/*******/
#include <arm_neon.h>
void neon_add(float * __restrict out, float * __restrict a, float * __restrict
b)
{
float32x2_t tmp1, tmp2;
tmp1 = vset_lane_f32(*a, tmp1, 0);
tmp2 = vset_lane_f32(*b, tmp2, 0);
tmp1 = vadd_f32(tmp1, tmp2);
*out = vget_lane_f32(tmp1, 0);
}
/*******/
00000000 <neon_add>:
0: e5913000 ldr r3, [r1]
4: eddf0b07 vldr d16, [pc, #28] ; 28 <neon_add+0x28>
8: e5922000 ldr r2, [r2]
c: eddf1b05 vldr d17, [pc, #20] ; 28 <neon_add+0x28>
10: ee003b90 vmov.32 d16[0], r3
14: ee012b90 vmov.32 d17[0], r2
18: f2400da1 vadd.f32 d16, d16, d17
1c: f4c0080f vst1.32 {d16[0]}, [r0]
20: e12fff1e bx lr
24: e1a00000 nop (mov r0,r0)
gcc fails to use a single instruction
vld1.32 {d16[0]}, [r1]
instead of
0: e5913000 ldr r3, [r1]
4: eddf0b07 vldr d16, [pc, #28] ; 28 <neon_add+0x28>
10: ee003b90 vmov.32 d16[0], r3
--
Summary: Suboptimal code for the use of ARM NEON intrinsic
"vset_lane_f32"
Product: gcc
Version: 4.4.3
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: siarhei dot siamashka at gmail dot com
GCC build triplet: arm-unknown-linux-gnueabi
GCC host triplet: arm-unknown-linux-gnueabi
GCC target triplet: arm-unknown-linux-gnueabi
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43364
