Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: TSCP benchmark with gcc vc6 vc.net on Athlon XP and P4

Author: Matt Taylor

Date: 05:44:36 02/21/03

Go up one level in this thread


On February 21, 2003 at 07:16:59, martin fierz wrote:

>On February 20, 2003 at 18:53:25, Russell Reagan wrote:
>
>>On February 20, 2003 at 17:54:12, martin fierz wrote:
>>
>>>-------------------------------------------
>>>version        XP 2400+    P4 1.4GHz
>>>-------------------------------------------
>>>"original"     2260 MIPS   1587
>>>VC 6           2373        1579
>>>VC.net         2496        1587
>>>gcc for P4     2420        1480
>>>gcc for AMD 1  2523        1451, then crash
>>>gcc for AMD 2  2663        crash
>>>-------------------------------------------
>>
>>If I'm reading this right, gcc is better for AMD processors, and MSVC is better
>>for Intel processors. Any idea how the intel compiler compares?
>
>that's what i read into these numbers too. with the big IF that i'm not sure
>whether TSCP is a good benchmark for "real" chess programs - it may be too
>simple. i'll definitely try gcc with my checkers program and see what happens
>there.
>i don't have the intel compiler. but since i have included the comparison with
>the "original" version of TSCP which is downloadable, someone with the intel
>compiler could compare his version with what i got for the AMD; 18% faster than
>the original on an athlon XP.
>the downside is of course that the AMD version doesn't run on a P4...
>
>cheers
>  martin

Which is strange. What could they be doing that would crash on a P4? 3DNow? GCC
generates 3DNow code? Yeah, right... :-)

-Matt



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.