Files
Ciro Santilli 六四事件 法轮功 11359838a1 assembly: improve organization of simd add
2019-05-12 00:00:09 +00:00

32 lines
769 B
ArmAsm

/* https://github.com/cirosantilli/linux-kernel-module-cheat#x86-addpq-instruction
*
* Add a bunch of floating point numbers in one go.
*/
#include "common.h"
ENTRY
.bss
output: .skip 16
.data
addps_input0: .float 1.5, 2.5, 3.5, 4.5
addps_input1: .float 5.5, 6.5, 7.5, 8.5
addps_expect: .float 7.0, 9.0, 11.0, 13.0
addpd_input0: .double 1.5, 2.5
addpd_input1: .double 5.5, 6.5
addpd_expect: .double 7.0, 9.0
.text
#define TEST(size) \
movups addp ## size ## _input0, %xmm0; \
movups addp ## size ## _input1, %xmm1; \
addp ## size %xmm1, %xmm0; \
movups %xmm0, output; \
ASSERT_MEMCMP(output, addp ## size ## _expect, $0x10)
/* 4x 32-bit */
TEST(s)
/* 2x 64-bit */
TEST(d)
#undef TEST
EXIT