Author: Vincent Diepeveen
Date: 05:30:04 07/08/03
Go up one level in this thread
On July 07, 2003 at 16:08:14, Joachim Rang wrote: >On July 07, 2003 at 11:34:11, Vincent Diepeveen wrote: > >>On July 07, 2003 at 10:36:00, Joachim Rang wrote: >> >>>On July 07, 2003 at 09:54:44, Vincent Diepeveen wrote: >>> >>>>On July 07, 2003 at 06:27:01, DANIEL JOHNSON wrote: >>>> >>>>>I PLAN ON BUILDING A NEW DUAL PROCESSOR COMPUTER WHEN THE NEW .09 CHIPS COME OUT >>>>>WITH 1 MEG CACHE MEMORY AND HAS 800 MHZ FSB ABOUT 3GB MEMORY, DUAL SATA RAID 0 >>>>>HDD, SWIFTECH WATERCOOLING, BUT WHICH CPU WILL BE BEST AND FASTEST FOR CHESS >>>>>OPTERON OR XEON >>>> >>>>Opteron will be trivially in 0.09 >>>> >>>>The question will be however whether the Xeon is there first in 0.09 or Opteron. >>>> >>>>I bet at Xeon in 0.09 being there first. >>>> >>>>As you can see at specint crafty at 64 bits opteron is 1562 rated. this is just >>>>1.8Ghz opteron. >>>> >>>>I expect for other programs when compiled 64 bits and with 16 GPRs and all the >>>>other features to be equally fast (having same speed difference). Fritz is >>>>assembly. In order to run in 64 bits at opteron Frans Morsch must now already be >>>>working at that as for every new processor he has to rewrite nearly the entire >>>>program. Must be hard worker that guy! I do not know whether he is busy with >>>>that now. I can ask if you want to. >>>> >>>>For most of the others it is just a simple recompile though. The work i need in >>>>diep to profit from having more bits in each register (the other advantages you >>>>get for free or are compiler issues) is not little either. It will give a few % >>>>though. >>>> >>>>When opteron is there in 0.09 it will be way faster, the chip is actually >>>>designed to be 0.09. It is however interesting to see what speed the prescott >>>>core will have in 0.09, i guess they will be sooner there. >>>> >>>>Best regards, >>>>Vincent >>> >>>Vincent, >>> >>>will you use an opteron machine in Graz this year? >>> >>>regards Joachim >> >>I would if the government didn't provide me with Europes fastest machine for >>computerchess available: 1440 processor TERAS machine. 1024 MIPS R14000 >>processors and 416 processor Itanium2 1.3Ghz Madisons. >> >>I'll be running hopefully at 500 processor partition (500 x 500Mhz = 250Ghz). >> >>Thanks to the dutch government and NWO, NCF, computermanufacturer SGI, WGS, >>IKAT, University of Maastricht and another number of organisations i forgot to >>mention here (but not deliberately) i am sure diep will try to do its utmost >>best in world champs 2003. >> >>Best regards, >>Vincent > > >Do ypu think this time you will get enough time to test and program your engine >for such a monster? I remember that last year you switched back to your Dual, >because you hadn't enough time to develop your engine fo so many processors. >regards Joachim Previous time i could test a few afternoons on a quad SGI, thanks to SGI. However i didn't get access to the TERAS machines until 3 days ahead. The latency of such partitions is more than 10 times worse than at a quad. Trivially that caused problems. July is also the maintaining month of supercomputers. Every scientist is having holiday then, so it is a great time for them to upgrade supercomputers (in july 2002 they upgraded harddisks so compile times were temporarily about 30 minutes for DIEP, versus several seconds now). That happened too in these days. I have had enough time to rewrite DIEP for this machine now. DIEP is ready for the world champs 2003! Of course still many small things you keep doing. It will be great now! Testing at 500 processors at the same time is pretty hard though. Majority of tests will be with like 130 cpu's. What i usually do with a new version is first by hand start it at 1 cpu. Then dual very shortly. Then a minute or 2 i start by hand 4 processors. Sometimes i try by hand half a minute 8 processors then just to see whether it doesn't crash. If it doesn't then i schedule first some 32 processor tests and after that a single 128 or 130 processor test. From now on that will be only 130 processor tests not 128. The TERAS supercomputer consists out of 2 machines. 1 machine with 512 cpu's that forms together 500 processors and 1 machine with several 'partitions' of 32 32 64 128 256 processors. Only when you schedule a program at 130 processors to be tested for say 10 minutes, then trivially it will be run on the 'big machine'. A major problem with testing is that it is very hard to just modify small things, because if you do so, waiting for a 130 processor test of 10 minutes to be run, this can take a full week. On average 3 days. That tells basically how succesful the NWO/NCF is. The machine nowadays is fully loaded. Previous years that wasn't the case. But 2003 a lot of scientists make use of the machine now. Therefore the new machine with 416 processors (32 + 64 + 64 + 256 processor partitions) is a great thing to have for the NWO/NCF. Best regards, Vincent
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.