[Beowulf] evaluating FLOPS capacity of our cluster

Gus Correa gus at ldeo.columbia.edu
Mon May 11 11:26:31 PDT 2009


Hi Tom, Rahul, list

Of course Tom is right about Barcelona and Shanghai being the first to 
have 4 flops/cycle max.
(Is this what AMD calls "3rd generation Opterons"?)
On my first email I should have mentioned that I was talking about
Opteron "Shanghai" 2376.
I suppose 4 flops/cycle max is when using SSE/SSE2/SSE3..., right?

Please, correct me if I am wrong,
but I found out that this information
(max number of floating point operations per cycle)
is not really very visible on the AMD web site for
their various processors.
I had to dig to find it.
I haven't checked the Intel site for their processors, though.

Is there any web site that lists this information clearly and correctly
for the most common marketplace processors (current, recent, old ones)?

If not, it would be a good service to the community if somebody
could put together a list like this.

Otherwise, how can I tell my fancy
Dell Precision 410 Pentium-III cluster Rpeak Gflops, Rahul?  :)

Thank you,
Gus Correa
---------------------------------------------------------------------
Gustavo Correa
Lamont-Doherty Earth Observatory - Columbia University
Palisades, NY, 10964-8000 - USA
---------------------------------------------------------------------



Tom Elken wrote:
>> On Behalf Of Rahul Nabar
>  
>> On Mon, May 11, 2009 at 12:23 PM, Gus Correa <gus at ldeo.columbia.edu>
>> wrote:
>>> Theoretical maximum Gflops (Rpeak in Top500 parlance), for instance,
>>> on cluster with AMD quad-core 2.3GHz processor is:
>>>
>>> 2.3 GHz x
>>> 4 floating point operations/cycle x
>>> 4 cores/CPU socket x
>>> number of CPU sockets per node x
>>> number of nodes.
>> Excellent. Thanks Gus. That sort of estimate is exactly what I needed.
>> I do have AMD Athelons.
>                  Athlons
> 
> AMD quad-core Opteron processors (code-named Barcelona and Shanghai for servers) were the first to have 4 FLOPs/cycle.  Earlier Opterons and Athlon64's had 2 FLOPS/cycle.
> 
> I don't know their current desktop processors as well, but I would guess that 3- or 4-core Phenoms also have a peak of 4 flops/cycle.  I am not sure about current Athlons.  But you may have to do more searching or give the list your precise Athlon (or Athlon64) model name and number to find out the actual # of FLOPS/cycle for your cluster.  IIRC, Athlon before Athlon64 had 1 FLOP/cycle -- though there have been a lot of branding and re-branding for AMD/Intel desktop CPUs over the years.
> 
> -Tom
> 
>> In fact, this is super usefule for some of our oldest legacy hardware
>> too. We used to just use Dell Desktops clustered together. I have
>> easily accessible all the other info. that goes into your equation.
>> Except the floating point operations / cycle numbers.
>>
>> Let me dig those out.
>>
>> Thanks!
>>
>> --
>> Rahul
>> _______________________________________________
>> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin
>> Computing
>> To change your subscription (digest mode or unsubscribe) visit
>> http://www.beowulf.org/mailman/listinfo/beowulf




More information about the Beowulf mailing list