Re: streams mp code - show and tell!

From: tom spelce (spelce@kistler.engr.sgi.com)
Date: Tue Dec 31 1996 - 16:39:52 CST


John,

Here are some streams runs for my speedracer (Rev B heart ... ) let me know if
these numbers vary from yours.

Thanks,

Tom

2 cpu speedracer:

 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLE PRECISION word
----------------------------------------------
 Array size = 8800000
 Offset = 0
 The total memory requirement is 201 MB
 You are running each test 5 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) Min time Max time Mean time RMS time Median
Copy: 469.33 0.3000 0.3100 0.3080 0.3080 0.3100
Scale: 454.19 0.3100 0.3200 0.3140 0.3140 0.3100
Add: 515.12 0.4100 0.4200 0.4140 0.4140 0.4100
Triad: 515.12 0.4100 0.4200 0.4120 0.4120 0.4200
-----------------------------------------------------------------------------
 Sum of a is = 151874.9999938380
 Sum of b is = 30375.00000156571
 Sum of c is = 40499.99999812090

---------------------------------------------------------------------------------

1cpu of 2 cpu speedracer ...
sangiacomo 28% ./stream
----------------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLE PRECISION word
----------------------------------------------
 Array size = 8800000
 Offset = 0
 The total memory requirement is 201 MB
 You are running each test 5 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 320000 microseconds
    (= 32 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) Min time Max time Mean time RMS time Median
Copy: 320.00 0.4400 0.4500 0.4440 0.4440 0.4400
Scale: 312.89 0.4500 0.4500 0.4500 0.4500 0.4500
Add: 352.00 0.6000 0.6100 0.6040 0.6040 0.6000
Triad: 357.97 0.5900 0.5900 0.5900 0.5900 0.5900
-----------------------------------------------------------------------------
 Sum of a is = 303750.0000219342
 Sum of b is = 60750.00000657157
 Sum of c is = 80999.99998345258
--------------------------------------------------------------------------------------

Impact 10000
  ----------------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLE PRECISION word
----------------------------------------------
 Array size = 8800000
 Offset = 0
 The total memory requirement is 201 MB
 You are running each test 5 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 820000 microseconds
    (= 82 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) Min time Max time Mean time RMS time Median
Copy: 105.86 1.3300 1.3600 1.3400 1.3400 1.3300
Scale: 105.86 1.3300 1.3500 1.3380 1.3380 1.3300
Add: 118.65 1.7800 1.8200 1.8100 1.8101 1.8200
Triad: 116.69 1.8100 1.8300 1.8200 1.8200 1.8300
-----------------------------------------------------------------------------
 Sum of a is = 303750.0000219342
 Sum of b is = 60750.00000657157
 Sum of c is = 80999.99998345258

s

-- 
Tom Spelce                        DSD Aurora Engineering
Silicon Graphics Inc.             spelce@sgi.com
2011 N. Shoreline Blvd.           1-415-933-4254
Mountain View, CA 94043-1389      1-415-964-8187(fax)



This archive was generated by hypermail 2b29 : Tue Apr 18 2000 - 05:23:06 CDT