Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Crafty 64-bit speedup

Author: Eugene Nalimov

Date: 10:08:59 03/26/04

Go up one level in this thread


On March 26, 2004 at 11:50:24, Slater Wold wrote:

>On March 26, 2004 at 10:59:30, Eugene Nalimov wrote:
>
>>On March 26, 2004 at 01:04:54, Slater Wold wrote:
>>
>>>All week I have been hacking away at making the fastest 32-bit & 64-bit binary
>>>with the AMD64 compiler I have access to.  I have been told that they are the
>>>same compiler; just one has the AMD64 additions.  Here are my results.
>>>
>>>
>>>Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 13.10.2179 for 80x86
>>>Copyright (C) Microsoft Corporation 1984-2002. All rights reserved.
>>>Microsoft (R) Incremental Linker Version 7.10.2179
>>>Copyright (C) Microsoft Corporation.  All rights reserved.
>>>
>>>Using /O2, /Ot, /Og:
>>>
>>>Total nodes: 89911558
>>>Raw nodes per second: 1550199
>>>Total elapsed time: 58
>>>SMP time-to-ply measurement: 11.034483
>>>
>>>Microsoft (R) C/C++ Optimizing Compiler Version 14.00.2207 for AMD64
>>>Copyright (C) Microsoft Corporation 1984-2002. All rights reserved.
>>>Microsoft (R) Incremental Linker Version 8.00.2207
>>>Copyright (C) Microsoft Corporation.  All rights reserved.
>>>
>>>Using /O2, /Ot, /Og:
>>>
>>>Total nodes: 89913841
>>>Raw nodes per second: 2140805
>>>Total elapsed time: 42
>>>SMP time-to-ply measurement: 15.238095
>>>
>>>
>>>
>>>The fastest binary I can find *anywhere*, has been Russell's.  He compiled it
>>>using all optimizations using .NET 2003 (he used VC_INLINE_ASM & ASM_INLINED,
>>>which my compiler cannot use).  Here is its result:
>>>
>>>Total nodes: 89908309
>>>Raw nodes per second: 1696383
>>>Total elapsed time: 53
>>>SMP time-to-ply measurement: 12.075472
>>>
>>>
>>>
>>>
>>>More engines to come soon...
>>
>>Please add /GL to the command line.
>>
>>Thanks,
>>Eugene
>
>WPO slows down the 64-bit compile (from 214 to 209).  And speeds up the 32-bit
>compile (from 155 to 163).

You probably have very old 64-bit compiler. During early development everybody
works on correctness, not on quality of generated code...

Thanks,
Eugene



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.