ref: 4e6b5079c65ccf396c0c4a606dd7e13ccb7da54f
parent: 8f6b9431001b96fa9154271a13ea1d175ca2796f
author: Scott LaVarnway <[email protected]>
date: Wed Aug 5 06:47:06 EDT 2015
VPX: remove scaled calls from FUN_CONV_1D and FUN_CONV_2D macros. The predict lut now handles this case. The encoder now calls vpx_scaled_2d() instead of vpx_convolve8() for scaling. Change-Id: Ia1c8af8a31e4cb4887a587143108cb45835f7df7
--- a/test/convolve_test.cc
+++ b/test/convolve_test.cc
@@ -41,9 +41,14 @@
ConvolveFunc h8, ConvolveFunc h8_avg,
ConvolveFunc v8, ConvolveFunc v8_avg,
ConvolveFunc hv8, ConvolveFunc hv8_avg,
+ ConvolveFunc sh8, ConvolveFunc sh8_avg,
+ ConvolveFunc sv8, ConvolveFunc sv8_avg,
+ ConvolveFunc shv8, ConvolveFunc shv8_avg,
int bd)
: copy_(copy), avg_(avg), h8_(h8), v8_(v8), hv8_(hv8), h8_avg_(h8_avg),
- v8_avg_(v8_avg), hv8_avg_(hv8_avg), use_highbd_(bd) {}+ v8_avg_(v8_avg), hv8_avg_(hv8_avg), sh8_(sh8), sv8_(sv8), shv8_(shv8),
+ sh8_avg_(sh8_avg), sv8_avg_(sv8_avg), shv8_avg_(shv8_avg),
+ use_highbd_(bd) {}ConvolveFunc copy_;
ConvolveFunc avg_;
@@ -53,6 +58,12 @@
ConvolveFunc h8_avg_;
ConvolveFunc v8_avg_;
ConvolveFunc hv8_avg_;
+ ConvolveFunc sh8_; // scaled horiz
+ ConvolveFunc sv8_; // scaled vert
+ ConvolveFunc shv8_; // scaled horiz/vert
+ ConvolveFunc sh8_avg_; // scaled avg horiz
+ ConvolveFunc sv8_avg_; // scaled avg vert
+ ConvolveFunc shv8_avg_; // scaled avg horiz/vert
int use_highbd_; // 0 if high bitdepth not used, else the actual bit depth.
};
@@ -597,8 +608,8 @@
DECLARE_ALIGNED(256, const int16_t, filter8[8]) = {0, 0, 0, 128, 0, 0, 0, 0};ASM_REGISTER_STATE_CHECK(
- UUT_->h8_(in, kInputStride, out, kOutputStride, filter8, 16, filter8, 16,
- Width(), Height()));
+ UUT_->sh8_(in, kInputStride, out, kOutputStride, filter8, 16, filter8, 16,
+ Width(), Height()));
CheckGuardBlocks();
@@ -615,8 +626,8 @@
DECLARE_ALIGNED(256, const int16_t, filter8[8]) = {0, 0, 0, 128, 0, 0, 0, 0};ASM_REGISTER_STATE_CHECK(
- UUT_->v8_(in, kInputStride, out, kOutputStride, filter8, 16, filter8, 16,
- Width(), Height()));
+ UUT_->sv8_(in, kInputStride, out, kOutputStride, filter8, 16, filter8, 16,
+ Width(), Height()));
CheckGuardBlocks();
@@ -633,8 +644,8 @@
DECLARE_ALIGNED(256, const int16_t, filter8[8]) = {0, 0, 0, 128, 0, 0, 0, 0};ASM_REGISTER_STATE_CHECK(
- UUT_->hv8_(in, kInputStride, out, kOutputStride, filter8, 16, filter8, 16,
- Width(), Height()));
+ UUT_->shv8_(in, kInputStride, out, kOutputStride, filter8, 16, filter8,
+ 16, Width(), Height()));
CheckGuardBlocks();
@@ -690,8 +701,6 @@
for (int filter_bank = 0; filter_bank < kNumFilterBanks; ++filter_bank) {const InterpKernel *filters =
vp9_filter_kernels[static_cast<INTERP_FILTER>(filter_bank)];
- const InterpKernel *const eighttap_smooth =
- vp9_filter_kernels[EIGHTTAP_SMOOTH];
for (int filter_x = 0; filter_x < kNumFilters; ++filter_x) { for (int filter_y = 0; filter_y < kNumFilters; ++filter_y) {@@ -700,7 +709,7 @@
ref, kOutputStride,
Width(), Height());
- if (filters == eighttap_smooth || (filter_x && filter_y))
+ if (filter_x && filter_y)
ASM_REGISTER_STATE_CHECK(
UUT_->hv8_(in, kInputStride, out, kOutputStride,
filters[filter_x], 16, filters[filter_y], 16,
@@ -710,11 +719,16 @@
UUT_->v8_(in, kInputStride, out, kOutputStride,
kInvalidFilter, 16, filters[filter_y], 16,
Width(), Height()));
- else
+ else if (filter_x)
ASM_REGISTER_STATE_CHECK(
UUT_->h8_(in, kInputStride, out, kOutputStride,
filters[filter_x], 16, kInvalidFilter, 16,
Width(), Height()));
+ else
+ ASM_REGISTER_STATE_CHECK(
+ UUT_->copy_(in, kInputStride, out, kOutputStride,
+ kInvalidFilter, 0, kInvalidFilter, 0,
+ Width(), Height()));
CheckGuardBlocks();
@@ -769,8 +783,6 @@
for (int filter_bank = 0; filter_bank < kNumFilterBanks; ++filter_bank) {const InterpKernel *filters =
vp9_filter_kernels[static_cast<INTERP_FILTER>(filter_bank)];
- const InterpKernel *const eighttap_smooth =
- vp9_filter_kernels[EIGHTTAP_SMOOTH];
for (int filter_x = 0; filter_x < kNumFilters; ++filter_x) { for (int filter_y = 0; filter_y < kNumFilters; ++filter_y) {@@ -779,7 +791,7 @@
ref, kOutputStride,
Width(), Height());
- if (filters == eighttap_smooth || (filter_x && filter_y))
+ if (filter_x && filter_y)
ASM_REGISTER_STATE_CHECK(
UUT_->hv8_avg_(in, kInputStride, out, kOutputStride,
filters[filter_x], 16, filters[filter_y], 16,
@@ -787,13 +799,18 @@
else if (filter_y)
ASM_REGISTER_STATE_CHECK(
UUT_->v8_avg_(in, kInputStride, out, kOutputStride,
- filters[filter_x], 16, filters[filter_y], 16,
+ kInvalidFilter, 16, filters[filter_y], 16,
Width(), Height()));
- else
+ else if (filter_x)
ASM_REGISTER_STATE_CHECK(
UUT_->h8_avg_(in, kInputStride, out, kOutputStride,
- filters[filter_x], 16, filters[filter_y], 16,
+ filters[filter_x], 16, kInvalidFilter, 16,
Width(), Height()));
+ else
+ ASM_REGISTER_STATE_CHECK(
+ UUT_->avg_(in, kInputStride, out, kOutputStride,
+ kInvalidFilter, 0, kInvalidFilter, 0,
+ Width(), Height()));
CheckGuardBlocks();
@@ -868,8 +885,6 @@
for (int filter_bank = 0; filter_bank < kNumFilterBanks; ++filter_bank) {const InterpKernel *filters =
vp9_filter_kernels[static_cast<INTERP_FILTER>(filter_bank)];
- const InterpKernel *const eighttap_smooth =
- vp9_filter_kernels[EIGHTTAP_SMOOTH];
for (int filter_x = 0; filter_x < kNumFilters; ++filter_x) { for (int filter_y = 0; filter_y < kNumFilters; ++filter_y) {wrapper_filter_block2d_8_c(in, kInputStride,
@@ -876,7 +891,7 @@
filters[filter_x], filters[filter_y],
ref, kOutputStride,
Width(), Height());
- if (filters == eighttap_smooth || (filter_x && filter_y))
+ if (filter_x && filter_y)
ASM_REGISTER_STATE_CHECK(
UUT_->hv8_(in, kInputStride, out, kOutputStride,
filters[filter_x], 16, filters[filter_y], 16,
@@ -886,11 +901,16 @@
UUT_->v8_(in, kInputStride, out, kOutputStride,
kInvalidFilter, 16, filters[filter_y], 16,
Width(), Height()));
- else
+ else if (filter_x)
ASM_REGISTER_STATE_CHECK(
UUT_->h8_(in, kInputStride, out, kOutputStride,
filters[filter_x], 16, kInvalidFilter, 16,
Width(), Height()));
+ else
+ ASM_REGISTER_STATE_CHECK(
+ UUT_->copy_(in, kInputStride, out, kOutputStride,
+ kInvalidFilter, 0, kInvalidFilter, 0,
+ Width(), Height()));
for (int y = 0; y < Height(); ++y)
for (int x = 0; x < Width(); ++x)
@@ -918,10 +938,10 @@
for (int frac = 0; frac < 16; ++frac) { for (int step = 1; step <= 32; ++step) {/* Test the horizontal and vertical filters in combination. */
- ASM_REGISTER_STATE_CHECK(UUT_->hv8_(in, kInputStride, out, kOutputStride,
- eighttap[frac], step,
- eighttap[frac], step,
- Width(), Height()));
+ ASM_REGISTER_STATE_CHECK(UUT_->shv8_(in, kInputStride, out, kOutputStride,
+ eighttap[frac], step,
+ eighttap[frac], step,
+ Width(), Height()));
CheckGuardBlocks();
@@ -1450,6 +1470,9 @@
wrap_convolve_copy_c_8, wrap_convolve_avg_c_8,
wrap_convolve8_horiz_c_8, wrap_convolve8_avg_horiz_c_8,
wrap_convolve8_vert_c_8, wrap_convolve8_avg_vert_c_8,
+ wrap_convolve8_c_8, wrap_convolve8_avg_c_8,
+ wrap_convolve8_horiz_c_8, wrap_convolve8_avg_horiz_c_8,
+ wrap_convolve8_vert_c_8, wrap_convolve8_avg_vert_c_8,
wrap_convolve8_c_8, wrap_convolve8_avg_c_8, 8);
INSTANTIATE_TEST_CASE_P(C_8, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve8_c),
@@ -1469,6 +1492,9 @@
wrap_convolve_copy_c_10, wrap_convolve_avg_c_10,
wrap_convolve8_horiz_c_10, wrap_convolve8_avg_horiz_c_10,
wrap_convolve8_vert_c_10, wrap_convolve8_avg_vert_c_10,
+ wrap_convolve8_c_10, wrap_convolve8_avg_c_10,
+ wrap_convolve8_horiz_c_10, wrap_convolve8_avg_horiz_c_10,
+ wrap_convolve8_vert_c_10, wrap_convolve8_avg_vert_c_10,
wrap_convolve8_c_10, wrap_convolve8_avg_c_10, 10);
INSTANTIATE_TEST_CASE_P(C_10, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve10_c),
@@ -1488,6 +1514,9 @@
wrap_convolve_copy_c_12, wrap_convolve_avg_c_12,
wrap_convolve8_horiz_c_12, wrap_convolve8_avg_horiz_c_12,
wrap_convolve8_vert_c_12, wrap_convolve8_avg_vert_c_12,
+ wrap_convolve8_c_12, wrap_convolve8_avg_c_12,
+ wrap_convolve8_horiz_c_12, wrap_convolve8_avg_horiz_c_12,
+ wrap_convolve8_vert_c_12, wrap_convolve8_avg_vert_c_12,
wrap_convolve8_c_12, wrap_convolve8_avg_c_12, 12);
INSTANTIATE_TEST_CASE_P(C_12, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve12_c),
@@ -1510,7 +1539,10 @@
vpx_convolve_copy_c, vpx_convolve_avg_c,
vpx_convolve8_horiz_c, vpx_convolve8_avg_horiz_c,
vpx_convolve8_vert_c, vpx_convolve8_avg_vert_c,
- vpx_convolve8_c, vpx_convolve8_avg_c, 0);
+ vpx_convolve8_c, vpx_convolve8_avg_c,
+ vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
+ vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
+ vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
INSTANTIATE_TEST_CASE_P(C, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve8_c),
@@ -1534,16 +1566,25 @@
wrap_convolve_copy_c_8, wrap_convolve_avg_c_8,
wrap_convolve8_horiz_sse2_8, wrap_convolve8_avg_horiz_sse2_8,
wrap_convolve8_vert_sse2_8, wrap_convolve8_avg_vert_sse2_8,
+ wrap_convolve8_sse2_8, wrap_convolve8_avg_sse2_8,
+ wrap_convolve8_horiz_sse2_8, wrap_convolve8_avg_horiz_sse2_8,
+ wrap_convolve8_vert_sse2_8, wrap_convolve8_avg_vert_sse2_8,
wrap_convolve8_sse2_8, wrap_convolve8_avg_sse2_8, 8);
const ConvolveFunctions convolve10_sse2(
wrap_convolve_copy_c_10, wrap_convolve_avg_c_10,
wrap_convolve8_horiz_sse2_10, wrap_convolve8_avg_horiz_sse2_10,
wrap_convolve8_vert_sse2_10, wrap_convolve8_avg_vert_sse2_10,
+ wrap_convolve8_sse2_10, wrap_convolve8_avg_sse2_10,
+ wrap_convolve8_horiz_sse2_10, wrap_convolve8_avg_horiz_sse2_10,
+ wrap_convolve8_vert_sse2_10, wrap_convolve8_avg_vert_sse2_10,
wrap_convolve8_sse2_10, wrap_convolve8_avg_sse2_10, 10);
const ConvolveFunctions convolve12_sse2(
wrap_convolve_copy_c_12, wrap_convolve_avg_c_12,
wrap_convolve8_horiz_sse2_12, wrap_convolve8_avg_horiz_sse2_12,
wrap_convolve8_vert_sse2_12, wrap_convolve8_avg_vert_sse2_12,
+ wrap_convolve8_sse2_12, wrap_convolve8_avg_sse2_12,
+ wrap_convolve8_horiz_sse2_12, wrap_convolve8_avg_horiz_sse2_12,
+ wrap_convolve8_vert_sse2_12, wrap_convolve8_avg_vert_sse2_12,
wrap_convolve8_sse2_12, wrap_convolve8_avg_sse2_12, 12);
INSTANTIATE_TEST_CASE_P(SSE2, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve8_sse2),
@@ -1594,7 +1635,10 @@
#endif // CONFIG_USE_X86INC
vpx_convolve8_horiz_sse2, vpx_convolve8_avg_horiz_sse2,
vpx_convolve8_vert_sse2, vpx_convolve8_avg_vert_sse2,
- vpx_convolve8_sse2, vpx_convolve8_avg_sse2, 0);
+ vpx_convolve8_sse2, vpx_convolve8_avg_sse2,
+ vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
+ vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
+ vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
INSTANTIATE_TEST_CASE_P(SSE2, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve8_sse2),
@@ -1618,7 +1662,10 @@
vpx_convolve_copy_c, vpx_convolve_avg_c,
vpx_convolve8_horiz_ssse3, vpx_convolve8_avg_horiz_ssse3,
vpx_convolve8_vert_ssse3, vpx_convolve8_avg_vert_ssse3,
- vpx_convolve8_ssse3, vpx_convolve8_avg_ssse3, 0);
+ vpx_convolve8_ssse3, vpx_convolve8_avg_ssse3,
+ vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
+ vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
+ vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
INSTANTIATE_TEST_CASE_P(SSSE3, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve8_ssse3),
@@ -1641,7 +1688,10 @@
vpx_convolve_copy_c, vpx_convolve_avg_c,
vpx_convolve8_horiz_avx2, vpx_convolve8_avg_horiz_ssse3,
vpx_convolve8_vert_avx2, vpx_convolve8_avg_vert_ssse3,
- vpx_convolve8_avx2, vpx_convolve8_avg_ssse3, 0);
+ vpx_convolve8_avx2, vpx_convolve8_avg_ssse3,
+ vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
+ vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
+ vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
INSTANTIATE_TEST_CASE_P(AVX2, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve8_avx2),
@@ -1665,13 +1715,19 @@
vpx_convolve_copy_neon, vpx_convolve_avg_neon,
vpx_convolve8_horiz_neon, vpx_convolve8_avg_horiz_neon,
vpx_convolve8_vert_neon, vpx_convolve8_avg_vert_neon,
- vpx_convolve8_neon, vpx_convolve8_avg_neon, 0);
+ vpx_convolve8_neon, vpx_convolve8_avg_neon,
+ vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
+ vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
+ vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
#else // HAVE_NEON
const ConvolveFunctions convolve8_neon(
vpx_convolve_copy_neon, vpx_convolve_avg_neon,
vpx_convolve8_horiz_neon, vpx_convolve8_avg_horiz_neon,
vpx_convolve8_vert_neon, vpx_convolve8_avg_vert_neon,
- vpx_convolve8_neon, vpx_convolve8_avg_neon, 0);
+ vpx_convolve8_neon, vpx_convolve8_avg_neon,
+ vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
+ vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
+ vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
#endif // HAVE_NEON_ASM
INSTANTIATE_TEST_CASE_P(NEON, ConvolveTest, ::testing::Values(
@@ -1695,7 +1751,10 @@
vpx_convolve_copy_dspr2, vpx_convolve_avg_dspr2,
vpx_convolve8_horiz_dspr2, vpx_convolve8_avg_horiz_dspr2,
vpx_convolve8_vert_dspr2, vpx_convolve8_avg_vert_dspr2,
- vpx_convolve8_dspr2, vpx_convolve8_avg_dspr2, 0);
+ vpx_convolve8_dspr2, vpx_convolve8_avg_dspr2,
+ vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
+ vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
+ vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
INSTANTIATE_TEST_CASE_P(DSPR2, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve8_dspr2),
@@ -1718,7 +1777,10 @@
vpx_convolve_copy_msa, vpx_convolve_avg_msa,
vpx_convolve8_horiz_msa, vpx_convolve8_avg_horiz_msa,
vpx_convolve8_vert_msa, vpx_convolve8_avg_vert_msa,
- vpx_convolve8_msa, vpx_convolve8_avg_msa, 0);
+ vpx_convolve8_msa, vpx_convolve8_avg_msa,
+ vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
+ vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
+ vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
INSTANTIATE_TEST_CASE_P(MSA, ConvolveTest, ::testing::Values(
make_tuple(4, 4, &convolve8_msa),
--- a/vp9/encoder/vp9_encoder.c
+++ b/vp9/encoder/vp9_encoder.c
@@ -2586,13 +2586,13 @@
kernel[y_q4 & 0xf], 16 * src_h / dst_h,
16 / factor, 16 / factor, bd);
} else {- vpx_convolve8(src_ptr, src_stride, dst_ptr, dst_stride,
+ vpx_scaled_2d(src_ptr, src_stride, dst_ptr, dst_stride,
kernel[x_q4 & 0xf], 16 * src_w / dst_w,
kernel[y_q4 & 0xf], 16 * src_h / dst_h,
16 / factor, 16 / factor);
}
#else
- vpx_convolve8(src_ptr, src_stride, dst_ptr, dst_stride,
+ vpx_scaled_2d(src_ptr, src_stride, dst_ptr, dst_stride,
kernel[x_q4 & 0xf], 16 * src_w / dst_w,
kernel[y_q4 & 0xf], 16 * src_h / dst_h,
16 / factor, 16 / factor);
--- a/vpx_dsp/x86/convolve.h
+++ b/vpx_dsp/x86/convolve.h
@@ -31,6 +31,7 @@
const int16_t *filter_x, int x_step_q4, \
const int16_t *filter_y, int y_step_q4, \
int w, int h) { \+ assert(filter[3] != 128); \
if (step_q4 == 16 && filter[3] != 128) { \ if (filter[0] || filter[1] || filter[2]) { \ while (w >= 16) { \@@ -102,11 +103,6 @@
} \
} \
} \
- if (w) { \- vpx_scaled_##name(src, src_stride, dst, dst_stride, \
- filter_x, x_step_q4, filter_y, y_step_q4, \
- w, h); \
- } \
}
#define FUN_CONV_2D(avg, opt) \
@@ -115,6 +111,8 @@
const int16_t *filter_x, int x_step_q4, \
const int16_t *filter_y, int y_step_q4, \
int w, int h) { \+ assert(filter_x[3] != 128); \
+ assert(filter_y[3] != 128); \
assert(w <= 64); \
assert(h <= 64); \
if (x_step_q4 == 16 && y_step_q4 == 16) { \@@ -136,9 +134,6 @@
filter_x, x_step_q4, filter_y, \
y_step_q4, w, h); \
} \
- } else { \- vpx_scaled_##avg##2d(src, src_stride, dst, dst_stride, \
- filter_x, x_step_q4, filter_y, y_step_q4, w, h); \
} \
}