[SciPy-user] Confusing BLAS/LAPACK situation on Mac OS X

Arnd Baecker arnd.baecker at web.de
Sun Jan 29 04:18:22 CST 2006


On Sat, 28 Jan 2006, Robert Kern wrote:

> Arnd Baecker wrote:
>
> > It might also be a bad choice, as ATLAS can be slower than
> > other optimized LAPACK routines, e.g. MKL or ACML.
> > Do any benchmarks of `Accelerate` vs. ATLAS exist?
>
> The BLAS in Accelerate.framework *is* ATLAS.

Thanks for the clarification!
((I did a quick google on Accelerate.framework before which lead me to
http://www.cocoabuilder.com/archive/message/cocoa/2004/10/17/119581
"""
If you're really paranoid about speed then you might want to look at
ATLAS <http://math-atlas.sourceforge.net/>. I'm not sure what the state
of play is nowadays, but there was a point when it produced a faster
BLAS implementation than the one in Apple's Accelerate framework.
"""
))

Best, Arnd



More information about the SciPy-user mailing list