Image processing using Open MPI
Filters will be applied depending on the number of processes: - one process: the "root" process will read the binary file received as input and will apply the respective matrix to the received filters as an argument for each pixel; the "root" process will write the resulting image to output file;
- for several processes: the "root" process will read the received binary file as an input and will send the pixel array to all processes; every process (including the "root" process) will apply the respective matrix to the filters received as an argument on the pixels of a number N of lines in the matrix of pixels: N = number of lines / number of processes if the partition is not exact, the "root" process will handle the rest of the lines; the "root" process will receive the rest of the image from the others processes and writes the resulting image to the output file;
Computer specifications:
Architecture: x86_64 CPU op-mode (s): 32-bit, 64-bit Byte Order: Little Endian CPU (s): 4 Online CPU (s) list: 0-3 Thread (s) per core: 2 Core (s) per socket: 2 Socket (s): 1 (Linux subsystem)
Execution time (s):
Image: landscape.pgm (325 KB) no. process -> 1 2 3 4 5 6 7 8 applied filters \ / smooth 0.875 0.76 0.698 0.656 0.535 0.512 0.486 0.442
bssembssem 0.918 0.854 0.767 0.709 0.638 0.595 0.5058 0.472
Image: rorschach.pgm (7526 KB) no. process -> 1 2 3 4 5 6 7 8 applied filters \ / smooth 1,741 1,680 1,634 1,616 1,442 1,141 1,373 1,279
bssembssem 4.733 4.268 3.97 3.622 3.48 3.370 2.977 2.696
Image: scroll-of-truth.pnm (756 KB) no. process -> 1 2 3 4 5 6 7 8 applied filters \ / smooth 0.832 0.779 0.697 0.664 0.613 0.542 0.527 0.451
bssembssem 1,227 1,088 0.972 0.917 0.811 0.775 0.691 0.622
Image: landscape.pnm (24301 KB)
smooth 4.449 4.297 4.226 4.145 4.011 3.901 3.481 3.34
bssembssem 12.93 12.44 12.011 12.004 11.705 11.683 11.339 10.801