Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: MSVC code gen bug?

Author: Rafael Andrist

Date: 12:21:22 01/01/02

Go up one level in this thread


On January 01, 2002 at 13:38:57, Gian-Carlo Pascutto wrote:

>On January 01, 2002 at 13:25:25, Vincent Diepeveen wrote:
>
>>I guess he has only 1 pawn hash entry structure (not object).
>>
>>also more 8 bits code gives more register stalls. gives more buggy
>>code, and what is more important. Bugs or using a few bytes more?
>>
>>>
>>>Also, generally using more memory will cause more cache misses, and those
>>>cost many cycles.  Sometimes the panelty is smaller than the gain, but
>>>sometimes not.  I have sometimes had significant slowdowns because of using
>>>some more KB of data (happened to me with the random value arrays for the
>>>zobrist hash).
>>
>>you talk k6 times here ?
>>
>>L2 cache is like 512KB. not small. With DIEP i'm outside that anyway.
>
>Never the less, on the Athlon XP Sjeng runs faster when I make something
>as often used as my move data structure 8 bits instead of 32 !
>
>I seems that saving space in the cache is much more important than
>some extra instructions.

yes. the conversion from Byte to Doubleword needs one additional cycle and
waiting for data from RAM clearly takes longer

Rafael B. Andrist



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.