Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hammer info. And som SMP musings.

Author: Vincent Diepeveen

Date: 11:25:20 03/26/02

Go up one level in this thread


On March 26, 2002 at 13:50:02, Tom Kerrigan wrote:

>On March 26, 2002 at 12:01:30, Vincent Diepeveen wrote:
>
>>If you manage to turn on SMT at the P4 chip, sure :)
>>Intel has put on its homepage however that 'currently it is turned
>>off at the chip'.
>>hehehehehehe
>
>I don't see why you're laughing, because you pretty much just admitted that
>you're an idiot. How can you go around declaring that HyperThreading is a load
>of marketing crap that yields no performance gain when you have absolutely NO
>data on the issue?

You are the idiot here believing that Hyperthreading gives something
when they admit that the current P4s 'currently have it turned off'.

How can a feature work when it is 'turned off' in hardware, to use
intels own words?

Apart from that there is the theoretic discussion whether it would
*ever* work for a P4. the answer is no in all respects.

Too small L1 datacache and small L2 cache (with around 5 clocks penalty)




>>that in DIEP i can do (multiprocessing)
>>if( quickboard[sq_d3] == whitebishop )
>>where with multithreading you always need indirection:
>>if( thisthread->quickboard[sq_d3] == whitebishop )
>>SEE WHAT I MEAN?
>
>This is a chess-specific issue, not an issue with threading in general, and it's
>not even a performance problem depending on the addressing modes that your
>processor supports.
>
>But it doesn't even matter, because the P4 exposes itself as two logical
>processors, so you can just run two Diep processes on it. No need for threading.
>You would have known this if you had paid attention to my previous posts. Now
>you just look like an idiot.
>
>>>I can do math, but have you taken a computer organization class? It doesn't
>>>matter how big your stupid data structures are, the only thing that matters is
>>>how big your working set is.
>>I'm not the fool here Tom, please see above. You forget an overhead
>>which is pretty big and causes processors like the P4 to be not happy.
>
>You are the fool, because you can't follow a simple argument about working sets.
>What does your indirection argument have to do with cache thrashing??
>
>>>Really, how about you give us some measurements? Oh, right, you don't have any.
>>>You make these numbers up.
>>I'm at least measuring things where possible. You have had 1 theoretic
>>class where probably teachers told you thread = process. Well it isn't.
>
>Well, if you're measuring things where possible, then you should have no problem
>telling us EXACTLY how many instructions/clock Diep retires. So...?
>
>>In fact if you read in the API from M$ the first sentence then it will
>>show you even more trouble when multithreading under windows. It says
>>that when a thread is doing graphics, that all other threads are
>>getting blocked when one of the threads is performing work.
>
>Then don't do graphics with your chess threads.
>
>-Tom



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.