I was actually going to post something about this to gauge interest. fftw looks to have really good performance across all platforms. Do you have to build in support for the SIMD versions on each platform or is it automagical detection at runtime?
iirc, you'll are able to configure fftw's makefiles to use sse (single precision) and sse2 (double precision) ... since pd represents samples as floats, we only need to build the single precision library... gcc doesn't support optimized code for different architectures in one binary ... icc does, but results bigger binaries...
you won't have to use SIMD optimization for pd, to get the full performance out of the fft objects, unless fftw uses them. on the other hand i added some optimize flags to pd's configure script ... at least p4 will get the best optimization i know of... but even without all the SIMD optimizations fftw uses a faster algorithm to compute the fft, and most of the performance will come from that, so it's not that a big difference if you use SIMD or not...
cheers...
Tim mailto:TimBlechmann@gmx.de ICQ: 96771783 -- The only people for me are the mad ones, the ones who are mad to live, mad to talk, mad to be saved, desirous of everything at the same time, the ones who never yawn or say a commonplace thing, but burn, burn, burn, like fabulous yellow roman candles exploding like spiders across the stars and in the middle you see the blue centerlight pop and everybody goes "Awww!" Jack Kerouac