[llvm-bugs] [Bug 43420] __builtin_constant_p resolves differently with loop unrolling at -O3

via llvm-bugs llvm-bugs at lists.llvm.org
Mon Sep 23 17:06:24 PDT 2019


Eli Friedman <efriedma at quicinc.com> changed:

           What    |Removed                     |Added
         Resolution|---                         |WONTFIX
                 CC|                            |efriedma at quicinc.com
             Status|NEW                         |RESOLVED

--- Comment #2 from Eli Friedman <efriedma at quicinc.com> ---
Apparently, gcc actually has multiple unrolling passes; at least one runs
before __builtin_constant_p is resolved, and at least one runs after.  (You can
see this by modifying the trip count of the loop.)

We aren't going to try to match the exact unroll threshold gcc uses for "early"
unrolling at -O3.

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/20190924/c6b10d0f/attachment.html>

More information about the llvm-bugs mailing list