https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65968
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |missed-optimization
Status|UNCONFIRMED |NEW
Last reconfirmed| |2015-12-23
Ever confirmed|0 |1
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.
AARCH64 it is not as bad as x86_64 though:
With -UEASY:
.L2:
ldr q0, [x0]
smull v2.4s, v0.4h, v0.4h
smull2 v0.4s, v0.8h, v0.8h
xtn v1.4h, v2.4s
xtn2 v1.8h, v0.4s
str q1, [x0], 16
cmp x1, x0
bne .L2
With -DEASY:
.L2:
ldr q0, [x0]
mul v0.8h, v0.8h, v0.8h
str q0, [x0], 16
cmp x1, x0
bne .L2