[llvm-bugs] [Bug 26428] LICM: Ignore loop exits which don't exit on first iteration when computing isGuaranteedToExecute

via llvm-bugs llvm-bugs at lists.llvm.org
Tue Jun 12 17:27:17 PDT 2018


listmail at philipreames.com changed:

           What    |Removed                     |Added
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED
           Assignee|ovmold at gmail.com            |listmail at philipreames.com

--- Comment #6 from listmail at philipreames.com ---
Closing the bug as support was implemented a while back.  There are some next
steps which would be nice to pursue, let me summarize them:
1) pipe SCEV through LICM and down into MustExecute.  This lets us handle much
more complicated cases than the current Constants and SimplifyInstruction
2) figure out a way to efficiently handle implicit control flow (i.e. handle
cases where potentially faulting instructions only run on the second iteration
down some rare path...)  I'd been working on this, but have basically given up
now and approached my original problem from a different angel.

You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20180613/764ba467/attachment.html>

More information about the llvm-bugs mailing list