[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Working patch for FFTW 3.0.x and Nd FFT's
From: |
David Bateman |
Subject: |
Re: Working patch for FFTW 3.0.x and Nd FFT's |
Date: |
Wed, 18 Feb 2004 10:10:29 +0100 |
User-agent: |
Mutt/1.4.1i |
According to Daniel J Sebald <address@hidden> (on 02/17/04):
> It is common for the meaning of an Octave function to
> be dependent upon the type of variable passed into it, or not passed in.
>
> fft_wisdom('mywisdom.wis');
> would set up Octave to use a wisdom file other than the default, fine.
At the moment this call loads/saves the wisdom dependent on the existence
of the file 'mywisdom.wis', which as you stated, potentially allows the
user to fall into the trap of using an incorrect wisdom file.
> fft_wisdom();
> fft_wisdom() -- prints the contents of the wisdom file?
> would run wisdom and overwrite the default file. Maybe this is a bit dodgy.
I don't see the point of printing out the wisdom. Stuff like,
(fftw-3.0.1 fftw_wisdom
(fftw_codelet_t2_8 1 #xc050 #x76a287c1 #x6467f8c0 #xbc5b4077 #x6b73469d)
(fftw_dft_vrank_geq1_register 0 #xc050 #x27f03af7 #xbb4692da #x6548866d #xcee8
b671)
(fftw_codelet_n1_8 0 #xc050 #x2b3c6956 #xa6db414f #xd90f414e #x335c0c22)
(fftw_codelet_hc2r_8 1 #xc050 #x52f0d2c4 #xc3a97c59 #xf5c2f3ed #xa4aa2a33)
(fftw_codelet_q1_4 0 #xc050 #xc1cdcb4f #x1970fc77 #x4b55d74c #x3ac37e7b)
(fftw_codelet_q1_8 0 #xc050 #x228110fa #x9f6c00de #x83523c60 #xcfa30b3f)
(fftw_codelet_m1_32 0 #xc050 #x6f893bf5 #xd45f2cbd #xa050da62 #xf302784b)
)
doesn't really tell the user much. Maybe it could print out the FFTs for
which wisdom already exists, but except for writing a parser for the wisdom
files themselves without Octave, I don't quite see how to access this
information.
As for making 'fftw_wisdom()' call the canonical wisdom command to generate
the system wisdom, this is a really really bad idea. It took 12 hours to
do this on my machine, and in that time you wouldn't be able to interrupt
the process.
> save 'mywisdom.wis' fft_wisdom();
> I'm not sure about the saving portion of it.
This might be possible, if we created a class to hold the wisdom info and
then did "wisdom = fftw_wisdom(); save -ascii 'mywisdom.wis' wisdom",
where fftw_wisdom would be of the octave_wisdom class. This is a lot of
extra work for no gain in functionality, since we can easily have a
seperate function to load/save the FFTW wisdom. Also the wisdom would be
stored twice in this case; once in the octave class octave_wisdom, and
again within FFTW itself.
Cheers
David
--
David Bateman address@hidden
Motorola CRM +33 1 69 35 48 04 (Ph)
Parc Les Algorithmes, Commune de St Aubin +33 1 69 35 77 01 (Fax)
91193 Gif-Sur-Yvette FRANCE
The information contained in this communication has been classified as:
[x] General Business Information
[ ] Motorola Internal Use Only
[ ] Motorola Confidential Proprietary
- Re: Working patch for FFTW 3.0.x and Nd FFT's, (continued)
- Re: Working patch for FFTW 3.0.x and Nd FFT's, Paul Kienzle, 2004/02/18
- Re: Working patch for FFTW 3.0.x and Nd FFT's, John W. Eaton, 2004/02/18
- Re: Working patch for FFTW 3.0.x and Nd FFT's, Paul Kienzle, 2004/02/19
- Re: Working patch for FFTW 3.0.x and Nd FFT's, John W. Eaton, 2004/02/20
- Re: Working patch for FFTW 3.0.x and Nd FFT's, David Bateman, 2004/02/18
- Message not available
- Message not available
- Re: Working patch for FFTW 3.0.x and Nd FFT's, Daniel J Sebald, 2004/02/17
- Re: Working patch for FFTW 3.0.x and Nd FFT's,
David Bateman <=
- Re: Working patch for FFTW 3.0.x and Nd FFT's, Daniel J Sebald, 2004/02/18
- Re: Working patch for FFTW 3.0.x and Nd FFT's, David Bateman, 2004/02/18
- Re: Working patch for FFTW 3.0.x and Nd FFT's, Dmitri A. Sergatskov, 2004/02/18
- Re: Working patch for FFTW 3.0.x and Nd FFT's, John W. Eaton, 2004/02/18
- Re: Working patch for FFTW 3.0.x and Nd FFT's, Daniel J Sebald, 2004/02/18