Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: GCC 2.95 big step forward

Author: Bernhard Bauer

Date: 23:54:46 08/09/99

Go up one level in this thread


On August 09, 1999 at 13:34:27, José de Jesús García Ruvalcaba wrote:

>On August 09, 1999 at 03:49:03, Bernhard Bauer wrote:
>
>[big snip]
>>
>>Here just some observations using gcc2.95.
>>On a Sun Ultra Sparc (2 proc) I tried to compile Crafty 16.15.
>>For the following compiler flags I got:
>>
>>-O  worked.
>>-O2 core dumped.
>>-O3 core dumped.
>>
>>So I'm able to produce a working crafty, but no higher optimization than -O.
>>
>>Kind regards
>>Bernhard
>
>	What NPS do you get with the 'bench' command? Have you tried the version that
>is in Dann Corbit's ftp site? If so, what NPS does it get in your machine? Just
>be careful with Solaris inconsistencies in reporting time.
>José.

Yes, I tried the version from Dann's site.
->crafty-sparc-16.15
->mt 2
->bench
gives
Total nodes: 49775279
Raw nodes per second: 94450
SMP time-to-ply measurement: 7.637681
The nps numbers are

              time=1:23   cpu=396%  mat=0  n=7466313   fh=90%  nps=89599
              time=2:09   cpu=396%  mat=0  n=13177598  fh=93%  nps=101592
              time=40.14  cpu=393%  mat=0  n=4053830   fh=90%  nps=100992
              time=37.42  cpu=395%  mat=0  n=4016970   fh=92%  nps=107348
              time=1:15   cpu=396%  mat=0  n=7316099   fh=94%  nps=96403
              time=2:40   cpu=397%  mat=0  n=13744469  fh=91%  nps=85613
which look not too bad.


My own version generated with gcc2.95 gives:
Total nodes: 49174317
Raw nodes per second: 111506
SMP time-to-ply measurement: 6.391304
The nps numbers are:
              time=1:07   cpu=394%  mat=0  n=7331011   fh=90%  nps=108946
              time=1:51   cpu=397%  mat=0  n=13092288  fh=93%  nps=116926
              time=34.62  cpu=392%  mat=0  n=4135080   fh=90%  nps=119441
              time=31.46  cpu=393%  mat=0  n=3908866   fh=92%  nps=124248
              time=1:04   cpu=396%  mat=0  n=7346462   fh=94%  nps=114770
              time=2:12   cpu=397%  mat=0  n=13360610  fh=91%  nps=101094
which look better.

Kind regards
Bernhard




This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.