https://gcc.gnu.org/bugzilla/show_bug.cgi?id=46091
--- Comment #11 from Uroš Bizjak <ubizjak at gmail dot com> --- (In reply to Uroš Bizjak from comment #10) > I'll look into this a bit some more. However, these insn should be rare, so > do not expect any noticeable application speed-up ... From the Agner lists, it is not clear if "bitop mem, imm" is faster than "movabs r, imm + logic_op mem, r". Looking at several tables, it looks that that it is not; for some AMD and older Intel targets, RMW bitop looses considerably.
