Sounds interesting. BTW I don't need FFT, but I think such a package
would be a good addition to Phobos. Would be great if we could
generalize the code a bit to work with general ranges (random-access I
suppose, worth exploring to see if less powerful ranges would be enough).
Andrei
On 08/01/2010 03:32 PM, David Simcha wrote:
Speed-wise, I've just been goofing around for the past hour or so and
I've sped it up 2x. It now does 1 size 2 ^^ 20 double[] ->
Complex!double FFT in about 880 milliseconds.
On 8/1/2010 3:36 PM, Walter Bright wrote:
David Simcha wrote:
As I am no longer going to use FFTs in my kernel density lib,
improving this FFT code will be bumped down my hacking to-do list.
Does what I have now sound better than nothing by a large enough
margin to warrant inclusion in std.numeric, or does it sound too
primitive to be widely useful? If it sounds worth including I'll
clean up/document the code and send it to the mailing list for
review. If it sounds too primitive, I'll just scrap it.
I don't know enough about FFT's to make any sort of informed comment.
_______________________________________________
phobos mailing list
[email protected]
http://lists.puremagic.com/mailman/listinfo/phobos
_______________________________________________
phobos mailing list
[email protected]
http://lists.puremagic.com/mailman/listinfo/phobos
_______________________________________________
phobos mailing list
[email protected]
http://lists.puremagic.com/mailman/listinfo/phobos
