Author: Matt Taylor
Date: 15:22:53 12/19/02
Go up one level in this thread
On December 19, 2002 at 17:55:47, Robert Hyatt wrote: >On December 19, 2002 at 16:45:44, Matt Taylor wrote: > >>On December 19, 2002 at 11:48:42, Robert Hyatt wrote: >> >>>On December 19, 2002 at 05:00:05, Matt Taylor wrote: >>> >>>>On December 19, 2002 at 02:57:11, Brandon wrote: >>>> >>>>>AMD chips are generally better (mhz per mhz) for chess than Intel chips. I use a >>>>>dual xeon 2.2 ghz for chess on ICC and I get some pretty crazy nps (I use chess >>>>>programs that use multiple cpus). Generally speaking, if the machine is going to >>>>>play chess only and is going to be a single cpu system, I'd stick with AMD.. >>>>>best bang for the buck in this case. However, if you are going to be doing video >>>>>editing or using "graphic intensive programs", Intel generally is better. If you >>>>>want to have a multiple cpu system (like a dual cpu system), I would stick with >>>>>Intel, as they have been in the multiple cpu business for many years while AMD >>>>>is pretty recent in this area (~1.5 years experience, at the most). I have heard >>>>>reports of inefficiencies and problems with dual amd configurations, so research >>>>>it out carefully. www.tomshardware.com is a good place to start... good luck. >>>>> >>>>> - Brandon S. >>>>> >>>>>P.S. - Programs like Crafty have been compiled by several different sources to >>>>>provide optimizations for the P4 and the AMD, so that muddies up the whole issue >>>>>of "which cpu is faster with chess progs" and what not.. >>>> >>>>Tom's Hardware posts a lot of crap. Half the articles are poor attempts to cover >>>>up lack of knowledge. When they do make an error, instead of fixing the problem, >>>>they try to explain to you why it's not an error. >>>> >>>>I own an AthlonMP system at home and have one on my desk at work. I haven't seen >>>>any inefficiencies; they do lack the quad-pumping stuff Intel does. So does the >>>>P3 Xeon. >>>> >>>>Overall quite happy with my system. It doesn't perform as well as high-end P4 >>>>Xeon systems, but it's hard to beat with a pricetag of $1,100 and comparable >>>>hardware minus SCSI. >>>> >>>>-Matt >>> >>> >>>My only comment is that I will _never_ own another IDE-based system. too slow. >>>Hogs the bus. Devices are slow. >> >>I have IDE RAID. It's not SCSI, but it's SCSI-like. No more bus hogging, and >>slow devices aren't as big of an issue. The IDE RAID cost me an extra $300 for >>the controller and a pair of disks. >> >>>My 2.8 is using ultra-320 scsi with 15K drives and it can eat EGTBs like a >>>gorilla eats bananas... >>> >>>SCSI is also a nice way to offload queueing issues as well. Let the >>>controller decide which read/write to do next... Since an operating system >>>really can't uderstand variable device geometry anyway... >> >>Yes, which is why it's nice that my IDE RAID controller functions as a SCSI >>device. :-) >> >>-Matt > > >Now show me your 320mbyte/sec bursts and 15K rpm (2ms latency) rotational >speeds.. :) Around 700 MB/sec burst in 8-way. I only have 2-way, so I only get about 120 MB/sec burst. Of course, SCSI RAID hits even higher speeds, and my latency is still not as nice as SCSI. The latency is mostly due to spindle speed, I think. I haven't seen any 10k or 15k rpm IDE disks for sale. -Matt
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.