Combine Multiple AWS Instances into a 16-GPU Monster Machine

Re:Combine Multiple AWS Instances into a 16-GPU Monster Machine by mtanski on Thursday March 31, 2016 @09:02PM (#11401837)

Congrats guys this is a really neat hack, really impressive.

Did you guys think about building it further out to provide a GPU load balancer for multiple frontend machines running Cuda / OpenCL?

[ Parent | on HN ]

Re:Combine Multiple AWS Instances into a 16-GPU Monster Machine by mtweak on Thursday March 31, 2016 @09:43PM (#11401990)

Hmm, can you elaborate? Do you mean having multiple smaller instances talk to a single GPU instance?

[ Parent | on HN ]

Re:Combine Multiple AWS Instances into a 16-GPU Monster Machine by mtanski on Friday April 01, 2016 @12:01AM (#11402490)

I'm talking about time-sharing. It doesn't mater if it's smaller instances sharing a single GPU instance or many instances sharing many GPU instances. Essentially N:M sharing (with some scheduling).

Since the GPU client is now abstracted from the GPU devices by placing the GPUs across the network. It seams like time-sharing should be next logical step.

[ Parent | on HN ]

Re:Combine Multiple AWS Instances into a 16-GPU Monster Machine by mtweak on Friday April 01, 2016 @12:50AM (#11402722)

Got it, this is actually already supported. At the very end of the blog post there is a link to create a custom configuration. You can create any N:M configuration, that is any number of clients to servers and therefore the level of performance scaling or GPU pooling.

Check it out: https://console.aws.amazon.com/cloudformation/home?region=us...

[ Parent | on HN ]

Re:Combine Multiple AWS Instances into a 16-GPU Monster Machine by mtanski on Friday April 01, 2016 @11:07AM (#11405316)

Great, thanks for clearing that up.

[ Parent | on HN ]