[PATCH] D43142: Experimental pass to convert all floating point operations to the equivalent constrained intrinsics

[Andy Kaylor via Phabricator via llvm-commits]

andrew.w.kaylor created this revision.
andrew.w.kaylor added reviewers: spatel, hfinkel, scanon.
Herald added subscribers: hintonda, mgorny, mehdi_amini.

I'm not sure this should actually be committed. I'm mostly putting it here as a way to make it available to anyone who is interested.

If anyone would like to see it added to the code base, feel free to make comments and I'll clean it up accordingly.

Right now, it only supports the legacy pass manager. Otherwise, I think it works OK.


https://reviews.llvm.org/D43142

Files:
  include/llvm/InitializePasses.h
  include/llvm/Transforms/IPO.h
  lib/Transforms/IPO/CMakeLists.txt
  lib/Transforms/IPO/ForceConstrainedFP.cpp
  lib/Transforms/IPO/IPO.cpp
  lib/Transforms/IPO/PassManagerBuilder.cpp
  test/Transforms/force-constrained-fp.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D43142.133682.patch
Type: text/x-patch
Size: 14125 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180209/97f0d691/attachment.bin>