https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85077
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Priority|P3 |P1
Status|WAITING |ASSIGNED
CC| |jakub at gcc dot gnu.org
Assignee|unassigned at gcc dot gnu.org |jakub at gcc dot gnu.org
Target Milestone|--- |8.0
Summary|V[248][SD]F abs not |[8 Regression] V[248][SD]F
|optimized to |abs not optimized to
--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
This regressed with r248124, before that change the FE would fold the two
functions to ABS_EXPR <x> but now it emits VEC_COND_EXPR < x < {0.0, 0.0, 0.0,
0.0} , -NON_LVALUE_EXPR <x> , x > and nothing recognizes that later. I'll have
a look.
