[SciPy-Dev] Accuracy of single-precision FFT

Pauli Virtanen pav@iki...
Fri Jun 25 09:05:08 CDT 2010


Fri, 25 Jun 2010 14:07:56 +0200, Sturla Molden wrote:
> Den 24.06.2010 19:41, skrev Anne Archibald:
>> We can't use FFTW, unfortunately, but it's clear that prime-size FFTs
>> can do much better than we are.
>
> FFTW and FFTPACK are not the only options.
>
> We could use Intel MKL, not just for FFTs but also BLAS. It's not worse
> than asking Intel to sponsor NumPy/SciPy with an MKL license. That would
> allow redistribution of a binary build linked against MKL, similar to
> what Cristoph Gohlke and Enthought are doing:

That would require adding support for multi-backend FFT back to Scipy. If 
we're going to put it back, then it would be worth to think carefully 
about the design first, so that someone doesn't have to rip it out 
again :)

Supporting only MKL is not an option, since even if we managed to secure 
ourselves a redistribution license, it will disturb our redistribution 
chain because all third party packagers [eg. Python(x,y), Linux 
distributions] would also need a license of their own. Moreover, 
developers would also need to grab a copy of MKL etc...

-- 
Pauli Virtanen



More information about the SciPy-Dev mailing list