All geek questions in one place

Any new ideas on using openCL with multiple GPUs?

My question is:

Was there a new promotion (or perhaps a developed tool / library) for using openCL with multiple GPUs? I understand that if someone wants to write code in openCL in order to use several GPUs, then he can, but they told me that the way to organize communication between them is a little "primitive". I want to know if there is something there that can put a level of abstraction between the programmer and everything that is a communication device between GPUs.

I am working on stochastic simulators with fairly large gratings, and I would like to be able to break them down into different GPUs, each of which can perform calculations and report if necessary. Writing this so that it is efficient enough, so if I can avoid all the low level of work using the standard way of doing it through openCL, it will be a lot of help.

Thanks!

+4
source share
1 answer

Scientifically, this article from Seoul National University in South Korea:

Reaching an image of a single computing device in OpenCL for multiple GPUs , http://dl.acm.org/citation.cfm?id=1941591

The authors propose an automatic mechanism for dividing the core into several GPUs. Unfortunately, their framework has not yet been released.

+1
source

Source: https://habr.com/ru/post/1398325/

More articles:


All Articles