[Beowulf] Single threaded memory bandwidth compared between Nehalem-EP and Westmere-EX

Joseph Landman landman at scalableinformatics.com
Sat Oct 19 05:10:51 PDT 2013


Did you try taskset to pin the CPU and the memory on the same numa
node?  What does numastat report?



Please pardon brevity and typos ... Sent from my iPhone

On Oct 19, 2013, at 3:36 AM, Chris Samuel <samuel at unimelb.edu.au> wrote:

> On Fri, 18 Oct 2013 12:32:00 PM Reuti wrote:
>
>> Hi,
>>
>> Am 18.10.2013 um 05:34 schrieb Christopher Samuel:
>>>
>>> We've got a group who bought themselves a 16 core 1.5TB Westmere-EX
>>> system and are getting what they consider to be poor memory bandwidth
>>> compared to an older 8 core Nehalem-EP system.
>>
>> Can you specify the CPUs in question?
>
> X5550 for Nehalem-EP
> E7-8837 for Westmere-EX
>
>>> Same CPU clock speed (2.6GHz), same memory speed (1067MHz), but about
>>> half the memory bandwidth (single threaded). I've run my own tests
>>> with the Stream benchmark and can confirm the same.
>>
>> Is the speed limited by the memory chips or the mainboard, or the population
>> on the mainboard (i.e. often the speed drops when all DIMM banks are used).
>
> I suspect not for mainboard, this is a IBM x3690 X5's and we've also got a
> quad socket SGI Altix UV10 with 1TB of RAM which shows pretty much the same
> bandwidth.
>
> cheers,
> Chris
> --
> Christopher Samuel        Senior Systems Administrator
> VLSCI - Victorian Life Sciences Computation Initiative
> Email: samuel at unimelb.edu.au Phone: +61 (0)3 903 55545
> http://www.vlsci.org.au/      http://twitter.com/vlsci
>
> _______________________________________________
> 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