Author: Robert Hyatt
Date: 06:23:43 02/21/03
Go up one level in this thread
On February 21, 2003 at 03:32:44, Tom Kerrigan wrote: >On February 20, 2003 at 22:07:12, P. Massie wrote: > >>I'm not an expert on HT, but based on what I've read about it, and what I know >>about how computers work I suspect it will be somewhat better than a "normal" >>processor for this, but not nearly as good as a true dual. My suggestion would >>be a dual AMD or Xeon. > >Actually, any sort of stuttering/unusability you get from multitasking on one >CPU is because of a poor scheduling algorithm in your operating system (or at >least one that leaves room for improvement). Because HT presents one processor >as two to the OS, that scheduling problem goes away. It would not surprise me if >HT chips were dramatically more responsive (although not that much faster) when >multitasking, although I'm not saying this is a certainty. I have never used a >HT system myself. > >-Tom This really isn't true. It is _unavoidable_ that some "stuttering" is going to happen, particularly if you use IDE disk drives. What can you do once you start an I/O on a slow IDE disk, and then an I/O request comes in from something that is "higher priority"?? You are basically stuck. With SCSI this is far less significant as the I/O can be started and the SCSI controller can possibly pick that one up on the way to the current cylinder. But if the machine is heavily loaded, everything is going to run slower and there is nothing to do about that either but to either reduce the load, or get a faster (or more processors) box. I don't think there is _any_ system that can deliver seamless performance, even unix/linux using the "nice" value for background processes still runs into bottlenecks dealing with I/O throughput.
This page took 0.01 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.