Now it looks reasonable!

From: weyrich@compt.chemie.uni-konstanz.de
Date: Fri Jan 27 1995 - 17:18:48 CST


Dear Dr. McCalpin,

Thank you very much for your help. I have followed literally
your prescription and have received the following results:

LOG.serial:
***********
--------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 57.98809528350830 hundredths of a second
 Increase the size of the arrays if this is <30
  and your clock precision is =<1/100 second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 134.3584 0.5973 0.2382 1.3161
Scaling : 133.7915 0.6519 0.2392 1.1426
Summing : 133.5284 0.6801 0.3595 1.5240
SAXPYing : 134.5865 0.6277 0.3566 1.1684
 Sum of a is : 1.1533007812800417E+18
 Sum of b is : 2.3066015625365702E+17
 Sum of c is : 3.0754687500612557E+17

LOG.1:
******
--------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 68.77861022949219 hundredths of a second
 Increase the size of the arrays if this is <30
  and your clock precision is =<1/100 second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 124.3583 0.2736 0.2573 0.3583
Scaling : 122.9336 0.2925 0.2603 0.3646
Summing : 132.4390 0.4254 0.3624 0.4920
SAXPYing : 132.7547 0.3908 0.3616 0.4813
 Sum of a is : 1.1533007812800417E+18
 Sum of b is : 2.3066015625365702E+17
 Sum of c is : 3.0754687500612557E+17

LOG.2:
******
--------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 131.0092091560364 hundredths of a second
 Increase the size of the arrays if this is <30
  and your clock precision is =<1/100 second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 216.7389 1.6026 0.1476 1.9973
Scaling : 233.7164 1.8901 0.1369 2.0187
Summing : 39.6001 2.4917 1.2121 2.9848
SAXPYing : 23.7588 2.6674 2.0203 2.9719
 Sum of a is : 1.1533007812480837E+18
 Sum of b is : 2.3066015625131405E+17
 Sum of c is : 3.0754687500425120E+17

LOG.3:
******
--------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 130.0364017486572 hundredths of a second
 Increase the size of the arrays if this is <30
  and your clock precision is =<1/100 second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 332.1917 1.2407 0.0963 1.9323
Scaling : 313.8209 1.2810 0.1020 1.9710
Summing : 320.0939 1.7044 0.1500 1.9950
SAXPYing : 332.4929 1.6362 0.1444 2.0086
 Sum of a is : 1.1533007812480940E+18
 Sum of b is : 2.3066015624897107E+17
 Sum of c is : 3.0754687500237683E+17

LOG.4:
******
--------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 139.6764039993286 hundredths of a second
 Increase the size of the arrays if this is <30
  and your clock precision is =<1/100 second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 385.7701 0.7763 0.0830 1.0157
Scaling : 365.6646 1.0159 0.0875 1.9298
Summing : 424.8539 1.2586 0.1130 1.9661
SAXPYing : 411.7521 1.3243 0.1166 1.9808
 Sum of a is : 1.1533007812481252E+18
 Sum of b is : 2.3066015624862480E+17
 Sum of c is : 3.0754687500050240E+17

LOG.5:
******
--------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 145.8331942558289 hundredths of a second
 Increase the size of the arrays if this is <30
  and your clock precision is =<1/100 second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 518.1091 0.8176 0.0618 1.0206
Scaling : 428.8284 1.0022 0.0746 1.4778
Summing : 461.7826 1.2218 0.1039 1.9587
SAXPYing : 486.7564 0.9611 0.0986 1.0444
 Sum of a is : 1.1533007812481564E+18
 Sum of b is : 2.3066015624878099E+17
 Sum of c is : 3.0754687500000000E+17

LOG.6:
******
--------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 146.2658047676086 hundredths of a second
 Increase the size of the arrays if this is <30
  and your clock precision is =<1/100 second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 634.0027 0.5602 0.0505 1.0086
Scaling : 542.4648 0.8850 0.0590 1.4662
Summing : 615.0137 1.1503 0.0780 1.9288
SAXPYing : 606.6119 0.9993 0.0791 1.9374
 Sum of a is : 1.1533007812481876E+18
 Sum of b is : 2.3066015624893718E+17
 Sum of c is : 3.0754687500000006E+17

LOG.7:
******
--------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 146.9164967536926 hundredths of a second
 Increase the size of the arrays if this is <30
  and your clock precision is =<1/100 second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 765.7333 0.7300 0.0418 0.9923
Scaling : 676.1462 0.6427 0.0473 1.0135
Summing : 643.4832 1.0677 0.0746 1.5109
SAXPYing : 672.7582 0.9236 0.0713 1.9450
 Sum of a is : 1.1533007812482189E+18
 Sum of b is : 2.3066015624909341E+17
 Sum of c is : 3.0754687500000000E+17

The maximum number of CPUs of our PCh is at present 7, indeed.
$i=8 gave an error message telling that, so I interrupted the
process.

I have no idea what has happened for $i=2 with Summing and
SAXPYing.

Thanks a lot once again! I learned quite a bit from you.

Very best regards,

Wolf Weyrich.

PS: I now better go home (0:14 CET); your coffee at 6:00 am that
you reported about is probably not such a good advice!



This archive was generated by hypermail 2b29 : Tue Apr 18 2000 - 05:23:04 CDT