Re: Stream FP Exception on OpenBSD on DEC Alpha Multia

From: Bart Grantham <bart@chumps.net>
Date: Fri Nov 15 2002 - 02:02:35 CST

I spoke too soon! It gave a FP exception with compiler flags "gcc -O3
stream_d.c second_cpu.c -o stream -lm" but worked when I backed off to
only -O . Jeez. I'll have to see what asm gcc is kicking out on
alpha. Anyways, here's the results on a DEC Multia with a 166MHz Alpha
AXP :

bash-2.05$ gcc -O stream_d.c second_cpu.c -o stream -lm
bash-2.05$ ./stream
-------------------------------------------------------------
This system uses 8 bytes per DOUBLE PRECISION word.
-------------------------------------------------------------
Array size = 2000000, Offset = 0
Total memory required = 45.8 MB.
Each test is run 10 times, but only
the *best* time for each is used.
-------------------------------------------------------------
Your clock granularity/precision appears to be 9999 microseconds.
Each test below will take on the order of 1420000 microseconds.
   (= 142 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
Copy: 26.4463 1.2160 1.2100 1.2300
Scale: 26.6666 1.2150 1.2000 1.2300
Add: 26.6667 1.8110 1.8000 1.8200
Triad: 26.5194 1.8140 1.8100 1.8200

Enjoy!

Bart Grantham wrote:

> Subject says it all. :-/ Using the latest sources. Would you like
> me to look into it, and specifically, do you have any things that I
> should try to track down the problem?
>
> Don't know if you care, but the bytebench for the machine is:
>
> http://www.bartgrantham.com/bytebench.txt
>
Received on Fri Nov 15 02:02:35 2002

This archive was generated by hypermail 2.1.8 : Wed Feb 11 2004 - 14:53:14 CST