FFT from CERN's genlib 
Author Message
 FFT from CERN's genlib

I need to port a fortran program containing the following calls
            CALL FFT(X,Y,1,N,1)
            CALL FFT(X,Y,1,N,-1)
They are routines from a old version of CERN's genlib we've on our
DEC Station.

Now we're migrating to UNIX so I need to recompile the sources, but the new
version of CERN's libraries contains no more these subroutines.

From the program I deduce:
        1) It's real FFT
        2) X is the input vector
        3) Y is the output vector
        4) N is the number of elements
        5) 1/-1 means direct/inverse transform

I don't know what the first "1" means.

I tried to substitute it with subroutines from FFTPACK and from Numerical
Recepies. Results of FFTPACK are identic to those of Numerical Recepies, but
they are different from the results I get on VAX: the resulting analisys
diagram is completely useless.

The author is no more reacheable.

Has someone ever used this genlib's FFT? Has someone the CERN's man page
for that old subroutine, so that I can see how to substitute it?

Thank you very much
Kind regards
--Angelo Pagan



Sun, 12 Jan 1997 02:07:27 GMT  
 FFT from CERN's genlib


Quote:
>I need to port a FORTRAN program containing the following calls
>        CALL FFT(X,Y,1,N,1)
>        CALL FFT(X,Y,1,N,-1)
>From the program I deduce:
>    1) It's real FFT
>    2) X is the input vector
>    3) Y is the output vector
>    4) N is the number of elements
>    5) 1/-1 means direct/inverse transform

>I don't know what the first "1" means.

Might it be possible that the first "1" is the index of the first array element
to be considered, and N is the index of the last element to be considered?

...dav
--
Ph: (h) +61 2 319 4001          D A Vincent  Software Development Engineer
Ph: (w) +61 2 337 0434             Atlas Elektronik (Australia) Pty Ltd

      "I hate it all, but there must be more to hate than this!"  



Wed, 15 Jan 1997 07:17:23 GMT  
 FFT from CERN's genlib

Quote:


>>I need to port a FORTRAN program containing the following calls
>>            CALL FFT(X,Y,1,N,1)
>>            CALL FFT(X,Y,1,N,-1)
>>From the program I deduce:
>>        1) It's real FFT
>>        2) X is the input vector
>>        3) Y is the output vector
>>        4) N is the number of elements
>>        5) 1/-1 means direct/inverse transform

>>I don't know what the first "1" means.
>Might it be possible that the first "1" is the index of the first array element
>to be considered, and N is the index of the last element to be considered?

It might also distance between elements to use in the FFT.  This makes more
sense in the complex to complex case where the FFT is used for the second
pass of a multiple pass FFT.


Fri, 17 Jan 1997 12:01:32 GMT  
 
 [ 3 post ] 

 Relevant Pages 

1. Compiler (people Retired) API's ;FFT's;FHT's

2. FFT's in FORTH?

3. Alliance CAD 3.0, How to configure Genlib?

4. FFT asm code in Dr. Dobb's Journal

5. Fast FFT's for 80x87

6. need GENLIB-libraries

7. ? what's wrong with this n dim fft code

8. Numeric:FFT inverse_real_fft doesn't trevni!

9. n-d FFT's and f90

10. Compiler Sources (actually CERN and the army)

11. Regina Cern Extensions for RxSocket?

12. CERN-Server and VRML

 

 
Powered by phpBB® Forum Software