Comments about FFT PATCH ready



On Sep 1, 2009, at 3:12 PM, Raymond Toy wrote:

> ?iga Lenar?i? wrote:
>>
>> On Aug 31, 2009, at 4:58 PM, Raymond Toy wrote:
>>>>
>>>> I'm concerned about the speed of fft on GCL - particularly, does  
>>>> GCL
>>>> (on windows) compile the code that is loaded with 'load(fft);' in
>>>> Maxima? Otherwise all 'optimizations' are rather pointless for  
>>>> the 90%
>>>> of userbase that runs GCL on Windows.
>>> load does not normally compile anything.  So it will be rather slow.
>>>
>>> Ray
>>>
>>
>> Perhaps fft should compile automatically on first load like lapack?
>> It's seems reasonable since it's numerical code. Users shouldn't be
>> troubled with 'manual' compilation.
>>
> Yes, that's probably a good idea.  I certainly expect fft to be
> reasonably fast.
>
> Ray
>

I can try to 'duplicate' the system that's used in lapack's case,  
though I don't understand it completely.

Another thing - I think fft should use 'flonum' type instead of  
'double-float', right (to support compilation with double-doubles and  
so on)?

Regards,
Ziga