[Bug tree-optimization/94919] Failure to recognize max pattern

["gabravier at gmail dot com" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94919

--- Comment #2 from Gabriel Ravier <gabravier at gmail dot com> ---
Essentially, what I've been doing in my spare time for the past few weeks is
looking at random pieces of code all over the internet, looking at the results
trunk gcc/clang give (usually on x86-64 (though I've been thinking about
looking at arm and other architectures), with flags like -Ofast
-march=tigerlake) and seeing whether GCC or LLVM optimizes them better than the
other (and making appropriate bug reports). Some of these are from actual
projects, some are from websites listing bithacks, some are from books like
Hacker's Delight, some were written by myself when studying the code from those
above sources, and finally a few of them were from testsuites in either LLVM or
GCC (though those are rarer, especially for LLVM, since it takes time to
convert them from LLVM bytecode to something I can compile with both GCC and
Clang).

For this particular bug report, I believe this is *probably* from somewhere in
Hacker's Delight.