Author: Robert Hyatt
Date: 20:01:04 03/18/03
Go up one level in this thread
On March 18, 2003 at 18:18:44, Brian Richardson wrote: >On March 18, 2003 at 18:08:17, Robert Hyatt wrote: > >>On March 18, 2003 at 16:57:17, Brian Richardson wrote: >> >>>On March 18, 2003 at 16:31:58, Robert Hyatt wrote: >>> >>>>On March 18, 2003 at 15:17:30, Dann Corbit wrote: >>>> >>>>>On March 18, 2003 at 14:04:56, Vincent Diepeveen wrote: >>>>> >>>>>>On March 17, 2003 at 19:49:30, leonid wrote: >>>>>> >>>>>>>On March 17, 2003 at 19:30:35, Dann Corbit wrote: >>>>>>> >>>>>>>>On March 17, 2003 at 19:17:38, leonid wrote: >>>>>>>>[snip] >>>>>>>>>It is truly sad that this kind of chip (Alpha) went to its end! Still, other >>>>>>>>>fact of migration to the new Itanium chip is very positive. This signify that >>>>>>>>>Intel's 64 chip had certain success and its quantity production is coming in >>>>>>>>>real. Then prices cuts should become tangible very soon. >>>>>>>> >>>>>>>>Intel is in no hurry to get the Itanic floating, and has said so themselves. >>>>>>>> >>>>>>>>Look for AMD to beat them out the door with cost effective volumes by a >>>>>>>>landslide. >>>>>>>> >>>>>>>>Microsoft also prefers the AMD approach, since all the old software will run >>>>>>>>without any modification. >>>>>>> >>>>>>>To be sincere I do know that AMD chip will be more accessible for me to buy but >>>>>>>Intel's chip more attractive to program. Intel's 128 registers do make me dream >>>>>>>without even mentioning its new architecture. Learning completely new Assembler >>>>>>>will be also interesting thing to do. >>>>>> >>>>>>$65000 for a 4 processor 1Ghz I2 box. >>>>>> >>>>>>Or if you buy 1 chip in a small 4 x 4 centimeter paper box it will be 'only >>>>>>$10000' or so. >>>>>> >>>>>>And that price won't get cheaper at all >>>>>> >>>>>>So forget itanium2 unless you have access to a supercomputer that has them. >>>>> >>>>>Pricing information (cheapest I could find): >>>>>http://www.hp.com/workstations/products/itanium/zx6000/summary.html >>>> >>>> >>>>Not bad. Dual for just over $6,000 would be a blast for a certain chess >>>>program. :) >>> >>>I have a dual Itanium2 1GHz system. Perhaps you forgot my earlier post. >>>With the optimizing compilers (tried both Intel and Microsoft), it runs >>>Tinker at about 2GHz Pentium speed (x86 binary code at only 30%). >>> >>>Has anyone been able to reproduce your (Eugene's?) results showing MUCH faster >>>Itanium2 performance? >>> >>>Brian >>> >>>PS No integrated Visual Studio environment or debugger, working with command >>>line interfaces, and so far pretty disappointing. >> >>The numbers from Intel were good, but they were on a single-cpu Itanium-2 >>machine. >> >>Eugene's were also from a single-cpu Itanium-2 if I recall correctly. >> >>One issue might be the compiler. Eugene obviously has the latest since he is in >>the >>compiler group at MS. Intel may be using a beta MS compiler or their own, I >>don't >>know. > >This is _the_ point. Itanium has been on the market for about 18 months now >(counting from Merced). Regular folks using the most recent generally available >compilers do not see the performance hoped for (other than floating point). >Saying Eugene can make it run fast is like saying, well Hsu could build a faster >hardware engine today...what about for the rest of us? >Bob, do you even have an Itanium system? It sounds like no. I have to think if >it was such a good thing, you would have found some $$$ for one. I've answered that in the past. No Itanium here at all, yet. I'm more interested in parallel programming issues than in the fastest possible single-cpu machine I can find. Otherwise I would also have a 21264 alpha here as well. I have an older alpha, but again it is a single cpu machine and that's not as interesting to me. Don't forget that with work, the Cray would be the fastest of all, but I don't have one of those either. :)
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.