Auotmatic reject > > Hi there, > > I have subscribe the list but still being rejected. Just > want to post my question during installation. > > --------------------------------------------------------- > -- > --------------------------------------------------------- > -- ------------------------------------------------ > > After install INTEL's MKL, I have set up the configuration > like this: > --------------------------------------------------------- > -- -- > MATHACCEL = intel > INCDIR = /Program/mkl_intel/include > LIBDIR = /Program/mkl_intel/lib/32 > LIBS = -lm -lmkl_lapack -lmkl_ia32 -lguide > -lpthread EXTRAFLAGS = > --------------------------------------------------------- > -- --- > My Linux distribution is Ubuntu 7.04, running on Intel > Core2 Duo. After makefile, I got error: > --------------------------------------------------------- > -- --------------------------------- > make > gcc -O3 -I/Program/mkl_intel/include > -L/Program/mkl_intel/lib/32 -DVERBOSE=1 -DMATHACCEL=1 -o > hotspot hotspot.o temperature.o RCutil.o > temperature_block.o temperature_grid.o flp.o flp_desc.o > npe.o shape.o util.o wire.o -lmkl_lapack -lmkl_ia32 > -lguide -lpthread -lm > /Program/mkl_intel/lib/32/libmkl_lapack.so: undefined > reference to `mkl_serv_load_fun' > /Program/mkl_intel/lib/32/libmkl_lapack.so: undefined > reference to `mkl_serv_load_dll' > collect2: ld returned 1 exit status > make: *** [hotspot] Error 1 > > --------------------------------------------------------- > -- -------------------------------- > I think the default configuration is for 32 bit. How > should I change the configuration for 64-bit? Many thanks. > Yan Zhang > young.chang at hotmail.com > ------------------------------------- > --------------------------------------------------------- > -- > --------------------------------------------------------- > -- ----------------------------------------------------- > > > > > __________________________________________________ > Want to do more with Windows Live? Learn ?0 hidden > secrets?from Jamie. Learn Now > > ---------- End Forwarded Message ---------- > > > > > > ---------- End Forwarded Message ---------- --On Wednesday, October 08, 2008 11:04 AM -0700 Sohaib Majzoub wrote: > Hi Everyone, > I am trying to compile Hotspot on Debian 64bit and i am > getting the error below. > It seems something incompatible with the Intel Library? > Appreciate any help, > Thanks > Sohaib > > gcc -O3 -I/intel/mkl/9.1.023/include > -L/intel/mkl/9.1.023/lib/32 -DVERBOSE=1 -DMATHACCEL=1 -o > hotspot hotspot.o temperature.o RCutil.o > temperature_block.o temperature_grid.o flp.o flp_desc.o > npe.o shape.o util.o wire.o -lmkl_lapack -lmkl_ia32 > -lguide -lpthread -lm > /usr/bin/ld: skipping incompatible > /intel/mkl/9.1.023/lib/32/libmkl_lapack.so when searching > for -lmkl_lapack > /usr/bin/ld: skipping incompatible > /intel/mkl/9.1.023/lib/32/libmkl_lapack.a when searching > for -lmkl_lapack > /usr/bin/ld: cannot find -lmkl_lapack > collect2: ld returned 1 exit status > make: *** [hotspot] Error 1 > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20081009/fbbc8f38/attachment-0001.html From ss_majzoub at yahoo.com Fri Oct 24 12:39:19 2008 From: ss_majzoub at yahoo.com (Sohaib Majzoub) Date: Fri, 24 Oct 2008 12:39:19 -0700 (PDT) Subject: [Hotspot] Hotspot, chip size resitrictions Message-ID: <230022.82403.qm@web50208.mail.re2.yahoo.com> Hi Everyone, What is the the restrictions (if any) on the chip size and/or number of blocks in the chip? I was trying a 30x30 blocks (900 blocks) and i am getting "error line too long." Is there a workaround to this problem? thanks Sohaib -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20081024/b2e62c66/attachment.html From ks4kk at cs.virginia.edu Sat Oct 25 08:46:48 2008 From: ks4kk at cs.virginia.edu (Karthik Sankaranarayanan) Date: Sat, 25 Oct 2008 11:46:48 -0400 (EDT) Subject: [Hotspot] Hotspot, chip size resitrictions In-Reply-To: <230022.82403.qm@web50208.mail.re2.yahoo.com> References: <230022.82403.qm@web50208.mail.re2.yahoo.com> Message-ID: Hi Sohaib, I think the issue is the buffer size in reading lines from the power trace file. Please increase the LINE_SIZE macro in util.h to a higher value. Hope this helps. -karthik On Fri, 24 Oct 2008, Sohaib Majzoub wrote: > Hi Everyone, > What is the the restrictions (if any) on the chip size and/or number of blocks in the chip? > I was trying a 30x30 blocks (900 blocks) and i am getting "error line too long." > Is there a workaround to this problem? > thanks > Sohaib From augusto.vega at bsc.es Mon Oct 27 03:13:45 2008 From: augusto.vega at bsc.es (Augusto Vega) Date: Mon, 27 Oct 2008 12:13:45 +0100 Subject: [Hotspot] Wattch question Message-ID: <4905A269.6000309@bsc.es> Hi, I know this is a mailing list about HotSpot. But I have some questions about Wattch and SimpleScalar and I was just wondering if somebody in this forum can help me. The question is why the `ialu_access' counter in Wattch is only updated when an integer instruction is executed? I think it should also be updated in the case of LD/ST instructions, because ALU is used for effective address calculation. Am I wrong? Thank you in advance, -- Augusto Vega BSC-CNS Barcelona Supercomputing Center Centro Nacional de Supercomputaci?n Telf.: +34 93 413 79 46 e-mail: augusto.vega at bsc.es From skadron at cs.virginia.edu Mon Oct 27 05:44:09 2008 From: skadron at cs.virginia.edu (Kevin Skadron) Date: Mon, 27 Oct 2008 07:44:09 -0600 Subject: [Hotspot] Wattch question In-Reply-To: <4905A269.6000309@bsc.es> References: <4905A269.6000309@bsc.es> Message-ID: <4905C5A9.4060502@cs.virginia.edu> As I recall, a load or store is split into an ADDQ in the reservation station and then the actual mem op is placed in the load/store queue. So the address generation behaves like an ialu op and should be captured by that counter. /K Augusto Vega wrote: > Hi, > > I know this is a mailing list about HotSpot. But I have some questions > about Wattch and SimpleScalar and I was just wondering if somebody in > this forum can help me. > > The question is why the `ialu_access' counter in Wattch is only updated > when an integer instruction is executed? I think it should also be > updated in the case of LD/ST instructions, because ALU is used for > effective address calculation. Am I wrong? > > Thank you in advance, > From uxc0983 at cacs.louisiana.edu Tue Oct 28 08:59:19 2008 From: uxc0983 at cacs.louisiana.edu (Chandran Unni) Date: Tue, 28 Oct 2008 11:59:19 -0500 (CDT) Subject: [Hotspot] Re : 3D modeling Message-ID: <60943.130.70.77.114.1225213159.squirrel@webmail.cacs.louisiana.edu> Hi, I was going through the layer.lcf and example.lcf files. I see that thermal interface material has the same floorplan file as the silicon layout. I was wondering if we should modify the TIM description ?? Thanx in advance, Unni From wh6p at cms.mail.virginia.edu Tue Oct 28 10:20:50 2008 From: wh6p at cms.mail.virginia.edu (Wei Huang) Date: Tue, 28 Oct 2008 14:20:50 -0400 Subject: [Hotspot] Re : 3D modeling In-Reply-To: <60943.130.70.77.114.1225213159.squirrel@webmail.cacs.louisiana.edu> References: <60943.130.70.77.114.1225213159.squirrel@webmail .cacs.louisiana.edu> Message-ID: <589A063FF1BE96B9106E23CB@mstu2> Hi, The TIM is just a passive layer that needs to be included in the model. Its floorplan can be arbitrary. However, setting the floorplan the same as its neighboring silicon layer is convenient because you can see the TIM temperatures beneath the silicon blocks, and sometimes that can be useful. You may just set the TIM layer with the same floorplan as the silicon layer above it in the lcf file. In the lcf file, it is something like this: silicon1 (flp1) TIM1 (flp1) silicon2 (flp2) TIM2 (flp2) etc... -Wei --On Tuesday, October 28, 2008 11:59 AM -0500 Chandran Unni wrote: > Hi, > I was going through the layer.lcf and example.lcf > files. I see that thermal interface material has the same > floorplan file as the silicon layout. I was wondering if > we should modify the TIM description ?? > > Thanx in advance, > Unni > _______________________________________________ > HotSpot mailing list > HotSpot at mail.cs.virginia.edu > http://www.cs.virginia.edu/mailman/listinfo/hotspot From uxc0983 at cacs.louisiana.edu Tue Oct 28 12:01:15 2008 From: uxc0983 at cacs.louisiana.edu (Chandran Unni) Date: Tue, 28 Oct 2008 15:01:15 -0500 (CDT) Subject: [Hotspot] Re : 3D modeling In-Reply-To: <589A063FF1BE96B9106E23CB@mstu2> References: <60943.130.70.77.114.1225213159.squirrel@webmail .cacs.louisiana.edu> <589A063FF1BE96B9106E23CB@mstu2> Message-ID: <3961.130.70.142.7.1225224075.squirrel@webmail.cacs.louisiana.edu> Thank You very much Wei. > Hi, > > The TIM is just a passive layer that needs to be included > in the model. Its floorplan can be arbitrary. However, > setting the floorplan the same as its neighboring silicon > layer is convenient because you can see the TIM > temperatures beneath the silicon blocks, and sometimes that > can be useful. > > You may just set the TIM layer with the same floorplan as > the silicon layer above it in the lcf file. In the lcf > file, it is something like this: > > silicon1 (flp1) > TIM1 (flp1) > silicon2 (flp2) > TIM2 (flp2) > etc... > > -Wei > > > --On Tuesday, October 28, 2008 11:59 AM -0500 Chandran Unni > wrote: > >> Hi, >> I was going through the layer.lcf and example.lcf >> files. I see that thermal interface material has the same >> floorplan file as the silicon layout. I was wondering if >> we should modify the TIM description ?? >> >> Thanx in advance, >> Unni >> _______________________________________________ >> HotSpot mailing list >> HotSpot at mail.cs.virginia.edu >> http://www.cs.virginia.edu/mailman/listinfo/hotspot > > > > > From uxc0983 at cacs.louisiana.edu Wed Oct 29 09:13:31 2008 From: uxc0983 at cacs.louisiana.edu (Chandran Unni) Date: Wed, 29 Oct 2008 12:13:31 -0500 (CDT) Subject: [Hotspot] Re : 3D modeling In-Reply-To: <589A063FF1BE96B9106E23CB@mstu2> References: <60943.130.70.77.114.1225213159.squirrel@webmail .cacs.louisiana.edu> <589A063FF1BE96B9106E23CB@mstu2> Message-ID: <1111.130.70.155.174.1225300411.squirrel@webmail.cacs.louisiana.edu> Hello Wei, While generating the transient and steady state temperatures for a 3D IC, does hotspot calculate vertical and horizontal temperature values ? Thanks Unni > Hi, > > The TIM is just a passive layer that needs to be included > in the model. Its floorplan can be arbitrary. However, > setting the floorplan the same as its neighboring silicon > layer is convenient because you can see the TIM > temperatures beneath the silicon blocks, and sometimes that > can be useful. > > You may just set the TIM layer with the same floorplan as > the silicon layer above it in the lcf file. In the lcf > file, it is something like this: > > silicon1 (flp1) > TIM1 (flp1) > silicon2 (flp2) > TIM2 (flp2) > etc... > > -Wei > > > --On Tuesday, October 28, 2008 11:59 AM -0500 Chandran Unni > wrote: > >> Hi, >> I was going through the layer.lcf and example.lcf >> files. I see that thermal interface material has the same >> floorplan file as the silicon layout. I was wondering if >> we should modify the TIM description ?? >> >> Thanx in advance, >> Unni >> _______________________________________________ >> HotSpot mailing list >> HotSpot at mail.cs.virginia.edu >> http://www.cs.virginia.edu/mailman/listinfo/hotspot > From wh6p at cms.mail.virginia.edu Wed Oct 29 09:56:05 2008 From: wh6p at cms.mail.virginia.edu (Wei Huang) Date: Wed, 29 Oct 2008 13:56:05 -0400 Subject: [Hotspot] Re : 3D modeling In-Reply-To: <1111.130.70.155.174.1225300411.squirrel@webmail.cacs.louisiana.edu> References: <60943.130.70.77.114.1225213159.squirrel@webmail .cacs.louisiana.edu> <589A063FF1BE96B9106E23CB@mstu2> <1111.130.70.155.174.1225300411.squirrel@webmail.cacs.louisia na.edu> Message-ID: <4A5514444692FE3AC241ED54@mstu2> Hi Unni, I am not sure what you mean by vertical and horizontal temperatures. Basically, HotSpot gives temperatures (and traces of temperatures if using transient mode) for units that appear in your floorplans for each silicon layer. -Wei --On Wednesday, October 29, 2008 12:13 PM -0500 Chandran Unni wrote: > Hello Wei, > While generating the transient and steady state > temperatures for a 3D IC, does hotspot calculate vertical > and horizontal temperature values ? > > Thanks > Unni > > >> Hi, >> >> The TIM is just a passive layer that needs to be included >> in the model. Its floorplan can be arbitrary. However, >> setting the floorplan the same as its neighboring silicon >> layer is convenient because you can see the TIM >> temperatures beneath the silicon blocks, and sometimes >> that can be useful. >> >> You may just set the TIM layer with the same floorplan as >> the silicon layer above it in the lcf file. In the lcf >> file, it is something like this: >> >> silicon1 (flp1) >> TIM1 (flp1) >> silicon2 (flp2) >> TIM2 (flp2) >> etc... >> >> -Wei >> >> >> --On Tuesday, October 28, 2008 11:59 AM -0500 Chandran >> Unni wrote: >> >>> Hi, >>> I was going through the layer.lcf and example.lcf >>> files. I see that thermal interface material has the >>> same floorplan file as the silicon layout. I was >>> wondering if we should modify the TIM description ?? >>> >>> Thanx in advance, >>> Unni >>> _______________________________________________ >>> HotSpot mailing list >>> HotSpot at mail.cs.virginia.edu >>> http://www.cs.virginia.edu/mailman/listinfo/hotspot >> >