Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Intel Hyperthreading and Ponder (Permanent Brain)

Author: Matthew McKnight

Date: 08:47:30 01/02/04

Go up one level in this thread


On January 02, 2004 at 11:45:57, Matthew McKnight wrote:

>On January 02, 2004 at 11:24:19, Steven J. Brann wrote:
>
>>On January 02, 2004 at 11:05:41, Anthony Cozzie wrote:
>>
>>>On January 02, 2004 at 09:57:38, Jasmine Baer wrote:
>>>
>>>>I've seen it written that under the following conditions:
>>>>
>>>>1.  Engine vs. Engine match or tournament
>>>>2.  Held on a single computer with a single processor
>>>>
>>>>having ponder=ON(or Permanent Brain in the Fritz GUI) will impact the play of
>>>>the engines since the each individual engine would not have full access to the
>>>>processor during its own turn.
>>>>
>>>>First, is this true?
>>>>
>>>>Second, is this issue, if it actually is an issue, something that is eliminated
>>>>by running a two-processor system?
>>>>
>>>>And, finally, does anyone have any solid insight on how ponder=off/on or
>>>>Permanent Brain works on a Pentium 4 with Hyperthreading?
>>>>
>>>>Thanks.
>>>
>>>Ponder means that the engine thinks while its opponent moves.  Since there is
>>>only 1 cpu, and both engines are thinking, they get half the cpu.
>>>
>>>HT is garbage for computer chess.  A pentium 4 is ONE core.  HT is designed for
>>>applications that spend most of their time in the memory system.
>>>
>>>anthony
>>
>>To me, my 3.0G HT machine is NOT garbage for computer chess.  When it is
>>thinking about a position it takes up 50% of the CPU and is still much much
>>faster than my 1.9G P4 machine.  When analyzing a position with my 1.9G P4, the
>>machine would be rendered useless for using any other application while it was
>>thinking about a position.
>>
>>So, HT enables me to accomplish other things on the machine at the same time ...
>>email, reviewing this site, chat with video, Word, Excel ... and
>>performance-wise its as if the chess program isn't running at all.  I'm
>>analyzing a position as I write this.
>
>Yes this is what HT does, but it's also what every other modern CPU does...
>It's just Intel's latest buzzword for a well-known technology called pipelining.
> I believe what the author meant by the word "garbage" is simply that you
>shouldn't believe you _need_ a machine with HT for computer chess.
>
>Interestingly enough, pipelining benefits from you running more than one program
>at a time.  That is to say that if you were to run two processes, each using 50%
>of the cpu for 1 hour, they would get more work done than the sum of work done
>from both processes individually at 100% for one hour each.

DOH! I meant half an hour each.

>The increased
>difference between the types of programs in the experiment, increases the
>results.  This is a product of the enormous clock rate!! The only way to get
>such a speed is to try and run instructions simultaneously, and this is why you
>do better when you run two completely different programs (i.e. working on
>different things).
>
>Personally, I would love to have a PIV 3.0ghz HT regardless of HT's
>significance, that machine is smokin.
>
>>
>>That means a lot to ME.  Certainly not garbage in my opinion.
>>
>>Steve



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.