[SciPy-user] FFT speed ?

Nils Wagner nwagner@iam.uni-stuttgart...
Wed Feb 7 08:13:05 CST 2007


Samuel GARCIA wrote:
> Yes it is pretty fast.
> The difference of time is not a problem for the end user (between 0.2
> and 1 s).
> But specialy the value of 186889 is really a problem, the computation
> is infinit.
> Did you try tis value especialy ?
>
You may try

a = rand((186889))
t1=time.time()
fft(a,pow(2,18))
t2=time.time()
print t2-t1
 
Nils



More information about the SciPy-user mailing list