From: Kyrylo Tkachov <[email protected]>
The test hard-codes two "fcmla ..., #0", which was the number trunk
happened to emit when it was added. BB SLP now also vectorises the two
plain-multiply subroutines: r17-141-g76b8869f08a6 ("tree-optimization/
124222 - rewrite BB SLP costing scalar coverage") taught
vect_bb_slp_scalar_cost to attribute the statements an SLP pattern node
covers, so the subgraph for c_add_ab and c_sub_ab is no longer costed as
unprofitable. They form .COMPLEX_MUL, and cmul<mode>3 expands each to
"fcmla #0" plus "fcmla #90", taking the #0 total from two to four.
For c_add_ab the loop body goes from a scalar pair of complex multiplies
ldr d26, [x5, x0]
ldr d27, [x3, x0]
ldr d29, [x2, x0]
fmul d24, d30, d26
ldr d28, [x1, x0]
fnmsub d24, d31, d27, d24
fmul d25, d30, d27
fmadd d25, d31, d26, d25
fadd d24, d29, d24
fadd d25, d25, d28
str d24, [x2, x0]
str d25, [x1, x0]
add x0, x0, 16
cmp x4, x0
bne .L3
to the vectorised form
ldr q29, [x2, x0]
movi v27.4s, 0
ldr q28, [x3, x0]
fcmla v27.2d, v28.2d, v31.2d, #0
fcmla v27.2d, v28.2d, v31.2d, #90
fadd v27.2d, v27.2d, v29.2d
str q27, [x2, x0]
add x0, x0, 16
cmp x0, x1
bne .L3
i.e. ten instructions instead of fifteen, and the whole rotation mix
across the file changes from
#0=2 #90=0 #180=0 #270=2
to
#0=4 #90=2 #180=0 #270=2
The loop vectoriser is unchanged: its dumps are identical before and
after, the two conjugate subroutines still form .COMPLEX_MUL_CONJ, and
the #270 directive that actually tests PR122408 still passes. The
runtime companion pr122408_2.f90 also still runs clean.
Update the counts, and add a #90 count and a #180 scan-assembler-not.
PR122408 was about picking the wrong rotation pair after operand
swapping, so pinning the full rotation mix guards the regression more
directly than the #0 count did.
Ok for trunk?
Thanks,
Kyrill
gcc/testsuite/ChangeLog:
* gfortran.target/aarch64/pr122408_1.f90: Update the fcmla
rotation counts.
Signed-off-by: Kyrylo Tkachov <[email protected]>
---
gcc/testsuite/gfortran.target/aarch64/pr122408_1.f90 | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/gcc/testsuite/gfortran.target/aarch64/pr122408_1.f90
b/gcc/testsuite/gfortran.target/aarch64/pr122408_1.f90
index 8a3416231ff..c1996b74869 100644
--- a/gcc/testsuite/gfortran.target/aarch64/pr122408_1.f90
+++ b/gcc/testsuite/gfortran.target/aarch64/pr122408_1.f90
@@ -57,5 +57,10 @@ subroutine c_sub_a_conjb(n, a, c, b) ! C -= A * conj(B)
end do
end subroutine c_sub_a_conjb
-! { dg-final { scan-assembler-times {fcmla\s+v[0-9]+.2d, v[0-9]+.2d,
v[0-9]+.2d, #0} 2 } }
+! The two plain multiplies form .COMPLEX_MUL (#0 + #90) and the two conjugate
+! multiplies form .COMPLEX_MUL_CONJ (#0 + #270). PR122408 is about detecting
+! the conjugate form, so the #270 and #180 counts are the ones that guard it.
+! { dg-final { scan-assembler-times {fcmla\s+v[0-9]+.2d, v[0-9]+.2d,
v[0-9]+.2d, #0} 4 } }
+! { dg-final { scan-assembler-times {fcmla\s+v[0-9]+.2d, v[0-9]+.2d,
v[0-9]+.2d, #90} 2 } }
! { dg-final { scan-assembler-times {fcmla\s+v[0-9]+.2d, v[0-9]+.2d,
v[0-9]+.2d, #270} 2 } }
+! { dg-final { scan-assembler-not {fcmla\s+v[0-9]+.2d, v[0-9]+.2d, v[0-9]+.2d,
#180} } }
--
2.50.1 (Apple Git-155)
