https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89582
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Target| |x86_64-*-*
Status|UNCONFIRMED |NEW
Last reconfirmed| |2019-03-05
CC| |rguenth at gcc dot gnu.org
Blocks| |53947
Depends on| |84101
Ever confirmed|0 |1
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
There's a duplicate for this (or slightly different testcases with the same
underlying issue). The vectorizer knows nothing about the target ABI
and assumes the arguments reside in memory while they are actually passed in
two %xmm registers each. Same for the return value. PR84101 is one of
the related PRs.
Referenced Bugs:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=53947
[Bug 53947] [meta-bug] vectorizer missed-optimizations
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84101
[Bug 84101] [7/8/9 Regression] -O3 and -ftree-vectorize trying too hard for
function returning trivial pair-of-uint64_t-structure
