Author: Robert Hyatt
Date: 07:32:13 02/14/04
Go up one level in this thread
On February 14, 2004 at 08:45:37, Kurt Tank wrote: >On February 13, 2004 at 06:57:56, Vincent Diepeveen wrote: > >>On February 13, 2004 at 05:08:27, Peter Berger wrote: >> >>>On February 13, 2004 at 00:39:38, robert flesher wrote: >>> >>>>Nt >>> >>>Did he? >>> >>>Maybe this is a misunderstanding. You asked if the program will run using >>>multiple threads, and it will. >>> >>>You didn't ask about what performance improvement will be the result. HT will >>>bring an additional 0%-30% depending on engine and position, which is much lower >>>than what you would get from a multi-processor computer. As the computer still >>>only has one physical processor, that's also logical. >>> >>>I recently bought a PIV3.2 with HT ability, but have only tried it with Crafty >>>so far, when it is about "Deep" programs. For Crafty 19.10 +20% looks like a >>>reasonable speedup number on average. >>> >>>Peter >> >>Additionally it will not work well for Shredder for example which was the >>original question. > >I have dual xeon 2.8 and when i turn Hyper-threading off i have lower Kn/s for >Shredder or Deep junior 8 for example . >I think a Quad is far more competitive as it has 4 physical processors. I don't see why _any_ program that does reasonably well at a real dual would fail to do reasonably well with a hyper-threaded CPU, myself. I've not seen anyone post any data that suggests such that I can recall, other than when the "speculation" line is crossed. It certainly works for my program. I have a dual xeon that has SMT enabled, so that it looks like a quad... It is faster with SMT on than with SMT off, for Crafty...
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.