[ quantlib-Patches-2315587 ] Fast Fourier Transform changes.

[SourceForge.net]

Patches item #2315587, was opened at 2008-11-19 20:32
Message generated for change (Comment added) made by lballabio
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=2315587&group_id=12740

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Slava Mazur (shlagbaum)
Assigned to: Luigi Ballabio (lballabio)
Summary: Fast Fourier Transform changes.

Initial Comment:
I propose the following changes to ql/math/fastfouriertransform.hpp and related files:

1. Enabled it (ifdef changed to ifndef)

2. Made number of frequencies a run time parameter, which
is passed as a parameter of class ctor. Very often the frequency of FT is determined at run time, so it's better
to have it this way.

3. Improved performance. Sines and cosines for a given frequency are calculated ones at object construction. This also allows for reuse the same object for different transformations.

4. Changed interface of fft method. Additional template parameter added to distinguish between the types of input and output sequences. End of output sequence added to allow for different input and output length.

5. Added check on consistency of size of the input sequence. Replaced [] by *-type dereference, which is more appropriate for iterators.

6. Added operator () to make the class a functor.

7. Minor changes to bit_reverse method.

8. Added a test to the test suite (separate patch).

The differences between proposed changes and the current trunk version are attached.

Regards,

Slava Mazur

----------------------------------------------------------------------

>Comment By: Luigi Ballabio (lballabio)
Date: 2009-01-29 17:19

Message:
Slava, I guess we'll have to wait until your company settles on an option.
Sorry, but there's not much I can do...


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-01-29 16:11

Message:
Luigi, i also would prefer the second option, but i still don't have a
clear permission from my company to do that. The problem here is pretty
standard - i can put my name in the copyright because this violates company
policies and they relactant to put their because they probably afraid of
possible implications and because they will have to review the current
policies. So, if you think of option 1 is a viable one - go for it, i don't
care much about authorship of this stuff. Let me know if you can recommend
something better. I'll let you know if there are any changes on my side.
Sorry, if it looked like i was ignoring your messages. I just didn't have
anything certain and don't have it now.

----------------------------------------------------------------------

Comment By: Luigi Ballabio (lballabio)
Date: 2009-01-28 12:04

Message:
A question before I can apply the patch: who owns the copyright of the code
you contributed? Is it you, your employer, your own company...? Posting the
copyright attribution I should use, i.e., something like
    Copyright (C) 2009 Random J. Hacker
or
    Copyright (C) 2009 ACME inc.
would be best.


----------------------------------------------------------------------

Comment By: Luigi Ballabio (lballabio)
Date: 2008-12-15 09:38

Message:
yes, we do care. We're just slow.


----------------------------------------------------------------------

Comment By: Slava Mazur (shlagbaum)
Date: 2008-12-12 16:29

Message:
Anybody cares about this?

----------------------------------------------------------------------

You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=2315587&group_id=12740

------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
QuantLib-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-dev