On Wed, 6 Dec 2000, I wrote:
> I was just trying out FFT routines from Numerical Recipes.
> To my great surprise I made the observation that the old
> F77 routine 'realft' (for an array of 1024 real data points)
> is twice as fast as the newer F90 version of it. I compiled
> both with the same NAGWare f95 compiler and run it on a
> 233 MHz Pentium-I PC. I know that the F90 versions are
> written with regard to effective use on parallel processors,
> but didn't expect such a dramatic loss of performance on a
> single processor machine.
Thanks to all who replied to my question, in particular to
Malcom Cohen from NAG. It seems that it really is the code
of NR's f90 version of the FFT routines that is so much
slower than the f77 version. It apparently has nothing to
do with the compiler (which I didn't expect), since other
compilers show the same slow down.
Thanks again and best regards,
Roland
+----------------------------------------------------------------------+
| Roland Schilling Home-Office: +49(89)32929-670 |
| Max-Planck-Institut fuer Quantenoptik Phone: +49(89)32905-265 |
| Hans-Kopfermann-Str. 1 Fax: +49(89)32905-200 |
| D-85748 Garching E-mail: [log in to unmask] |
| Germany http://www.geo600.uni-hannover.de |
+----------------------------------------------------------------------+
|