On Sat, 16 Mar 1996, John D. McCalpin wrote:
> > I checked your FTP archive. A guy from DEC just added 2 figures from
> > 2 different runs to get a result for a 2 CPU system [...]
>
> That is acceptable (but not preferred) if the results are from
> two *simultaneous* runs and if the wall clock timer (rather than
> the cpu timer) is used. The should interfere with each other
> in much the same way as for a "true" multi-cpu run.
I see.
I also got some figures for older DECs we have.
The first one is DECstation 5000/240 (MIPS R3000 at 40 MHz) with 64 MB
running ULTRIX.
I used -O3 -D_POSIX_SOURCE to compile it. The latter flag is needed for
CLK_TCK to be defined.
-------------------------------------------------------------
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 16666 microseconds.
Each test below will take on the order of 350000 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
Assignment: 50.5263 0.3318 0.3167 0.3500
Scaling : 48.0000 0.3468 0.3333 0.3667
Summing : 27.1698 0.9102 0.8833 0.9333
SAXPYing : 27.1698 0.9034 0.8833 0.9167
**************************************************************
Another one is DECstation 5000/133 (R3000 at 33MHz) with 16M of RAM. I
had to decrease N to avoid swapping.
-------------------------------------------------------------
This system uses 8 bytes per DOUBLE PRECISION word.
-------------------------------------------------------------
Array size = 450000, Offset = 0
Total memory required = 10.3 MB.
Each test is run 10 times, but only
the *best* time for each is used.
-------------------------------------------------------------
Your clock granularity/precision appears to be 16666 microseconds.
Each test below will take on the order of 300000 microseconds.
(= 18 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
Assignment: 24.0000 0.3068 0.3000 0.3167
Scaling : 25.4118 0.2867 0.2833 0.3000
Summing : 20.2500 0.5534 0.5333 0.5667
SAXPYing : 20.9032 0.5287 0.5167 0.5833
****************************************************************
The last one is DEC 5000/25 (R3000 at 25MHz) with 12M of RAM.
-------------------------------------------------------------
This system uses 8 bytes per DOUBLE PRECISION word.
-------------------------------------------------------------
Array size = 250000, Offset = 0
Total memory required = 5.7 MB.
Each test is run 10 times, but only
the *best* time for each is used.
-------------------------------------------------------------
Your clock granularity/precision appears to be 16666 microseconds.
Each test below will take on the order of 266666 microseconds.
(= 16 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
Assignment: 18.4615 0.2486 0.2167 0.3333
Scaling : 17.1429 0.2531 0.2333 0.3667
Summing : 13.3333 0.4602 0.4500 0.4833
SAXPYing : 13.8462 0.4468 0.4333 0.4667
**************************************************************
Thank you for your time,
Igor
This archive was generated by hypermail 2b29 : Tue Apr 18 2000 - 05:23:05 CDT