[PATCH] D36963: Use sorting instead of a lot of stable_partitions for `ICF::segregate` function

Fri Aug 25 18:24:13 PDT 2017

alex.telishev updated this revision to Diff 112768.
alex.telishev added a comment.

Fix `for_each_n` policy to par instead of seq.  This patch speed is the same as before (much faster than without).

About testing - I linked clang and my project with old version, older patch and newer patch without /DEBUG (with /DEBUG lld crashed). For clang all 3 versions were consistently getting binary the same result. For my project all 3 versions were getting sometimes different results.

https://reviews.llvm.org/D36963

Files:
  lld/COFF/ICF.cpp

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D36963.112768.patch
Type: text/x-patch
Size: 7463 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170826/90ebb410/attachment.bin>