[Numpy-discussion] Fast threading solution thoughts

David Cournapeau david@ar.media.kyoto-u.ac...
Thu Feb 12 08:04:04 CST 2009


Matthieu Brucher wrote:
>
> For BLAS level 3, the MKL is parallelized (so matrix multiplication is).
>   

Same for ATLAS: thread support is one focus in the 3.9 serie, currently
in development. I have never used it, I don't know how it compare to the
MKL,

David


More information about the Numpy-discussion mailing list