Board index » delphi » 2d-FFT or the Need for MORE speed

2d-FFT or the Need for MORE speed

Dear Delphi-programmers,

Im working for sometime on a project that does complex 2d-FFT to show
the effects of optical-aberrations on the quality of astronomical
telescopes. The routines I use are based on Don Cross' algorithm
optimized for a 512x512 matrix. I was wondering if any of you could
point me towards a speedier routine/DLL than the one Im using at this
moment.

kind regards,

Cor Berrevoets
Ritthem, The Netherlands
Aberrator free star-testing software
http://home.wxs.nl/~erkma000/aberrator.htm

Sent via Deja.com http://www.deja.com/
Before you buy.

 

Re:2d-FFT or the Need for MORE speed


You might be able to find something on
http://www.efg2.com/Lab/Library/Delphi/MathFunctions/Engineering.htm#DSP...

<aylin....@wxs.nl> schreef in bericht news:8pcqqj$7s$1@nnrp1.deja.com...

Quote
> Dear Delphi-programmers,

> Im working for sometime on a project that does complex 2d-FFT to show
> the effects of optical-aberrations on the quality of astronomical
> telescopes. The routines I use are based on Don Cross' algorithm
> optimized for a 512x512 matrix. I was wondering if any of you could
> point me towards a speedier routine/DLL than the one Im using at this
> moment.

> kind regards,

> Cor Berrevoets
> Ritthem, The Netherlands
> Aberrator free star-testing software
> http://home.wxs.nl/~erkma000/aberrator.htm

> Sent via Deja.com http://www.deja.com/
> Before you buy.

Re:2d-FFT or the Need for MORE speed


Thankfs for the info but I allready found the EFG page. Im in need of a
more dedicated routine, possible even written in ASM.

In article <8phuha$m...@porthos.nl.uu.net>,
  "M.H. Avegaart" <avegaartNOS...@mccomm.nl> wrote:

Quote
> You might be able to find something on

http://www.efg2.com/Lab/Library/Delphi/MathFunctions/Engineering.htm#DSP
...

Quote

Cor Berrevoets
Ritthem, The Netherlands
Aberrator free star-testing software
http://home.wxs.nl/~erkma000/aberrator.htm

Sent via Deja.com http://www.deja.com/
Before you buy.

Re:2d-FFT or the Need for MORE speed


In article <8pl3qo$4h...@nnrp1.deja.com>,
   cor...@my-deja.com wrote:

Quote
>Thankfs for the info but I allready found the EFG page. Im
in need of a
>more dedicated routine, possible even written in ASM.

I doubt you'll find anything that beats the ISPL (Intel
Signal Processing Library, from Intel). It's heavily
optimized for all kinds of Intel CPUs and designed for
maximal performance.
No amateur work, to say the least.

REMOVE nospam. in my address

Please note our recent change of area code from
714 to 949 and the new name of our department

Quoc Thang NGUYEN
Laboratory of Cellular and
Molecular Neurobiology
Dept. of Neurobiology and Behavior
University of California, Irvine
Irvine, CA92717 USA
Ph: (949) 824-4730
Fx: (949) 824-3522

Re:2d-FFT or the Need for MORE speed


Dear Quoc Thang Nguyen,

Sorry for this question, but I could only get your answer concerning
this interesting question which 2D-FFT library should be used (on an
Intel computer). Is this ISPL (Intel
Signal Processing Library, from Intel) library free, at least for
university usage?

Thanks in advance,

Daniel Spangenberg

Physikalische Chemie und Elektrochemie I
Heinrich-Heine Universitaet Duesseldorf

Re:2d-FFT or the Need for MORE speed


I've torn through Intel's site and STILL can not locate the ISPL...any
ideas?

Jason

Re:2d-FFT or the Need for MORE speed


Finally found it...
http://support.intel.com/support/performancetools/libraries/spl/

Re:2d-FFT or the Need for MORE speed


In article <8q4y5.719$PW4.560...@typhoon.southeast.rr.com>,
   "Jason McClain" <jmccl...@nospam.cox.rr.com> wrote:
Quote
>Finally found it...
>http://support.intel.com/support/performancetools/libraries/spl/

Glad you did! There is a treasure trove on Intel's Web site
for developers dealing with numerical problems.

Good luck,

REMOVE nospam. in my address

Please note our recent change of area code from
714 to 949 and the new name of our department

Quoc Thang NGUYEN
Laboratory of Cellular and
Molecular Neurobiology
Dept. of Neurobiology and Behavior
University of California, Irvine
Irvine, CA92717 USA
Ph: (949) 824-4730
Fx: (949) 824-3522

Other Threads