OpenCL GEMM routines

[rahulg]

Hi. In some of the MAGMA presentations, and in one paper, autotuned OpenCL GEMM routines were mentioned. Are they distributed/available anywhere?

I actually have developed autotuned BLAS of my own, (to be open-sourced soon) and wanted to compare the results. I am also interested in porting clMAGMA to my BLAS instead of AMD's BLAS.

[mgates3]

We've worked on auto-tuned CUDA GEMM routines, which are available in the CUDA MAGMA 1.2 distribution. For OpenCL, we use AMD's BLAS.

-mark