Skip to content


Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP


Heterogeneous Systems #15

zeland opened this Issue · 1 comment

2 participants


What about Heterogeneous Systems? For example I have two ATI 5870 and NVIDIA 470 in my system.
how can i use GATLAS?


Unless it has changed in the last few months, it is not possible to link a single binary to the ATI and NVIDIA OpenCL libraries at the same time. Scheduling across heterogeneous GPUs must be done from multiple processes rather than threads due to this linkage limitation from the vendor libraries. Depending on your application, this is not as terrible as it sounds. Map-reduce style problems are not difficult to partition and schedule arbitrarily if each scheduled computation is independent. This corresponds to the flat data parallel case. It does happen quite a lot.

@cjang cjang closed this
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.