[Numpy-discussion] experiments with SSE vectorization

Nathaniel Smith njs at pobox.com
Fri May 17 00:56:41 EDT 2013


On 17 May 2013 05:19, "Christopher Jordan-Squire" <cjordan1 at uw.edu> wrote:
>
> I'd been under the impression that the easiest way to get SSE support
> was to have numpy use an optimized blas/lapack. Is that not the case?

Apples and oranges. That's the easiest (only) way to get SSE support for
operations that go through blas/lapack, but there are also lots of
operations in numpy that are implemented directly.

-n
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20130517/2f4d9468/attachment.html>


More information about the NumPy-Discussion mailing list