Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(232)

Issues Search
    My Issues | Starred     Open | Closed | All

Unified Diff: source/libvpx/vpx_dsp/mips/vpx_convolve8_avg_vert_msa.c

Issue 1302353004: libvpx: Pull from upstream (Closed) Base URL: https://chromium.googlesource.com/chromium/deps/libvpx.git@master
Patch Set: Created 5 years, 4 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « source/libvpx/vpx_dsp/mips/vpx_convolve8_avg_msa.c ('k') | source/libvpx/vpx_dsp/mips/vpx_convolve8_dspr2.c » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: source/libvpx/vpx_dsp/mips/vpx_convolve8_avg_vert_msa.c
diff --git a/source/libvpx/vpx_dsp/mips/vpx_convolve8_avg_vert_msa.c b/source/libvpx/vpx_dsp/mips/vpx_convolve8_avg_vert_msa.c
index 4d4b3d9e353864f9e897ce681fcb7639e71d69a1..0164e41aa161a3c5614b86c2ac71687089a29083 100644
--- a/source/libvpx/vpx_dsp/mips/vpx_convolve8_avg_vert_msa.c
+++ b/source/libvpx/vpx_dsp/mips/vpx_convolve8_avg_vert_msa.c
@@ -8,6 +8,7 @@
* be found in the AUTHORS file in the root of the source tree.
*/
+#include <assert.h>
#include "./vpx_dsp_rtcd.h"
#include "vpx_dsp/mips/vpx_convolve_msa.h"
@@ -283,7 +284,6 @@ static void common_vt_2t_and_aver_dst_4x4_msa(const uint8_t *src,
ILVR_D2_UB(src21_r, src10_r, src43_r, src32_r, src2110, src4332);
DOTP_UB2_UH(src2110, src4332, filt0, filt0, tmp0, tmp1);
SRARI_H2_UH(tmp0, tmp1, FILTER_BITS);
- SAT_UH2_UH(tmp0, tmp1, 7);
out = (v16u8)__msa_pckev_b((v16i8)tmp1, (v16i8)tmp0);
out = __msa_aver_u_b(out, dst0);
@@ -323,7 +323,6 @@ static void common_vt_2t_and_aver_dst_4x8_msa(const uint8_t *src,
DOTP_UB4_UH(src2110, src4332, src6554, src8776, filt0, filt0, filt0, filt0,
tmp0, tmp1, tmp2, tmp3);
SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, FILTER_BITS);
- SAT_UH4_UH(tmp0, tmp1, tmp2, tmp3, 7);
PCKEV_B2_UB(tmp1, tmp0, tmp3, tmp2, src2110, src4332);
AVER_UB2_UB(src2110, dst0, src4332, dst1, src2110, src4332);
ST4x4_UB(src2110, src2110, 0, 1, 2, 3, dst, dst_stride);
@@ -365,7 +364,6 @@ static void common_vt_2t_and_aver_dst_8x4_msa(const uint8_t *src,
DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, tmp0, tmp1,
tmp2, tmp3);
SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, FILTER_BITS);
- SAT_UH4_UH(tmp0, tmp1, tmp2, tmp3, 7);
PCKEV_AVG_ST8x4_UB(tmp0, dst0, tmp1, dst1, tmp2, dst2, tmp3, dst3,
dst, dst_stride);
}
@@ -402,7 +400,6 @@ static void common_vt_2t_and_aver_dst_8x8mult_msa(const uint8_t *src,
DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, tmp0, tmp1,
tmp2, tmp3);
SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, FILTER_BITS);
- SAT_UH4_UH(tmp0, tmp1, tmp2, tmp3, 7);
PCKEV_AVG_ST8x4_UB(tmp0, dst1, tmp1, dst2, tmp2, dst3, tmp3, dst4,
dst, dst_stride);
dst += (4 * dst_stride);
@@ -410,7 +407,6 @@ static void common_vt_2t_and_aver_dst_8x8mult_msa(const uint8_t *src,
DOTP_UB4_UH(vec4, vec5, vec6, vec7, filt0, filt0, filt0, filt0, tmp0, tmp1,
tmp2, tmp3);
SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, FILTER_BITS);
- SAT_UH4_UH(tmp0, tmp1, tmp2, tmp3, 7);
PCKEV_AVG_ST8x4_UB(tmp0, dst5, tmp1, dst6, tmp2, dst7, tmp3, dst8,
dst, dst_stride);
dst += (4 * dst_stride);
@@ -460,7 +456,6 @@ static void common_vt_2t_and_aver_dst_16w_msa(const uint8_t *src,
ILVL_B2_UB(src1, src0, src2, src1, vec1, vec3);
DOTP_UB2_UH(vec0, vec1, filt0, filt0, tmp0, tmp1);
SRARI_H2_UH(tmp0, tmp1, FILTER_BITS);
- SAT_UH2_UH(tmp0, tmp1, 7);
PCKEV_AVG_ST_UB(tmp1, tmp0, dst0, dst);
dst += dst_stride;
@@ -468,19 +463,16 @@ static void common_vt_2t_and_aver_dst_16w_msa(const uint8_t *src,
ILVL_B2_UB(src3, src2, src4, src3, vec5, vec7);
DOTP_UB2_UH(vec2, vec3, filt0, filt0, tmp2, tmp3);
SRARI_H2_UH(tmp2, tmp3, FILTER_BITS);
- SAT_UH2_UH(tmp2, tmp3, 7);
PCKEV_AVG_ST_UB(tmp3, tmp2, dst1, dst);
dst += dst_stride;
DOTP_UB2_UH(vec4, vec5, filt0, filt0, tmp0, tmp1);
SRARI_H2_UH(tmp0, tmp1, FILTER_BITS);
- SAT_UH2_UH(tmp0, tmp1, 7);
PCKEV_AVG_ST_UB(tmp1, tmp0, dst2, dst);
dst += dst_stride;
DOTP_UB2_UH(vec6, vec7, filt0, filt0, tmp2, tmp3);
SRARI_H2_UH(tmp2, tmp3, FILTER_BITS);
- SAT_UH2_UH(tmp2, tmp3, 7);
PCKEV_AVG_ST_UB(tmp3, tmp2, dst3, dst);
dst += dst_stride;
@@ -519,48 +511,40 @@ static void common_vt_2t_and_aver_dst_32w_msa(const uint8_t *src,
DOTP_UB2_UH(vec0, vec1, filt0, filt0, tmp0, tmp1);
SRARI_H2_UH(tmp0, tmp1, FILTER_BITS);
- SAT_UH2_UH(tmp0, tmp1, 7);
PCKEV_AVG_ST_UB(tmp1, tmp0, dst0, dst);
DOTP_UB2_UH(vec2, vec3, filt0, filt0, tmp2, tmp3);
SRARI_H2_UH(tmp2, tmp3, FILTER_BITS);
- SAT_UH2_UH(tmp2, tmp3, 7);
PCKEV_AVG_ST_UB(tmp3, tmp2, dst1, dst + dst_stride);
ILVR_B2_UB(src3, src2, src4, src3, vec4, vec6);
ILVL_B2_UB(src3, src2, src4, src3, vec5, vec7);
DOTP_UB2_UH(vec4, vec5, filt0, filt0, tmp0, tmp1);
SRARI_H2_UH(tmp0, tmp1, FILTER_BITS);
- SAT_UH2_UH(tmp0, tmp1, 7);
PCKEV_AVG_ST_UB(tmp1, tmp0, dst2, dst + 2 * dst_stride);
DOTP_UB2_UH(vec6, vec7, filt0, filt0, tmp2, tmp3);
SRARI_H2_UH(tmp2, tmp3, FILTER_BITS);
- SAT_UH2_UH(tmp2, tmp3, 7);
PCKEV_AVG_ST_UB(tmp3, tmp2, dst3, dst + 3 * dst_stride);
ILVR_B2_UB(src6, src5, src7, src6, vec0, vec2);
ILVL_B2_UB(src6, src5, src7, src6, vec1, vec3);
DOTP_UB2_UH(vec0, vec1, filt0, filt0, tmp0, tmp1);
SRARI_H2_UH(tmp0, tmp1, FILTER_BITS);
- SAT_UH2_UH(tmp0, tmp1, 7);
PCKEV_AVG_ST_UB(tmp1, tmp0, dst4, dst + 16);
DOTP_UB2_UH(vec2, vec3, filt0, filt0, tmp2, tmp3);
SRARI_H2_UH(tmp2, tmp3, FILTER_BITS);
- SAT_UH2_UH(tmp2, tmp3, 7);
PCKEV_AVG_ST_UB(tmp3, tmp2, dst5, dst + 16 + dst_stride);
ILVR_B2_UB(src8, src7, src9, src8, vec4, vec6);
ILVL_B2_UB(src8, src7, src9, src8, vec5, vec7);
DOTP_UB2_UH(vec4, vec5, filt0, filt0, tmp0, tmp1);
SRARI_H2_UH(tmp0, tmp1, FILTER_BITS);
- SAT_UH2_UH(tmp0, tmp1, 7);
PCKEV_AVG_ST_UB(tmp1, tmp0, dst6, dst + 16 + 2 * dst_stride);
DOTP_UB2_UH(vec6, vec7, filt0, filt0, tmp2, tmp3);
SRARI_H2_UH(tmp2, tmp3, FILTER_BITS);
- SAT_UH2_UH(tmp2, tmp3, 7);
PCKEV_AVG_ST_UB(tmp3, tmp2, dst7, dst + 16 + 3 * dst_stride);
dst += (4 * dst_stride);
@@ -605,48 +589,40 @@ static void common_vt_2t_and_aver_dst_64w_msa(const uint8_t *src,
ILVL_B2_UB(src1, src0, src2, src1, vec1, vec3);
DOTP_UB2_UH(vec0, vec1, filt0, filt0, tmp0, tmp1);
SRARI_H2_UH(tmp0, tmp1, FILTER_BITS);
- SAT_UH2_UH(tmp0, tmp1, 7);
PCKEV_AVG_ST_UB(tmp1, tmp0, dst0, dst);
DOTP_UB2_UH(vec2, vec3, filt0, filt0, tmp2, tmp3);
SRARI_H2_UH(tmp2, tmp3, FILTER_BITS);
- SAT_UH2_UH(tmp2, tmp3, 7);
PCKEV_AVG_ST_UB(tmp3, tmp2, dst1, dst + dst_stride);
ILVR_B2_UB(src4, src3, src5, src4, vec4, vec6);
ILVL_B2_UB(src4, src3, src5, src4, vec5, vec7);
DOTP_UB2_UH(vec4, vec5, filt0, filt0, tmp4, tmp5);
SRARI_H2_UH(tmp4, tmp5, FILTER_BITS);
- SAT_UH2_UH(tmp4, tmp5, 7);
PCKEV_AVG_ST_UB(tmp5, tmp4, dst2, dst + 16);
DOTP_UB2_UH(vec6, vec7, filt0, filt0, tmp6, tmp7);
SRARI_H2_UH(tmp6, tmp7, FILTER_BITS);
- SAT_UH2_UH(tmp6, tmp7, 7);
PCKEV_AVG_ST_UB(tmp7, tmp6, dst3, dst + 16 + dst_stride);
ILVR_B2_UB(src7, src6, src8, src7, vec0, vec2);
ILVL_B2_UB(src7, src6, src8, src7, vec1, vec3);
DOTP_UB2_UH(vec0, vec1, filt0, filt0, tmp0, tmp1);
SRARI_H2_UH(tmp0, tmp1, FILTER_BITS);
- SAT_UH2_UH(tmp0, tmp1, 7);
PCKEV_AVG_ST_UB(tmp1, tmp0, dst4, dst + 32);
DOTP_UB2_UH(vec2, vec3, filt0, filt0, tmp2, tmp3);
SRARI_H2_UH(tmp2, tmp3, FILTER_BITS);
- SAT_UH2_UH(tmp2, tmp3, 7);
PCKEV_AVG_ST_UB(tmp3, tmp2, dst5, dst + 32 + dst_stride);
ILVR_B2_UB(src10, src9, src11, src10, vec4, vec6);
ILVL_B2_UB(src10, src9, src11, src10, vec5, vec7);
DOTP_UB2_UH(vec4, vec5, filt0, filt0, tmp4, tmp5);
SRARI_H2_UH(tmp4, tmp5, FILTER_BITS);
- SAT_UH2_UH(tmp4, tmp5, 7);
PCKEV_AVG_ST_UB(tmp5, tmp4, dst6, (dst + 48));
DOTP_UB2_UH(vec6, vec7, filt0, filt0, tmp6, tmp7);
SRARI_H2_UH(tmp6, tmp7, FILTER_BITS);
- SAT_UH2_UH(tmp6, tmp7, 7);
PCKEV_AVG_ST_UB(tmp7, tmp6, dst7, dst + 48 + dst_stride);
dst += (2 * dst_stride);
@@ -664,19 +640,8 @@ void vpx_convolve8_avg_vert_msa(const uint8_t *src, ptrdiff_t src_stride,
int w, int h) {
int8_t cnt, filt_ver[8];
- if (16 != y_step_q4) {
- vpx_convolve8_avg_vert_c(src, src_stride, dst, dst_stride,
- filter_x, x_step_q4, filter_y, y_step_q4,
- w, h);
- return;
- }
-
- if (((const int32_t *)filter_y)[1] == 0x800000) {
- vpx_convolve_avg(src, src_stride, dst, dst_stride,
- filter_x, x_step_q4, filter_y, y_step_q4,
- w, h);
- return;
- }
+ assert(y_step_q4 == 16);
+ assert(((const int32_t *)filter_y)[1] != 0x800000);
for (cnt = 0; cnt < 8; ++cnt) {
filt_ver[cnt] = filter_y[cnt];
« no previous file with comments | « source/libvpx/vpx_dsp/mips/vpx_convolve8_avg_msa.c ('k') | source/libvpx/vpx_dsp/mips/vpx_convolve8_dspr2.c » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698