Author: Robert Hyatt
Date: 08:33:17 12/13/02
Go up one level in this thread
Since it looks like Vincent has "cut and run" from this discussion, it appears to be over. As always, cold hard data is something he can't cope with. For those that don't think hyper-threading exists today, all you have to do is visit Intel's web site to see that _all_ current xeon processors come with hyper-threading, and the 3.06ghz PIV also comes with it. (note that contrary to vincent's rambling nonsense, the PIV is _not_ the same as the xeon processor and the Intel note about the PIV at 3.06 being the _only_ hyper-threaded PIV is correct but totally unrelated to the xeon processor's ability to do SMT. HT (SMT) works just fine. I've experimented with it a bit and have found a couple of ways to make a single SMT cpu run twice as fast, although this is not going to happen with Crafty because it doesn't have nearly the type of instruction mix to make this work. But it can happen. Running Eugene's compression/decompression code (two copies) is one good way to see the thing run like a bat... There are others. And apparently two logical processors in one physical processor is just the beginning. As I said a few years ago when everyone was saying "who needs a parallel program? " it won't be _much_ longer before _every_ machine appears to have more than one CPU. It is coming. And it will work. Regardless of hand-waving nay-sayers. It is a logical development of removing one more time-critical piece of code from the Operating System into the microprocessor, namely the task scheduler. By the way, this machine that doesn't exist is beginning to impress me quite a bit. The first round of SMT looks interesting. Can't wait for the "second round".
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.