[RC5] More about 64bit clients

Fuzzy Logic fuzzymo at gmail.com
Wed Aug 24 18:25:27 EDT 2005


First of all, thank you for providing a useful reply, rather unlike
the angry one from Decio. I'm not sure how I managed to anger him.

I have to say that the results are interesting. Looks like the RC5-72
is a hair faster running 32-bit (might be a testing artifact), but
significantly faster for OGR on 64-bit.

What kind of numbers do we see for PPC-32 versus PPC-64? I couldn't
find statistics on the d.net site, and have no personal experience
with running d.net on them.

Fuzzy
-- 
"We could never learn to be brave and patient, if there were only joy
in the world." -- Helen Keller

On 8/24/05, Jim C. Nasby <decibel at distributed.net> wrote:

> On Wed, Aug 24, 2005 at 04:36:26PM -0400, Fuzzy Logic wrote:
> > Actually, I'm not surprised. Often, the 64-bit version of applications
> > performs slower than the 32-bit version.
> >
> > Unfortunately, I don't have any 64-bit boxes on which I am allowed to
> > run D.net clients, so I haven't really had a chance to try it out.
> 
> Opteron box, running 32bit RHEL:
> uname -a:
> Linux 2.6.9-1.648_ELsmp #1 SMP Tue Oct 26 12:55:36 EDT 2004 i686 athlon i386 GNU/Linux
> 
> [Aug 25 00:32:46 UTC] Automatic processor type detection found
>                       an AMD K8-5 (Opteron 8xx) processor.
> [Aug 25 00:32:46 UTC] RC5-72: using core #0 (SES 1-pipe).
> [Aug 25 00:33:05 UTC] RC5-72: Benchmark for core #0 (SES 1-pipe)
>                       0.00:00:16.77 [5,264,046 keys/sec]
> [Aug 25 00:33:05 UTC] RC5-72: using core #1 (SES 2-pipe).
> [Aug 25 00:33:24 UTC] RC5-72: Benchmark for core #1 (SES 2-pipe)
>                       0.00:00:17.01 [8,835,532 keys/sec]
> [Aug 25 00:33:24 UTC] RC5-72: using core #2 (DG 2-pipe).
> [Aug 25 00:33:43 UTC] RC5-72: Benchmark for core #2 (DG 2-pipe)
>                       0.00:00:16.37 [9,421,329 keys/sec]
> [Aug 25 00:33:43 UTC] RC5-72: using core #3 (DG 3-pipe).
> [Aug 25 00:34:01 UTC] RC5-72: Benchmark for core #3 (DG 3-pipe)
>                       0.00:00:16.25 [9,248,771 keys/sec]
> [Aug 25 00:34:01 UTC] RC5-72: using core #4 (DG 3-pipe alt).
> [Aug 25 00:34:20 UTC] RC5-72: Benchmark for core #4 (DG 3-pipe alt)
>                       0.00:00:16.84 [9,792,946 keys/sec]
> [Aug 25 00:34:20 UTC] RC5-72: using core #5 (SS 2-pipe).
> [Aug 25 00:34:40 UTC] RC5-72: Benchmark for core #5 (SS 2-pipe)
>                       0.00:00:16.87 [9,743,657 keys/sec]
> [Aug 25 00:34:40 UTC] RC5-72: using core #6 (GO 2-pipe).
> [Aug 25 00:34:58 UTC] RC5-72: Benchmark for core #6 (GO 2-pipe)
>                       0.00:00:16.27 [10,797,165 keys/sec]
> [Aug 25 00:34:58 UTC] RC5-72: using core #7 (SGP 3-pipe).
> [Aug 25 00:35:16 UTC] RC5-72: Benchmark for core #7 (SGP 3-pipe)
>                       0.00:00:16.15 [9,496,617 keys/sec]
> [Aug 25 00:35:16 UTC] RC5-72: using core #8 (MA 4-pipe).
> [Aug 25 00:35:35 UTC] RC5-72: Benchmark for core #8 (MA 4-pipe)
>                       0.00:00:16.52 [5,793,127 keys/sec]
> [Aug 25 00:35:35 UTC] RC5-72: using core #9 (MMX 4-pipe).
> [Aug 25 00:35:54 UTC] RC5-72: Benchmark for core #9 (MMX 4-pipe)
>                       0.00:00:16.51 [4,902,304 keys/sec]
> [Aug 25 00:35:54 UTC] OGR-P2: using core #0 (GARSP 6.0-A).
> [Aug 25 00:36:12 UTC] OGR-P2: Benchmark for core #0 (GARSP 6.0-A)
>                       0.00:00:16.41 [20,800,015 nodes/sec]
> [Aug 25 00:36:12 UTC] OGR-P2: using core #1 (GARSP 6.0-B).
> [Aug 25 00:36:31 UTC] OGR-P2: Benchmark for core #1 (GARSP 6.0-B)
>                       0.00:00:16.47 [14,591,146 nodes/sec]
> [Aug 25 00:36:31 UTC] OGR-P2: using core #2 (GARSP 6.0-asm-rt1-gen).
> [Aug 25 00:36:50 UTC] OGR-P2: Benchmark for core #2 (GARSP 6.0-asm-rt1-gen)
>                       0.00:00:16.38 [23,819,202 nodes/sec]
> 
> Same class machine, running 64bit:
> uname -a
> Linux 2.6.9-5.ELsmp #1 SMP Wed Jan 5 19:29:47 EST 2005 x86_64 x86_64 x86_64 GNU/Linux
> 
> [Aug 25 00:35:59 UTC] RC5-72: using core #0 (SNJL 3-pipe).
> [Aug 25 00:36:19 UTC] RC5-72: Benchmark for core #0 (SNJL 3-pipe)
>                       0.00:00:16.83 [9,788,559 keys/sec]
> [Aug 25 00:36:19 UTC] RC5-72: using core #1 (KBE-64 3-pipe).
> [Aug 25 00:36:38 UTC] RC5-72: Benchmark for core #1 (KBE-64 3-pipe)
>                       0.00:00:17.02 [10,298,565 keys/sec]
> [Aug 25 00:36:38 UTC] RC5-72: using core #2 (ANSI 4-pipe).
> [Aug 25 00:36:56 UTC] RC5-72: Benchmark for core #2 (ANSI 4-pipe)
>                       0.00:00:16.37 [4,288,957 keys/sec]
> [Aug 25 00:36:56 UTC] RC5-72: using core #3 (ANSI 2-pipe).
> [Aug 25 00:37:15 UTC] RC5-72: Benchmark for core #3 (ANSI 2-pipe)
>                       0.00:00:16.69 [3,711,458 keys/sec]
> [Aug 25 00:37:15 UTC] RC5-72: using core #4 (ANSI 1-pipe).
> [Aug 25 00:37:33 UTC] RC5-72: Benchmark for core #4 (ANSI 1-pipe)
>                       0.00:00:16.15 [2,964,119 keys/sec]
> [Aug 25 00:37:52 UTC] OGR-P2: Benchmark for core #0 (GARSP 6.0-64)
>                       0.00:00:16.26 [35,400,722 nodes/sec]


More information about the rc5 mailing list