Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Correction hydra hardware

Author: Matthew Hull

Date: 21:03:26 02/01/05

Go up one level in this thread



>>>
>>>Quote on:
>>>------------------------------------------
>>>
>>>It's 4 nodes.
>>>
>>>1 node = 8 processor Xeon.
>>>------------------------------------------
>>>quote off
>>>
>>>"one node = 8 processor xeon".  That is _clearly_ SMP.  And _that_ is the box I
>>>was talking about.  I've run on more than one of them.  For chess they are bad,
>>>running 2 copies of crafty produces 2x NPS of one copy.  Running 4 copies
>>>produces nearly 4x the NPS of one copy.  Running 8 copes produces barely 6x the
>>>1 copy speed.
>>
>>
>>Would this be considered a NUMA relationship, between the two 4-way nodes.  Does
>>your NUMA crafty handle this better than your older SMP-only crafty?
>>
>
>
>wasn't numa.  Used the Intel "Fusion" chipset to tie two 4-way processor groups
>together to provide a total of 8, but there was a bottleneck.  It was one of
>those "let's see what it will do" sort of things.  Big L2 cache processors could
>provide good performance if memory requirements were limited.
>I'll take a NUMA opteron system any day.


I see now.

What then was the issue with the dual Athlons?  IIRC, the memory was not
interleaved, so SMP crafty did not do as well on those as on dual Xeon.  Did
your NUMA changes help with dual Athlons?

Thanks,



>
>
>



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.