Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Experimentation with move ordering

Author: Jeremiah Penery

Date: 02:54:42 11/29/00

Go up one level in this thread


On November 29, 2000 at 04:38:33, Eugene Nalimov wrote:

>On November 29, 2000 at 02:53:42, Jeremiah Penery wrote:
>
>>On November 28, 2000 at 23:59:14, Ed Schröder wrote:
>>
>>>Especially multiplies has been improved dramatically in the latest
>>>generation of processors. Nowadays it is hardly an issue anymore. I
>>>still use << where ever I can but I have no problems to use * so now
>>>and then.
>>
>>It seems like compilers should produce the same assembly code for things like
>>a<<1 and a*2, but of course I'm not sure if they do.  Ditto for a>>1 and a/2.
>>(and also <<2 = *4, etc.)
>
>Of course a>>1 and a/2 should generate different code (hint: check value of both
>expressions for a == -1).

First, a question: If you divide 1 by 2 and have to return an int, is 1 or 0
returned?

If 0 is returned in the above, then to assembly for a>>1 and a/2 should be the
same if a is unsigned, correct?  If 1 is returned, then it seems like it would
work for all cases except when a==1.  Please correct me again if I'm wrong. :)



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.