Closed Martinsos closed 8 years ago
I should try swapping block calculate function with macro, maybe that would make things faster?
I gave some more thought to this and I do not think it would help! From profiling that I have done so far, it does overhead for function call is problem.
I am not sure what else can I do to speed up this part! I will leave this open for when I have more time, but I have no brilliant ideas.
I removed ifs from calculation of block, which is core of whole algorithm. That gave some speed, about 30% - 40% speedup. What I should do further: investigate if operations in block calculation can be further simplified! That would bring more speedup