Applying code optimization based on hardware resources, Making a simple image processing library to test different approaches
Optimized mini‐img‐processing libary implemented from scratch using C, Pthreads, And Intel‐intrinsics.
-
Sequential thresholding : linear pass
-
unrolled thresholding : using loop unrolling
-
Fast thresholding : using loop unrolling + precomputed addresses
-
SSE2 thresholding : using vectorization
-
parallel thresholding : using threads + SSE2
-
First excutable "imgModuleRelease" is compiled with '-msse2' '-O3' flags