Author: Vincent Diepeveen
Date: 19:59:38 03/01/04
Go up one level in this thread
On March 01, 2004 at 18:45:34, Andreas Guettinger wrote: >On March 01, 2004 at 18:05:55, Vincent Diepeveen wrote: > >> >>That apple show was the show of a fool. >> >>If i remember well they first compiled with GCC at x86 hardware a few >>benchmarks. >> >>Then they compared that with gcc at the G5. >> >>Real sick compare. >> >>So all we know it's a slow CPU, that's all. How slow is not clear simply. >> > >LOL- You say some thing about exagerating. :)) > >I think what he meant is actual data somebody posted some days ago: > >>machine is dual 2GHz Powermac G5 >> >>wit default hash, here's results for 1 and 2 cpus >>============================================= >>Crafty v19.10 (1 cpus) >> >>White(1): bench >>...Total nodes: 89942714 >>Raw nodes per second: 1249204 >>Total elapsed time: 72 >>SMP time-to-ply measurement: 8.888889 >>============================================= >>Crafty v19.10 (1 cpus) >> >>White(1): mt=2 >>max threads set to 2 >>White(1): bench >>... >>Total nodes: 83479388 >>Raw nodes per second: 2455276 >>Total elapsed time: 34 >>SMP time-to-ply measurement: 18.823529 >>============================================= > > >Not as good as an Athlon64 3200, I agree, but nevertheless. > >Ruffian does over 1.6MNPS on a single G5 2Ghz and it's optimized for G5. >Do you really think PerOla would show up with a G5 for CCT6 if the G5 is a slow >CPU. > >regards >Andy ?? Now compare with Opteron with 400Mhz RAM. Blows this cpu away left and right! I hope you realize the weakest chain is memory latency. Both crafty and that ruffian are way too memory oriented. G5 trivially is good there, just like Opteron with its on die memory controller. In 2005 this G5 is still clocked at 2Ghz. Opteron moves by end 2004 to 0.09 and will hit way higher clockrates. So i have no idea where you live in history. It's faster now when using the right compiler, and it sure will be a lot faster soon again. http://www.spec.org/osg/cpu2000/results/res2004q1/cpu2000-20040126-02725.html What i'm seeing here for Power4 really is not impressive despite unified caches.
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.