[lld] r286061 - [ELF] - Implemented threaded --build-id computation

[George Rimar via llvm-commits]

>Have you benchmarked this on windows? I was about to benchmark this on
>linux but noticed that parallel_for_each does nothing there since it
>uses a thread for each 1024 elements, so in this case it would only
>use more than one thread if the file was more than a GiB.

No I didn't benchmark on windows.
Do you want me to try ? (seems your patch already fixed this for linux case).

George.