https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85050
--- Comment #3 from H.J. Lu <hjl.tools at gmail dot com> --- (In reply to Richard Biener from comment #2) > IIRC there's a related bug with regard to double <-> float conversions and > SSE > vectorization. Can't find it right now. PR 36844?
