This file is indexed.

/usr/share/psychtoolbox-3/PsychOpenGL/PsychGLSLShaders/gpgpuGVF.m is in psychtoolbox-3-common 3.0.11.20140816.dfsg1-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
function gpgpuGVF(singlefbo, trials)

if nargin < 1
    singlefbo=1;
end;

if nargin < 2
    trials = 1;
end;

% Is the script running in OpenGL Psychtoolbox?
AssertOpenGL;

% Find the screen to use for display:
screenid=max(Screen('Screens'));

% Setup Psychtoolbox for OpenGL 3D rendering support and initialize the
% mogl OpenGL for Matlab wrapper:
InitializeMatlabOpenGL(1);

% Open a double-buffered full-screen window on the main displays screen.
[win , winRect] = Screen('OpenWindow', screenid);

% Setup the OpenGL rendering context of the onscreen window for use by
% OpenGL wrapper. After this command, all following OpenGL commands will
% draw into the onscreen window 'win':
Screen('BeginOpenGL', win);

% Read test input image...
%impath = [PsychtoolboxRoot '/PsychDemos/konijntjes1024x768.jpg']
impath = [PsychtoolboxRoot '/PsychDemos/OpenGL4MatlabDemos/earth_512by256.jpg']
inputimage = imread(impath);
% ...convert into grayscale image...
inputimage = transpose(rgb2gray(inputimage));
%inputimage = inputimage(1:128, 1:128);
% Pad it to be a square, power-of-two image:
msize = pow2(ceil(log2(max(size(inputimage)))))
tmpimage = zeros(msize, msize);
xpad = int16(floor((msize - size(inputimage,1))/2))+1;
ypad = int16(floor((msize - size(inputimage,2))/2))+1;
tmpimage(xpad:xpad+size(inputimage,1)-1, ypad:ypad+size(inputimage,2)-1) = inputimage(:,:);
inputimage = tmpimage;

% Compute edge-map:
inputimage = uint8(edge(inputimage, 'canny'));

% TEST TEST TEST
%inputimage(:,:)=0;
%inputimage(100:105,:)=255;

Screen('EndOpenGL', win);


% ...and convert it to texture:
size(inputimage)
inputtex = Screen('MakeTexture', win, inputimage);

% Retrieve an OpenGL handle to it:
texin = Screen('GetOpenGLTexture', win, inputtex);

Screen('BeginOpenGL', win);



maxattachnr = glGetIntegerv(GL.MAX_COLOR_ATTACHMENTS_EXT)

% Use 
if glGetIntegerv(GL.MAX_COLOR_ATTACHMENTS_EXT)<2 || singlefbo==0
    singlefbo=0;
else
    singlefbo=1;
end;

precision = GL.RGBA_FLOAT16_APPLE;

% Create framebuffer objects and color buffer textures:
if singlefbo==0
    % Hw only supports one color attachment per FBO. Need to create three
    % FBOs for pingpong:
    [fbos(1), tex(1)]=moglCreateFBO(msize, msize, 1, 4, precision);
    [fbos(2), tex(2)]=moglCreateFBO(msize, msize, 1, 4, precision);
    [fbos(3), tex(3)]=moglCreateFBO(msize, msize, 1, 4, precision);
else
    % We use one FBO with two color attachments for pingpong.
    [fbos(1), tex ]=moglCreateFBO(msize, msize, 2, 4, precision);
    % Set fbos(2) = 0 to signal not to use fbo(2).
    fbos(2)=0;
    % We create one FBO for the static b, c1 and c2 arrays:
    [fbos(3), tex(3) ]=moglCreateFBO(msize, msize, 1, 4, precision);
end;

% Load our bias and rescale shader:
glslnormalizer = LoadGLSLProgramFromFiles('ScaleAndBiasShader');
prebias = glGetUniformLocation(glslnormalizer, 'prescaleoffset');
postbias = glGetUniformLocation(glslnormalizer, 'postscaleoffset');
scalefactor = glGetUniformLocation(glslnormalizer, 'scalefactor');

% Activate it for following blit op:
glUseProgram(glslnormalizer);

% Set no bias to be applied:
glUniform1f(prebias, 0.0);
glUniform1f(postbias, 0.0);

% Multiply all luminance values by 255, so they are in usual range 0-255
% instead of 0-1. We do this to reduce numeric roundoff errors.
glUniform1f(scalefactor, 255.0);

% Bind FBO 1, colorbuffer 1:
moglChooseFBO(fbos(1), 1);

% Blit input image into it, scaling it from 0-1 to 0-255:
moglBlitTexture(texin);

% Disable shader:
glUseProgram(0);

% Read it back (only the RED channel, it is a grayscale image, so
% red, green and blue contain the same content. We read it back to make
% sure we can compare Matlabs results to the GPU results:
inputimage = glReadPixels(0, 0, msize, msize, GL_RED, GL_FLOAT);
imshow(inputimage);
figure;

% Unbind FBO, reset to normal framebuffer:
moglChooseFBO(0);

% Perform reduce operation to compute image properties: We need the maximum
% and minimum luminance value:
[minv, maxv, meanv] = moglComputeMinMaxMeanOfTexture(moglGetTexForFBO(fbos(1), 1), fbos(1), fbos(2), 1);
glFinish;
tic;
for c=1:trials
    % We use the texture of fbo(1) as input image, as we know that fbo-1
    % will not be used as drawing target in first iteration. We provide
    % fbo(1) and fbo(2) as ping-pong buffers for iterative reduce
    % operation:
    [dminv, dmaxv, dmeanv] = moglComputeMinMaxMeanOfTexture(moglGetTexForFBO(fbos(1), 1), fbos(1), fbos(2), 1);
end;
durationgpu = toc / trials;

minc=min(min(inputimage));
maxc=max(max(inputimage));
meanc=mean(mean(inputimage));

tic;
minc=min(min(inputimage));
maxc=max(max(inputimage));
meanc=mean(mean(inputimage));
durationcpu = toc;

fprintf('Minimum   CPU = %f  , OpenGL = %f\n', minc, minv);
fprintf('Maximum   CPU = %f  , OpenGL = %f\n', maxc, maxv);
fprintf('Mean      CPU = %f  , OpenGL = %f\n', meanc, meanv);
fprintf('Time/pass CPU = %f ms, OpenGL = %f ms\n', durationcpu * 1000, durationgpu * 1000);


% Compute GVF in C-Code:
mu = 0.1
numiters = 100

% GVFC needs double input, not float input!
gvfcinput = double(inputimage);

tic
[gvfc_v,gvfc_u] = GVFC(gvfcinput, mu, numiters);
gvfcduration = toc * 1000
gvfcdurationperiter = gvfcduration / numiters

% This performs stage 1 of GVF initialization: Normalize all edge map
% values, remapping it so they span the whole range of 
% [min ; max] to interval [0 ; 1]. Compute initial edge map gradient by
% taking central differences as derivatives.
% We use a special shader for this that combines these two operations,
% drawing the input edge map texture into FBO 1 again:
tic;

% Load our remap, normalize and gradient map shader:
glslgradientshader = LoadGLSLProgramFromFiles('ScaleBiasAndGradientShader');
prebias = glGetUniformLocation(glslgradientshader, 'prescaleoffset');
postbias = glGetUniformLocation(glslgradientshader, 'postscaleoffset');
scalefactor = glGetUniformLocation(glslgradientshader, 'scalefactor');


% Draw into FBO 1:
moglChooseFBO(fbos(1), 1);

% Activate gradient shader for following blit op:
glUseProgram(glslgradientshader);

% Compute bias and scale to be used:
minv = minv / 255;
maxv = maxv / 255;
sf = 1.0 / (maxv - minv);

% Set bias to be applied: This maps the minimum value to zero:
glUniform1f(prebias, -minv);
% No post-scale bias:
glUniform1f(postbias, 0.0);
% Scale factor to spread out everything into range 0.0 - 1.0:
glUniform1f(scalefactor, sf);

% Blit input image into it, remapping/normalizing it to range 0.0 - 1.0 and
% then computing the gradient image on it. The RED channel will contain gx,
% the GREEN channel will contain gy, BLUE and ALPHA are unused.
moglBlitTexture(texin);

% Disable shader:
glUseProgram(0);

% Read blitted textures back from FBOs:
% This shows the normalized 0-1 edge-map:
img = glReadPixels(0, 0, msize, msize, GL_RGB, GL_FLOAT);
% Red is dx:
imshow(img(:,:,1), [-0.5 0.5]);
figure

% Green is dy:
imshow(img(:,:,2), [-0.5 0.5]);
figure

% Blue is remapped intensity:
imshow(img(:,:,3), [0 1]);
figure

% GVF Step 2: Initial flow field (u,v) is stored in (R,G) channels of
% FBO1-1's texture. Compute constant magnitude, c1 and c2 arrays in FBO 3:

% Bind FBO 3-1 for drawing. We could do fast-binding here, as dimensions
% match previously bound FBO 1-1:
moglChooseFBO(fbos(3),1);

% Load and activate our b=1-magnitude, c1, c2 shader:
glslGVFInitshader = LoadGLSLProgramFromFiles('GVFInitShader');
glUseProgram(glslGVFInitshader);

% Blit the flow field in FBO 1-1 into it:
moglBlitTexture(moglGetTexForFBO(fbos(1), 1));

% Disable shader:
glUseProgram(0);

% Read blitted textures back from FBOs:
% This shows the b=1-mag, c1 and c2 fields:
img = glReadPixels(0, 0, msize, msize, GL_RGB, GL_FLOAT);
% Red is b = 1 - mag (Range 0.5 to 1.0)
imagesc(img(:,:,1));
figure

% Green is c1:
imagesc(img(:,:,2));
figure

% Blue is c2:
imagesc(img(:,:,3));
figure

% FBO 1-1 contains initial (u,v) flow field.
% FBO 3-1 contains (b,c1,c2) static field.
% GVF iteration will bounce data between FBO 1-1 and FBO 1-2, the
% intermediate result of last iteration bound as texture to texture unit 0,
% the static field bound as texture to texture unit 1:

% Load GVF Update shader:
glslGVFUpdateshader = LoadGLSLProgramFromFiles('GVFUpdateShader');
ingradient = glGetUniformLocation(glslGVFUpdateshader, 'ingradient');
bc1c2field = glGetUniformLocation(glslGVFUpdateshader, 'bc1c2field');

% Activate update shader:
glUseProgram(glslGVFUpdateshader);

glUniform1i(ingradient, 0);
glUniform1i(bc1c2field, 1);

% Retrieve handle to mu constant and set it to reasonable value:
fourmu = glGetUniformLocation(glslGVFUpdateshader, 'fourmu');
glUniform1f(fourmu, 4 * mu);

% Bind FBO 1-2 as target for first iteration:
% We could do fast-binding as dimensions of all buffers match!
moglChooseFBO(fbos(1), 2);

% Source textures for iterations are cached in buffertex:
buffertex(1) = moglGetTexForFBO(fbos(1), 1);
buffertex(2) = moglGetTexForFBO(fbos(1), 2);

% Need to setup 2nd texture unit for constant input image:
b_c1_c2_tex = moglGetTexForFBO(fbos(3), 1);

% Unbind our constant texture:
glBindTexture(GL.TEXTURE_RECTANGLE_EXT, 0);

% Select the 2nd texture unit (unit 1) for setup:
glActiveTexture(GL.TEXTURE1);

% Switch it into RECTANGLE texture mapping mode:
glDisable(GL.TEXTURE_2D);
glEnable(GL.TEXTURE_RECTANGLE_EXT);

% Bind our constant texture:
glBindTexture(GL.TEXTURE_RECTANGLE_EXT, b_c1_c2_tex);

% Make sure we use nearest neighbour sampling:
glTexParameteri(GL.TEXTURE_RECTANGLE_EXT, GL.TEXTURE_MIN_FILTER, GL.NEAREST);
glTexParameteri(GL.TEXTURE_RECTANGLE_EXT, GL.TEXTURE_MAG_FILTER, GL.NEAREST);

% And that we clamp to edge:
glTexParameteri(GL.TEXTURE_RECTANGLE_EXT, GL.TEXTURE_WRAP_S, GL.CLAMP);
glTexParameteri(GL.TEXTURE_RECTANGLE_EXT, GL.TEXTURE_WRAP_T, GL.CLAMP);

% Choose texture application function to be a neutral REPLACE:
glTexEnvfv(GL.TEXTURE_ENV,GL.TEXTURE_ENV_MODE,GL.REPLACE);

% Unit 1 is ready, switch back to Unit 0, the first unit:
glActiveTexture(GL.TEXTURE0);

% Switch it into RECTANGLE texture mapping mode:
glDisable(GL.TEXTURE_2D);
glEnable(GL.TEXTURE_RECTANGLE_EXT);

inputtex = buffertex(1);

glFinish;

gvfinitduration = toc * 1000

tic;

% GVF iterative ping-pong update loop:
bufferid = 0;
for i=1:numiters
%tic
    % Compute drawbuffers id and its texturehandle:
    bufferid = 1 - bufferid;
    futuretex = buffertex(bufferid + 1);

    % Bind proper FBO as target for intermediate reduce results:
    if singlefbo == 0
        % Use two FBO's for pingpong:
        if bufferid==0
            glBindFramebufferEXT(GL.FRAMEBUFFER_EXT, fbos(1));
        else
            glBindFramebufferEXT(GL.FRAMEBUFFER_EXT, fbos(2));
        end;
    else
        % Use one dual-buffer FBO for pingpong:
        % We only switch the draw-buffer to keep the overhead as low as
        % possible:
        glDrawBuffer(GL.COLOR_ATTACHMENT0_EXT + bufferid);
    end;
    
    % Do it: We use fast-blit mode (1) because we did texture setup
    % ourselves:
    moglBlitTexture(inputtex, 0, 0, msize, msize, 1);

    % Assign source texture for next pass:
    inputtex = futuretex;
%glFinish;
%gpupass = toc * 1000
end;

glFinish
gvfduration = toc * 1000
gvfperiteration = gvfduration / numiters

% Shut down the GVF iteration shader...
glUseProgram(0);
% ...and both texture units:
glDisable(GL.TEXTURE_RECTANGLE_EXT);
glBindTexture(GL.TEXTURE_RECTANGLE_EXT, 0);
glActiveTexture(GL.TEXTURE1);
glDisable(GL.TEXTURE_RECTANGLE_EXT);
glBindTexture(GL.TEXTURE_RECTANGLE_EXT, 0);
glActiveTexture(GL.TEXTURE0);

% Assign read buffer for GVF readout:
glReadBuffer(GL.COLOR_ATTACHMENT0_EXT + bufferid);

% Readout final result: RED = flow_x, GREEN = flow_y:
gvf_finalimg = glReadPixels(0, 0, msize, msize, GL_RGB, GL_FLOAT);

% Final flow_x:
imagesc(gvf_finalimg(:,:,1));
figure

% Final flow_y:
imagesc(gvf_finalimg(:,:,2));
figure

% Unbind FBO, reset to normal framebuffer:
moglChooseFBO(0);

% Delete all FBO's:
for i=1:length(fbos)
    moglDeleteFBO(fbos(i));
end;

% Shut down OpenGL rendering:
Screen('EndOpenGL', win);

% Close onscreen window and release all other ressources:
Screen('CloseAll');

%minu = min(min(gvfc_u))
minu = min(min(gvf_finalimg(:,:,1)))
%maxu = max(max(gvfc_u))
maxu = max(max(gvf_finalimg(:,:,1)))

return

imagesc(gvfc_u);
figure;
imagesc(gvfc_v);
figure;
diff_u = gvfc_u - gvf_finalimg(:,:,1);
diff_v = gvfc_v - gvf_finalimg(:,:,2);
imagesc(diff_u);
figure;
imagesc(diff_v);

maxdiff=max(max(abs(diff_u) + abs(diff_v)))
avgdiff=mean(mean(abs(diff_u) + abs(diff_v)))
% Well done!
return