cula with multiple GPUs?

General discussion for CULA. Use this forum for questions, examples, feedback, and feature requests.

cula with multiple GPUs?

Postby bravegag » Thu Jun 20, 2013 1:45 am

Hello,

If I have multiple GPU installed in the same host, how can I take advantage via CULA? Will CULA use all the available GPUs or I have to load balance myself? if I have to do the load balancing what's the CULA API function to specify the current Device Id?

e.g. in cuBLAS I can specify the device:
cudaSetDevice(deviceId)

TIA,
Best regards,
Giovanni
bravegag
 
Posts: 1
Joined: Thu Jun 20, 2013 1:30 am

Re: cula with multiple GPUs?

Postby john » Mon Jun 24, 2013 5:23 am

You have to load balance yourself if you're working on problems independently with each gpu. If you're using them cooperatively, then you could try out the pCULA interface.
john
Administrator
 
Posts: 587
Joined: Thu Jul 23, 2009 2:31 pm


Return to General CULA Discussion

Who is online

Users browsing this forum: No registered users and 1 guest

cron