https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112411
--- Comment #4 from Alexandre Oliva <aoliva at gcc dot gnu.org> --- yeah, its intended use case is for debugging -fcompare-debug fails, reducing superfluous differences between rtl dumps. In theory GCC may overflow insn uids regardless of this param, and if guarding against that is important, there should be code to check for it when incrementing insn uids, but I suppose we can live without that. Such limit values are definitely not useful for fuzz testing, but I'd be curious whether values that do not cause overflows would cause any codegen differences. I think not, and it really shouldn't, but this thought makes me hesitant to recommend against using this option for fuzz testing.