BLASX icon indicating copy to clipboard operation
BLASX copied to clipboard

Published 20 hours ago • verified publisher icon linnanwang

CPU Level Parallelism

Open romanlarionov opened this issue 9 years ago • 5 comments

Hello!

This library looks great, but I was wondering if it has CPU multi-threading blas capabilities. Reading through the code for some of the *gemm files, it almost appears to be the case.

I'm trying to perform a benchmark on AWS between g2 and c4 instances. I was hoping to find some way of writing a single code base that will perform the same function on the two different instances.

romanlarionov avatar Apr 21 '16 22:04 romanlarionov

For CPU multi-threading, it depends on what CPU BLAS you link, and how you configure them.

Please don't pay too much attention to CPU, this is a multiGPU BLAS. For more details, you can visit cuBLAS-XT. They are pretty same, but different approach.

linnanwang avatar Apr 22 '16 00:04 linnanwang

Alright, good to know and thanks for the lead.

Btw, are there any plans for supporting double complex based operations in the near future?

romanlarionov avatar Apr 22 '16 00:04 romanlarionov

good catch. I just merged ZGEMM a few seconds ago.

linnanwang avatar Apr 22 '16 00:04 linnanwang

Ohh I didn't know that was a valid pull request. Is the current master golden? I pretty much need to support zgemm and would rather use this over cublas-xt.

romanlarionov avatar Apr 22 '16 01:04 romanlarionov

Yes, the current master has ZGEMM.

linnanwang avatar Apr 22 '16 01:04 linnanwang