RE: 2.9.9+ for Windows

[Theresa Meuse]

Dear Michael Goffioul,

 

This is the same as my earlier email, but without the attached scripts (hence it is no longer 6MB).  Please note that I do not use cygwin (would I have better luck if I did?) and has MS Windows XP Professional 2002 SP 2 running on an Intel(R) Pentium® 4 CPU 3.60GHz 3.59GHz, 1.00 GB of RAM.

 

=======================================================================

 

I have been asked to make SSCA #3 v1.02, HPCS’s latest Synthetic Aperture Radar Benchmark, compatible with Octave under Windows, using this build:

 

GNU Octave 2.9.9+ version from Michael Goffioul http://www.dbateman.org/octave/octave_20070126(2.9.9+)_setup.exe

 

However it retrieves data erroneously from storage:

 

OCTAVE 2.9.9+

 

fwriteSDG.m – COMPUTE_MODE data write:

***DB: ftfMRow = 754, ftfNCol = 1.

***DB: kSize = 754, kuSize = 762.

***DB: uSize = 762, ucSize = 320.

 

freadStage1Start.m – COMPUTE_MODE data read:

***DB: ftfMRow = 754, ftfNCol = 1.

***DB: kSize = 754, kuSize = 762.

***DB: uSize = 762, ucSize = 320.

***DB: fastTimeFilterR = 82 x 1.

***DB: fastTimeFilterI = 0 x 1.

***DB: k = 0 x 1.

***DB: ku =  0 x 1.

***DB: u =  0 x 1.

***DB: uc =  0 x 1.

 

Whereas with MATLAB (also Linux Octave 2.9.5):

 

fwriteSDG.m – COMPUTE_MODE data write:

***DB: ftfMRow = 754, ftfNCol = 1.

***DB: kSize = 754, kuSize = 762.

***DB: uSize = 762, ucSize = 320.

 

freadStage1Start.m – COMPUTE_MODE data read:

***DB: ftfMRow = 754, ftfNCol = 1.

***DB: kSize = 754, kuSize = 762.

***DB: uSize = 762, ucSize = 320.

***DB: fastTimeFilterR = 754 x 1.

***DB: fastTimeFilterI = 754 x 1.

***DB: k = 1 x 754.

***DB: ku = 1 x 762.

***DB: u = 1 x 762.

***DB: uc = 1 x 320.

***DB: fastTimeFilter = 754 x 1.

 

The script snippets that produce these results are:

 

fwriteSDG.m – COMPUTE_MODE data write:

 

    % Store complex components, real part first, then imaginary.

    fwrite(fid, real(fastTimeFilter), 'float32');

    fwrite(fid, imag(fastTimeFilter), 'float32');

 

    fwrite(fid, k,  'float32');

    fwrite(fid, ku, 'float32');

 

    fwrite(fid, u,  'float32');

    fwrite(fid, uc, 'float32');

 

 

freadStage1Start.m – COMPUTE_MODE data read:

 

    % Retrieve complex raw SAR components real part first, then

    % imaginary. (Note that Matlab 7 SP2's FREAD promotes to double.)

    fastTimeFilterR = single(fread(fid, [ftfMRow,ftfNCol], 'float32'));

    fastTimeFilterI = single(fread(fid, [ftfMRow,ftfNCol], 'float32'));

 

    k  = single(fread(fid, [1,kSize],  'float32'));

    ku = single(fread(fid, [1,kuSize], 'float32'));

  

    u  = single(fread(fid, [1,uSize],  'float32'));

    uc = single(fread(fid, [1,ucSize], 'float32'));

 

I am wondering if you can recommend me another Octave for Windows build that I could try testing with.

 

Kind thanks,

 

Theresa

 

 

 

Theresa Meuse

MIT Lincoln Laboratory

244 Wood Street

Lexington, MA 02420

(781) 981-4066

www.ll.mit.edu