Author: Vincent Diepeveen
Date: 06:15:19 02/29/04
Go up one level in this thread
On February 28, 2004 at 10:48:28, Slater Wold wrote: It was poor testing, because no optimizations default triggers -O2 in both intel c++ and that specific visual c++ compiler, but of course with gcc and old visual c++ it doesn't trigger -O2 by default. So you compare -O in 32 bits with -O2 at 64 bits. -O means no optimizations and -O2 means full optimizations but not function inlining and no dangerous (buggy) optimizations. So your comparision goes wrong because of the default triggering of optimizations. >I was comparing compilers...not CPUs. >With Intel's compiler, with no optimizations, the FX51 gets over 1.2M nps. >Both compilers came from the DDK, and are, per MS, exactly the same. They just >use different libraries. > >It wasn't poor testing. You just misinterpeted it. > > >On February 28, 2004 at 07:57:16, Vincent Diepeveen wrote: > >>On February 28, 2004 at 03:33:00, Slater Wold wrote: >> >>>cl from the Windows 2003 Server DDK used in all compilations. No flags >>>(optimizations) were used. >> >>Your results are not representing truth. >> >>>64-bit: >>>Microsoft (R) C/C++ Optimizing Compiler Version 14.00.2207 for AMD64 >>>Copyright (C) Microsoft Corporation 1984-2002. All rights reserved. >>> >>>32-bit: >>>Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 12.00.8804 for 80x86 >>>Copyright (C) Microsoft Corp 1984-1998. All rights reserved. >> >>Everyone cannot take your 32 bits results serious. >> >>They are a deliberate way to make 64 bits look better than it is. >> >>We are not fighting a intel/amd war here. >> >>This is too sick for words to do so. You know very well how to compile. >> >>Crafty at my K7 2.127Ghz (MP2600) even with slow msvc 7.1 compiler gets way >>above 1 million nps here. >> >>In 32 bits it'll get another 50% more at least at the opteron you're testing. >> >>> >>>Olithink: >>> >>>64-bit: >>>12 8 3121 22997307 b1-c3 d7-d5 d2-d4 g8-f6 c1-f4 c8-f5 g1-f3 b8-c6 e2-e3 >>> e7-e6 f1-b5 d8-d7 >>> >>>32-bit: >>>12 8 3284 22997307 b1-c3 d7-d5 d2-d4 g8-f6 c1-f4 c8-f5 g1-f3 b8-c6 e2-e3 >>> e7-e6 f1-b5 d8-d7 >>> >>> >>>TSCP: >>> >>>64-bit: >>>Nodes: 550778 >>>Best time: 1703 ms >>>Nodes per second: 323416 (Score: 1.330) >>> >>>32-bit: >>>Nodes: 550778 >>>Best time: 2875 ms >>>Nodes per second: 191574 (Score: 0.788) >>> >>> >>>Crafty: >>> >>>64-bit: >>>Total nodes: 102625951 >>>Raw nodes per second: 2052519 >>>Total elapsed time: 50 >>>SMP time-to-ply measurement: 12.800000 >>> >>>32-bit: >>>Total nodes: 102625951 >>>Raw nodes per second: 733042 >>>Total elapsed time: 140 >>>SMP time-to-ply measurement: 4.571429
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.