[gmx-developers] BlueGene

Roland Schulz roland at utk.edu
Thu Jul 31 07:52:46 CEST 2008


Hi,

it was mostly wrong names in the #include statements.

It works on BlueGene/P and is much faster than the C non-bonded kernel
(about 2x). With this intrinsic C kernel one gets 18% of peak. Because
BlueGene doesn't have SIMD commands specific for single precision (thus it
does only 2 at a time instead of 4) the performance difference to Opteron is
greater than for double precision code. The performance difference to a
Barcelona 2.1Ghz core is 5.7x (at 20% peak).

Roland

On Wed, Jul 30, 2008 at 12:29 PM, Roland Schulz <roland at utk.edu> wrote:

> Hi,
>
> it seems that the blugene non-bonded kernel in the CVS version is
> incomplete. E.g. the nb_kernel_bluegene_test_asm.h is missing as far as I
> can see. The gromacs wiki for bluegene explains how to get the kernel
> generated by mknb to work. Also on the mailinglist all the messages I found
> about BlueGene were about mknb and not abou the bluegene specifik kernel.
>
> Could it be that it is incomplete do one has to do somethng to generate
> these files?
>
> Roland
>
>
> On Wed, Jul 30, 2008 at 11:45 AM, Berk Hess <hessb at mpip-mainz.mpg.de>wrote:
>
>> Hi,
>>
>> Mathias Puetz from IBM wrote non-bonded kernels for BlueGene/L.
>> I don't know if these are also used for BlueGene/P.
>> If there are problems, he would be the person to contact.
>>
>> Berk.
>>
>>
>> Roland Schulz wrote:
>>
>>> Hi,
>>>
>>> has someone looked at the non-bonded kernels for BlueGene/P? Based on the
>>> assembler produced by the compiler, it does no SIMD, runs out of registers
>>> and does unnecessary rounding.
>>>
>>> Are there plans for a PPC450 assembler kernel?
>>>
>>> Roland
>>> This email was Anti Virus checked by Astaro Security Gateway.
>>> http://www.astaro.com
>>> ------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> gmx-developers mailing list
>>> gmx-developers at gromacs.org
>>> http://www.gromacs.org/mailman/listinfo/gmx-developers
>>> Please don't post (un)subscribe requests to the list. Use the www
>>> interface or send it to gmx-developers-request at gromacs.org.
>>>
>>
>> _______________________________________________
>> gmx-developers mailing list
>> gmx-developers at gromacs.org
>> http://www.gromacs.org/mailman/listinfo/gmx-developers
>> Please don't post (un)subscribe requests to the list. Use the www
>> interface or send it to gmx-developers-request at gromacs.org.
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20080731/3fdd55f2/attachment.html>


More information about the gromacs.org_gmx-developers mailing list