Author: Robert Hyatt
Date: 21:29:46 12/18/03
Go up one level in this thread
On December 18, 2003 at 16:20:48, Rolf Tueschen wrote: >On December 18, 2003 at 16:07:44, Robert Hyatt wrote: > >> >>It's a different way of thinking about it when you think about it. Those >>extra CPUs don't just magically make the program faster without a _lot_ of >>design effort and programming work. > > >What is the specific difficulty? No, dont tell me but our actual problem childs >need an understanding because they seem to have incredible difficulties. How >about making extra classes with Gian Carlo and Vincent in a special room on ICC? > Or is this naive to think of? Is it impossible to advise someone whose code is >unknown to you? I really would like to see the two compete with the big ones. >You know it becomes boring with the three top engines. > >Of course the arrogant kids would yell "troll, piss off!" :) > The basic problem is just what I said. Making a parallel program work is not an easy task. There are lots of interactions between the processes (threads) that require specific design details to solve. That turns into time, time that could be spent on non-parallel things. So saying that a program has an unfair advantage when it runs on equal hardware with another program, but the other program can't use the extra CPU is simply wrong. Because _that_ program has an advantage also, the programmer spent more time on the serial engine while the parallel programmer was working on the parallel issues. Why that seems to elude people is beyond me... Having done _both_ gives me a dual perspective. And I'm not talking about two processors _there_. :) >Rolf
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.