Re: latest linus-2.5 BK broken

Jesse Pollard (pollard@tomcat.admin.navo.hpc.mil)
Fri, 21 Jun 2002 07:59:15 -0500 (CDT)


Martin Dalecki <dalecki@evision-ventures.com>:
>Yes HT gives 12%. naive SMP gives 50% and good SMP (aka corssbar bus)
>gives 70% for two CPU. All those numbers are well below the level
>where more then 2-4 makes hardly any sense... Amdahl bites you still if you
>read it like:
...

I think your numbers are a little low - I've seen between 50%-80% on
master/slave SMP depending on the job. 50% if both processess are heavily
syscall oriented, 75% (or therabouts) when both processes are more normally
balanced, and 80% if both processes are more compute bound.

Good SMP, with a crossbar switch buss should give close to 95%. Good SMP
alone should give about 75%.

My expierence with good crossbar switch is based on Cray UNICOS/YMP/SV
hardware. A well tuned hardware platform, and slightly less well tuned
SMP implementation, though the UNICOS 10 rewrite may have fixed the
SMP implementation.

-------------------------------------------------------------------------
Jesse I Pollard, II
Email: pollard@navo.hpc.mil

Any opinions expressed are solely my own.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/