[PATCH] [lld][Core] Implement parallel_for_each

[Shankar Kalpathi Easwaran]

I agree dividing the tasks by the hardware concurrency could get optimal performance. If you see parallel_sort(few lines above) the code tries to divide tasks by 1024. This file needs to be cleaned up based on your suggestion and I feel we could implement that later, just for allowing others to try various loops to be optimized based on this patch. I am fine anyways but I think creating an optimal solution would make others wait.


http://reviews.llvm.org/D8348

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/