[PATCH] [lld][Core] Implement parallel_for_each

[Shankar Kalpathi Easwaran]

I agree with David's comment, it doesnot show up with any considerable improvement.

I happen to get an improved number (0.5 second improvement) only **once** :-

19.55user 0.78system 0:11.83elapsed 171%CPU (0avgtext+0avgdata 7037040maxresident)k

I usually get time taken as below **without** this patch :-

20.06user 0.74system 0:12.55elapsed 165%CPU (0avgtext+0avgdata 7042880maxresident)k

This patch gives us a benefit of having a version of parallel_for_each on unix platforms without degrading performance.


http://reviews.llvm.org/D8348

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