[OMPI users] revisiting topic: prebuilt on windows7 not xp

[MM]

in http://www.open-mpi.org/community/lists/users/2011/07/16854.php, Shiqing
says it is impossible to use the prebuilt binaries with Windows xp, however
I manage to use them in Release mode  with vc2010, but not in debug mode.



Can this be some indication that there is a tweak to have both the release
and debug version of the libs when built on windows 7 work also for xp?



I would definitely rather use prebuilt openmpi libs as it's easier to change
versions,



thanks,

MM

Re: [OMPI users] gpudirect p2p?

[Chris Cooper]

Ah I see.  Thanks for the info!

On Sat, Oct 15, 2011 at 12:06 AM, Rolf vandeVaart
 wrote:
>>-Original Message-
>>From: users-boun...@open-mpi.org [mailto:users-boun...@open-mpi.org]
>>On Behalf Of Chris Cooper
>>Sent: Friday, October 14, 2011 1:28 AM
>>To: us...@open-mpi.org
>>Subject: [OMPI users] gpudirect p2p?
>>
>>Hi,
>>
>>Are the recent peer to peer capabilities of cuda leveraged by Open MPI when
>>eg you're running a rank per gpu on the one workstation?
>
> Currently, no.  I am actively working on adding that capability.
>
>>
>>It seems in my testing that I only get in the order of about 1GB/s as per
>>http://www.open-mpi.org/community/lists/users/2011/03/15823.php,
>>whereas nvidia's simpleP2P test indicates ~6 GB/s.
>>
>>Also, I ran into a problem just trying to test.  It seems you have to do
>>cudaSetDevice/cuCtxCreate with the appropriate gpu id which I was wanting
>>to derive from the rank.  You don't however know the rank until after
>>MPI_Init() and you need to initialize cuda before.  Not sure if there's a
>>standard way to do it?  I have a workaround atm.
>>
>
> The recommended way is to put the GPU in exclusive mode first.
>
> #nvidia-smi -c 1
>
> Then, have this kind of snippet at the beginning of the program. (this is 
> driver
> API, probably should use runtime API)
>
> res = cuInit(0);
> if (CUDA_SUCCESS != res) {
>    exit(1);
> }
>
> if(CUDA_SUCCESS != cuDeviceGetCount(&cuDevCount)) {
>    exit(2);
> }
> for (device = 0; device < cuDevCount; device++) {
>    if (CUDA_SUCCESS != (res = cuDeviceGet(&cuDev, device))) {
>        exit(3);
>    }
>    if (CUDA_SUCCESS != cuCtxCreate(&ctx, 0, cuDev)) {
>     /* Another process must have grabbed it.  Go to the next one. */
>    } else {
>        break;
>    }
>    i++;
> }
>
>
>
>>Thanks,
>>Chris
>>___
>>users mailing list
>>us...@open-mpi.org
>>http://www.open-mpi.org/mailman/listinfo.cgi/users
> ---
> This email message is for the sole use of the intended recipient(s) and may 
> contain
> confidential information.  Any unauthorized review, use, disclosure or 
> distribution
> is prohibited.  If you are not the intended recipient, please contact the 
> sender by
> reply email and destroy all copies of the original message.
> ---
>
> ___
> users mailing list
> us...@open-mpi.org
> http://www.open-mpi.org/mailman/listinfo.cgi/users
>