Author: William Bryant
Date: 19:38:39 11/26/04
Go up one level in this thread
First, Thanks a lot. Note: When I use -03 instead of -02 I get an extra 100K NPS. (Which is intersting since on my previous computer I was getting 100K NPS). On the latests bench, I got 3.11M NPS on the bench test. I am going to like my new apple. (Dual 2.5gHz G5). William On November 26, 2004 at 20:23:47, Andreas Guettinger wrote: >On November 25, 2004 at 18:40:29, William Bryant wrote: > >>I have a new computer, and wanted to bench mark crafty. >> >>The crafty ftp site is missing the crafty doc's at the moment. >> >>I want help understanding the output of bench: >> >>Running benchmark. . . >>...... >>Total nodes: 76469107 >>Raw nodes per second: 0 >>Total elapsed time: 25 >>SMP time-to-ply measurement: 25.600000 >>White(1): >> >>I assume the lower the time, and the higher the SMP time-to-ply the better. >> >>The raw nodes per second at 0, is this an interger data overflow problem? > >No, it's an integer conversion error. The line 159 in bench.c had to be changed >for ppc compatibility: > >Print(4095, "Raw nodes per second: %d\n", (int) (nodes / (total_time_used / >100))); > >> >>Could I get a quick explination of the SMP time-to-ply value? >> >>Also, >>I would appreciate a updated make file for crafty for a dual G5. >>I made the changes based on previous posts but could use some help with command >>line programming. >>Make files, include files, and file chanages would help. >> >>William > >A Makefile optimized for Dual G5 plus the above bench.c fix plus some inline >functions for ppc can be found below. I didn't change any other crafty code >except of this enhancements and fixes for ppc. > >http://home.datacomm.ch/aguettinger/files/crafty-19.17ppc.tar.gz > >cu >Andy
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.