http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53712
--- Comment #2 from Uros Bizjak <ubizjak at gmail dot com> 2012-06-18 10:03:11 UTC --- (In reply to comment #1) > the testcase works as expected. Still with the "redundant"(?) instruction > though. Thus your source is invalid but the missed-optimization looks > odd (though it's only there at -O0). It also misses to combine the > unaligned load into the cmpistri instruction. The "redundant" insn is there by design, the splitter for pcmpistri insn checks for dead outputs. Since at -O0 liveness analysis isn't run, we split to both forms of pcmpistri instructions. I will add _unaligned forms of pcmistri that will consume UNSPEC_MOVU operator.
