https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80724
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |missed-optimization
Status|UNCONFIRMED |NEW
Last reconfirmed| |2017-07-26
CC| |pinskia at gcc dot gnu.org
Target Milestone|--- |8.0
Summary|gcc.target/aarch64/pr62178. |[8 Regression]
|c failed because of r247885 |gcc.target/aarch64/pr62178.
| |c failed because of r247885
Ever confirmed|0 |1
--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
The vectorized loop looks like:
.L7:
ldr w2, [x1, 4]!
ldr q1, [x0], 124
fmov s3, w2
mla v0.4s, v1.4s, v3.s[0]
cmp x0, x3
bne .L7
Which is worse than before in GCC 7:
.L7:
ld1r {v1.4s}, [x1], 4
ldr q2, [x0], 124
mla v0.4s, v2.4s, v1.4s
cmp x0, x2
bne .L7
