re: Memory bandwidth testing

From: Dave Olson (olson@anchor.esd.sgi.com)
Date: Mon Sep 30 1991 - 18:47:34 CDT


Here are the results from the 35 and the indigo, both running 4.0
with 16 Mbytes of memory, compiled as f77 -O3. I ran it 3 times
in a row on each.

=========== out.SGI_35_s =====
 Single precision appears to have 7 digits of accuracy
 Assuming 4 bytes per default REAL word
--------------------------------------
 Timing calibration ; time = 50.00000 hundredths of a second
 Timing calibration ; time = 53.00000 hundredths of a second
 Timing calibration ; time = 52.00000 hundredths of a second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 50.0000 0.0875 0.0800 0.1100
          : 50.0000 0.0832 0.0800 0.1000
          : 50.0000 0.0831 0.0800 0.0900
Scaling : 28.5715 0.1452 0.1400 0.1600
          : 28.5715 0.1421 0.1400 0.1500
          : 28.5715 0.1481 0.1400 0.1600
Summing : 37.5000 0.1661 0.1600 0.1800
          : 37.5001 0.1620 0.1600 0.1700
          : 37.5000 0.1661 0.1600 0.1800
SAXPYing : 25.0000 0.2538 0.2400 0.3100
          : 25.0000 0.2410 0.2400 0.2500
          : 25.0000 0.2501 0.2400 0.2600
=========== out.SGI_35_d =====
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 45.99999990314245 hundredths of a second
 Timing calibration ; time = 45.00000104308128 hundredths of a second
 Timing calibration ; time = 45.00000085681677 hundredths of a second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 53.3334 0.1129 0.0900 0.2000
          : 53.3335 0.0910 0.0900 0.1000
          : 53.3335 0.0900 0.0900 0.0900
Scaling : 36.9233 0.1352 0.1300 0.1500
          : 36.9231 0.1351 0.1300 0.1400
          : 36.9231 0.1381 0.1300 0.1500
Summing : 48.0001 0.1698 0.1500 0.2000
          : 45.0000 0.1632 0.1600 0.1900
          : 45.0000 0.1641 0.1600 0.1700
SAXPYing : 42.3530 0.1834 0.1700 0.2100
          : 42.3530 0.1825 0.1700 0.2200
          : 42.3530 0.1842 0.1700 0.2000
=========== out.SGI_Indigo_s ====
 Single precision appears to have 7 digits of accuracy
 Assuming 4 bytes per default REAL word
--------------------------------------
 Timing calibration ; time = 58.00000 hundredths of a second
 Timing calibration ; time = 60.00000 hundredths of a second
 Timing calibration ; time = 61.00000 hundredths of a second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 28.5714 0.1512 0.1400 0.1600
          : 28.5715 0.1462 0.1400 0.1600
          : 28.5715 0.1463 0.1400 0.1700
Scaling : 20.0000 0.2184 0.2000 0.2400
          : 21.0526 0.2136 0.1900 0.2400
          : 21.0526 0.2114 0.1900 0.2300
Summing : 25.0000 0.2574 0.2400 0.2800
          : 26.0870 0.2463 0.2300 0.2700
          : 26.0869 0.2525 0.2300 0.2800
SAXPYing : 18.1818 0.3637 0.3300 0.3900
          : 18.7500 0.3536 0.3200 0.3800
          : 18.1818 0.3523 0.3300 0.3800
========== out.SGI_Indigo_d ====
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLEPRECISION word
--------------------------------------
 Timing calibration ; time = 53.00000030547380 hundredths of a second
 Timing calibration ; time = 50.99999718368053 hundredths of a second
 Timing calibration ; time = 56.00000526756048 hundredths of a second
 ---------------------------------------------------
Function Rate (MB/s) RMS time Min time Max time
Assignment: 32.0000 0.1683 0.1500 0.2200
          : 34.2859 0.1647 0.1400 0.1900
          : 32.0000 0.1656 0.1500 0.2000
Scaling : 22.8571 0.2378 0.2100 0.2800
          : 22.8571 0.2346 0.2100 0.2700
          : 22.8571 0.2396 0.2100 0.2700
Summing : 27.6924 0.2936 0.2600 0.3500
          : 27.6923 0.2849 0.2600 0.3300
          : 28.8000 0.2785 0.2500 0.3100
SAXPYing : 26.6667 0.3269 0.2700 0.4000
          : 25.7143 0.3079 0.2800 0.3600
          : 25.7143 0.3015 0.2800 0.3400

     Dave Olson
----============----
Life would be so much easier if we could just look at the source code.



This archive was generated by hypermail 2b29 : Tue Apr 18 2000 - 05:23:01 CDT