[llvm-bugs] [Bug 48408] New: Failure to optimize out comparison when very narrow range is already known

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=48408

            Bug ID: 48408
           Summary: Failure to optimize out comparison when very narrow
                    range is already known
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Scalar Optimizations
          Assignee: unassignedbugs at nondot.org
          Reporter: gabravier at gmail.com
                CC: llvm-bugs at lists.llvm.org

#ifndef __clang__
__attribute__((always_inline)) inline void __builtin_assume(bool condition)
{
    if (!condition)
        __builtin_unreachable();
}
#endif

bool f(int a)
{
    __builtin_assume((a == -1) || (a == 0));
    return a < -1;
}

This can be optimized to `return false;`. This transformation is done by GCC,
but not by LLVM.

See also Godbolt comparison here : https://godbolt.org/z/8x5x89

-- 
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/20201206/03384a77/attachment.html>