generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Merge "sad4d neon: 64x[32,64]"

Browse Source
This commit is contained in:
Johann Koenig 2017-07-12 20:15:00 +00:00 committed by Gerrit Code Review
commit e0d79bc7b5
4 changed files with 66 additions and 82 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
test/sad_test.cc
View File

@ -676,6 +676,7 @@ INSTANTIATE_TEST_CASE_P(NEON, SADavgTest, ::testing::ValuesIn(avg_neon_tests));
const SadMxNx4Param x4d_neon_tests[] = { const SadMxNx4Param x4d_neon_tests[] = {
SadMxNx4Param(64, 64, &vpx_sad64x64x4d_neon), SadMxNx4Param(64, 64, &vpx_sad64x64x4d_neon),
SadMxNx4Param(64, 32, &vpx_sad64x32x4d_neon),
SadMxNx4Param(32, 64, &vpx_sad32x64x4d_neon), SadMxNx4Param(32, 64, &vpx_sad32x64x4d_neon),
SadMxNx4Param(32, 32, &vpx_sad32x32x4d_neon), SadMxNx4Param(32, 32, &vpx_sad32x32x4d_neon),
SadMxNx4Param(32, 16, &vpx_sad32x16x4d_neon), SadMxNx4Param(32, 16, &vpx_sad32x16x4d_neon),

136
vpx_dsp/arm/sad4d_neon.c
View File

@ -176,93 +176,67 @@ void vpx_sad32x64x4d_neon(const uint8_t *src, int src_stride,
sad32x_4d(src, src_stride, ref, ref_stride, res, 64); sad32x_4d(src, src_stride, ref, ref_stride, res, 64);
} }
static INLINE unsigned int horizontal_long_add_16x8(const uint16x8_t vec_lo, static INLINE void sum64x(const uint8x16_t a_0, const uint8x16_t a_1,
const uint16x8_t vec_hi) { const uint8x16_t b_0, const uint8x16_t b_1,
const uint32x4_t vec_l_lo = uint16x8_t *sum) {
vaddl_u16(vget_low_u16(vec_lo), vget_high_u16(vec_lo)); *sum = vabal_u8(*sum, vget_low_u8(a_0), vget_low_u8(b_0));
const uint32x4_t vec_l_hi = *sum = vabal_u8(*sum, vget_high_u8(a_0), vget_high_u8(b_0));
vaddl_u16(vget_low_u16(vec_hi), vget_high_u16(vec_hi)); *sum = vabal_u8(*sum, vget_low_u8(a_1), vget_low_u8(b_1));
const uint32x4_t a = vaddq_u32(vec_l_lo, vec_l_hi); *sum = vabal_u8(*sum, vget_high_u8(a_1), vget_high_u8(b_1));
const uint64x2_t b = vpaddlq_u32(a);
const uint32x2_t c = vadd_u32(vreinterpret_u32_u64(vget_low_u64(b)),
vreinterpret_u32_u64(vget_high_u64(b)));
return vget_lane_u32(c, 0);
} }
// Calculate the absolute difference of 64 bytes from vec_src_00, vec_src_16, static INLINE void sad64x_4d(const uint8_t *a, int a_stride,
// vec_src_32, vec_src_48 and ref. Accumulate partial sums in vec_sum_ref_lo const uint8_t *const b[4], int b_stride,
// and vec_sum_ref_hi. uint32_t *result, const int height) {
static void sad_neon_64(const uint8x16_t vec_src_00, int i;
const uint8x16_t vec_src_16, uint16x8_t sum_0 = vdupq_n_u16(0);
const uint8x16_t vec_src_32, uint16x8_t sum_1 = vdupq_n_u16(0);
const uint8x16_t vec_src_48, const uint8_t *ref, uint16x8_t sum_2 = vdupq_n_u16(0);
uint16x8_t *vec_sum_ref_lo, uint16x8_t sum_3 = vdupq_n_u16(0);
uint16x8_t *vec_sum_ref_hi) { uint16x8_t sum_4 = vdupq_n_u16(0);
const uint8x16_t vec_ref_00 = vld1q_u8(ref); uint16x8_t sum_5 = vdupq_n_u16(0);
const uint8x16_t vec_ref_16 = vld1q_u8(ref + 16); uint16x8_t sum_6 = vdupq_n_u16(0);
const uint8x16_t vec_ref_32 = vld1q_u8(ref + 32); uint16x8_t sum_7 = vdupq_n_u16(0);
const uint8x16_t vec_ref_48 = vld1q_u8(ref + 48); const uint8_t *b_loop[4] = { b[0], b[1], b[2], b[3] };
*vec_sum_ref_lo = vabal_u8(*vec_sum_ref_lo, vget_low_u8(vec_src_00), for (i = 0; i < height; ++i) {
vget_low_u8(vec_ref_00)); const uint8x16_t a_0 = vld1q_u8(a);
*vec_sum_ref_hi = vabal_u8(*vec_sum_ref_hi, vget_high_u8(vec_src_00), const uint8x16_t a_1 = vld1q_u8(a + 16);
vget_high_u8(vec_ref_00)); const uint8x16_t a_2 = vld1q_u8(a + 32);
*vec_sum_ref_lo = vabal_u8(*vec_sum_ref_lo, vget_low_u8(vec_src_16), const uint8x16_t a_3 = vld1q_u8(a + 48);
vget_low_u8(vec_ref_16)); a += a_stride;
*vec_sum_ref_hi = vabal_u8(*vec_sum_ref_hi, vget_high_u8(vec_src_16), sum64x(a_0, a_1, vld1q_u8(b_loop[0]), vld1q_u8(b_loop[0] + 16), &sum_0);
vget_high_u8(vec_ref_16)); sum64x(a_2, a_3, vld1q_u8(b_loop[0] + 32), vld1q_u8(b_loop[0] + 48),
*vec_sum_ref_lo = vabal_u8(*vec_sum_ref_lo, vget_low_u8(vec_src_32), &sum_1);
vget_low_u8(vec_ref_32)); b_loop[0] += b_stride;
*vec_sum_ref_hi = vabal_u8(*vec_sum_ref_hi, vget_high_u8(vec_src_32), sum64x(a_0, a_1, vld1q_u8(b_loop[1]), vld1q_u8(b_loop[1] + 16), &sum_2);
vget_high_u8(vec_ref_32)); sum64x(a_2, a_3, vld1q_u8(b_loop[1] + 32), vld1q_u8(b_loop[1] + 48),
*vec_sum_ref_lo = vabal_u8(*vec_sum_ref_lo, vget_low_u8(vec_src_48), &sum_3);
vget_low_u8(vec_ref_48)); b_loop[1] += b_stride;
*vec_sum_ref_hi = vabal_u8(*vec_sum_ref_hi, vget_high_u8(vec_src_48), sum64x(a_0, a_1, vld1q_u8(b_loop[2]), vld1q_u8(b_loop[2] + 16), &sum_4);
vget_high_u8(vec_ref_48)); sum64x(a_2, a_3, vld1q_u8(b_loop[2] + 32), vld1q_u8(b_loop[2] + 48),
&sum_5);
b_loop[2] += b_stride;
sum64x(a_0, a_1, vld1q_u8(b_loop[3]), vld1q_u8(b_loop[3] + 16), &sum_6);
sum64x(a_2, a_3, vld1q_u8(b_loop[3] + 32), vld1q_u8(b_loop[3] + 48),
&sum_7);
b_loop[3] += b_stride;
}
result[0] = vget_lane_u32(horizontal_add_long_uint16x8(sum_0, sum_1), 0);
result[1] = vget_lane_u32(horizontal_add_long_uint16x8(sum_2, sum_3), 0);
result[2] = vget_lane_u32(horizontal_add_long_uint16x8(sum_4, sum_5), 0);
result[3] = vget_lane_u32(horizontal_add_long_uint16x8(sum_6, sum_7), 0);
}
void vpx_sad64x32x4d_neon(const uint8_t *src, int src_stride,
const uint8_t *const ref[4], int ref_stride,
uint32_t *res) {
sad64x_4d(src, src_stride, ref, ref_stride, res, 32);
} }
void vpx_sad64x64x4d_neon(const uint8_t *src, int src_stride, void vpx_sad64x64x4d_neon(const uint8_t *src, int src_stride,
const uint8_t *const ref[4], int ref_stride, const uint8_t *const ref[4], int ref_stride,
uint32_t *res) { uint32_t *res) {
int i; sad64x_4d(src, src_stride, ref, ref_stride, res, 64);
uint16x8_t vec_sum_ref0_lo = vdupq_n_u16(0);
uint16x8_t vec_sum_ref0_hi = vdupq_n_u16(0);
uint16x8_t vec_sum_ref1_lo = vdupq_n_u16(0);
uint16x8_t vec_sum_ref1_hi = vdupq_n_u16(0);
uint16x8_t vec_sum_ref2_lo = vdupq_n_u16(0);
uint16x8_t vec_sum_ref2_hi = vdupq_n_u16(0);
uint16x8_t vec_sum_ref3_lo = vdupq_n_u16(0);
uint16x8_t vec_sum_ref3_hi = vdupq_n_u16(0);
const uint8_t *ref0, *ref1, *ref2, *ref3;
ref0 = ref[0];
ref1 = ref[1];
ref2 = ref[2];
ref3 = ref[3];
for (i = 0; i < 64; ++i) {
const uint8x16_t vec_src_00 = vld1q_u8(src);
const uint8x16_t vec_src_16 = vld1q_u8(src + 16);
const uint8x16_t vec_src_32 = vld1q_u8(src + 32);
const uint8x16_t vec_src_48 = vld1q_u8(src + 48);
sad_neon_64(vec_src_00, vec_src_16, vec_src_32, vec_src_48, ref0,
&vec_sum_ref0_lo, &vec_sum_ref0_hi);
sad_neon_64(vec_src_00, vec_src_16, vec_src_32, vec_src_48, ref1,
&vec_sum_ref1_lo, &vec_sum_ref1_hi);
sad_neon_64(vec_src_00, vec_src_16, vec_src_32, vec_src_48, ref2,
&vec_sum_ref2_lo, &vec_sum_ref2_hi);
sad_neon_64(vec_src_00, vec_src_16, vec_src_32, vec_src_48, ref3,
&vec_sum_ref3_lo, &vec_sum_ref3_hi);
src += src_stride;
ref0 += ref_stride;
ref1 += ref_stride;
ref2 += ref_stride;
ref3 += ref_stride;
}
res[0] = horizontal_long_add_16x8(vec_sum_ref0_lo, vec_sum_ref0_hi);
res[1] = horizontal_long_add_16x8(vec_sum_ref1_lo, vec_sum_ref1_hi);
res[2] = horizontal_long_add_16x8(vec_sum_ref2_lo, vec_sum_ref2_hi);
res[3] = horizontal_long_add_16x8(vec_sum_ref3_lo, vec_sum_ref3_hi);
} }

9
vpx_dsp/arm/sum_neon.h
View File

@ -30,6 +30,15 @@ static INLINE uint32x2_t horizontal_add_uint16x8(const uint16x8_t a) {
vreinterpret_u32_u64(vget_high_u64(c))); vreinterpret_u32_u64(vget_high_u64(c)));
} }
static INLINE uint32x2_t horizontal_add_long_uint16x8(const uint16x8_t a,
const uint16x8_t b) {
const uint32x4_t c = vpaddlq_u16(a);
const uint32x4_t d = vpadalq_u16(c, b);
const uint64x2_t e = vpaddlq_u32(d);
return vadd_u32(vreinterpret_u32_u64(vget_low_u64(e)),
vreinterpret_u32_u64(vget_high_u64(e)));
}
static INLINE uint32x2_t horizontal_add_uint32x4(const uint32x4_t a) { static INLINE uint32x2_t horizontal_add_uint32x4(const uint32x4_t a) {
const uint64x2_t b = vpaddlq_u32(a); const uint64x2_t b = vpaddlq_u32(a);
return vadd_u32(vreinterpret_u32_u64(vget_low_u64(b)), return vadd_u32(vreinterpret_u32_u64(vget_low_u64(b)),

2
vpx_dsp/vpx_dsp_rtcd_defs.pl
View File

@ -860,7 +860,7 @@ add_proto qw/void vpx_sad64x64x4d/, "const uint8_t *src_ptr, int src_stride, con
specialize qw/vpx_sad64x64x4d avx2 neon msa sse2 vsx/; specialize qw/vpx_sad64x64x4d avx2 neon msa sse2 vsx/;
add_proto qw/void vpx_sad64x32x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array"; add_proto qw/void vpx_sad64x32x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
specialize qw/vpx_sad64x32x4d msa sse2 vsx/; specialize qw/vpx_sad64x32x4d neon msa sse2 vsx/;
add_proto qw/void vpx_sad32x64x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array"; add_proto qw/void vpx_sad32x64x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
specialize qw/vpx_sad32x64x4d neon msa sse2 vsx/; specialize qw/vpx_sad32x64x4d neon msa sse2 vsx/;