Author: Robert Hyatt
Date: 16:12:24 02/16/03
Go up one level in this thread
On February 16, 2003 at 18:17:32, Jeremiah Penery wrote: >On February 16, 2003 at 16:50:48, Brian Richardson wrote: > >>I have been running some informal Itanium2 tests with Tinker and Crafty (18.15). >>The results are not encouraging. I know Bob Hyatt has posted better numbers >for > >Every piece of data I've seen contradicts Bob's claim. As a wild guess, I might >say he saw multi-CPU numbers that either got reported as single-CPU, or someone >just assumed they were single-CPU numbers. However, I haven't actually seen the >numbers Bob saw - there is some chance that what he reported is accurate. Eugene supplied some of the numbers. Someone inside intel supplied others. Since, as I have said, I have never had my hands on one of these things, I don't have any test results that I can publish as "run by me" as I can post for my various hardware choices I do have access to. However, the alpha numbers in SPEC are way below the numbers that I _do_ have logs for for a 21264 at 600mhz: total positions searched.......... 300 number right...................... 300 number wrong...................... 0 percentage right.................. 100 percentage wrong.................. 0 total nodes searched.............. 236973211.0 average search depth.............. 4.5 nodes per second.................. 783641 That was one cpu. The next is for a dual 21264 at 600mhz: total positions searched.......... 300 number right...................... 300 number wrong...................... 0 percentage right.................. 100 percentage wrong.................. 0 total nodes searched.............. 330905102.0 average search depth.............. 4.5 nodes per second.................. 1266767 This version was _prior_ to our "lockless hash algorithm" which made the SMP scaling much better on the alpha, in terms of raw NPS... The above I can produce the log files for, for anyone interested in seeing the results for all 300 wac positions... > >>Itanium2, but I can’t come even close. Then again, I am no professional >>compiler developer or performance engineer. > >As you pointed out near the very end, the Intel/HP people who submitted the SPEC >scores didn't do much better. I would hope that they _are_ professional >performance engineers. > >> Anyway, here is the data (Knps). > >[SNIP] > >>I was not able to get profiling to work, nor did I try Vtune. > >From what I understand, profiling is quite important for Itanium to get good >results. This is why even the base SPEC scores were achieved with profiling >runs, AFAIK. > >>The results are more or less in line with the SPEC CPU2000int results for Crafty >>(see www.spec.org), some of which are (run times in seconds): >> >>Hewlett-Packard Company hp server rx2600 (1000 MHz, Itanium 2) >>186.crafty 128 >> >>Advanced Micro Devices Epox 8KHA+ Motherboard, AMD Athlon (TM) XP 1900+ (CPU >>MHz: 1600) >>186.crafty 106 >> >>Intel Corporation Intel D850EMVR motherboard (3.06 GHz, Pentium 4 processor with >>HT Technology) >>186.crafty 86 >> >>Advanced Micro Devices ASUS A7N8X (REV 1.02) Motherboard, AMD Athlon (TM) XP >>2800+ (CPU MHz: 2250) >>186.crafty 76.3
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.