diff options
| author | gxw <guxiwei-hf@loongson.cn> | 2019-08-07 17:52:00 +0800 |
|---|---|---|
| committer | Michael Niedermayer <michael@niedermayer.cc> | 2019-08-13 16:48:38 +0200 |
| commit | a3e572d96fd1dd6291f6b28e173db858c08ff8d8 (patch) | |
| tree | 85807f6ec1442cc362cf8946e67f564c92267e07 /libavcodec/mips/h264idct_msa.c | |
| parent | 8f92eb05e063e6c4d6e36521020620d4e6e1c21d (diff) | |
| download | ffmpeg-a3e572d96fd1dd6291f6b28e173db858c08ff8d8.tar.gz | |
avutil/mips: refine msa macros CLIP_*.
Changing details as following:
1. Remove the local variable 'out_m' in 'CLIP_SH' and store the result in
source vector.
2. Refine the implementation of macro 'CLIP_SH_0_255' and 'CLIP_SW_0_255'.
Performance of VP8 decoding has speed up about 1.1%(from 7.03x to 7.11x).
Performance of H264 decoding has speed up about 0.5%(from 4.35x to 4.37x).
Performance of Theora decoding has speed up about 0.7%(from 5.79x to 5.83x).
3. Remove redundant macro 'CLIP_SH/Wn_0_255_MAX_SATU' and use 'CLIP_SH/Wn_0_255'
instead, because there are no difference in the effect of this two macros.
Reviewed-by: Shiyou Yin <yinshiyou-hf@loongson.cn>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Diffstat (limited to 'libavcodec/mips/h264idct_msa.c')
| -rw-r--r-- | libavcodec/mips/h264idct_msa.c | 7 |
1 files changed, 3 insertions, 4 deletions
diff --git a/libavcodec/mips/h264idct_msa.c b/libavcodec/mips/h264idct_msa.c index 7851bfdf4b..fbf7795e27 100644 --- a/libavcodec/mips/h264idct_msa.c +++ b/libavcodec/mips/h264idct_msa.c @@ -233,8 +233,7 @@ static void avc_idct8_addblk_msa(uint8_t *dst, int16_t *src, int32_t dst_stride) res0, res1, res2, res3); ADD4(res4, tmp4, res5, tmp5, res6, tmp6, res7, tmp7, res4, res5, res6, res7); - CLIP_SH4_0_255(res0, res1, res2, res3); - CLIP_SH4_0_255(res4, res5, res6, res7); + CLIP_SH8_0_255(res0, res1, res2, res3, res4, res5, res6, res7); PCKEV_B4_SB(res1, res0, res3, res2, res5, res4, res7, res6, dst0, dst1, dst2, dst3); ST_D8(dst0, dst1, dst2, dst3, 0, 1, 0, 1, 0, 1, 0, 1, dst, dst_stride) @@ -263,8 +262,8 @@ static void avc_idct8_dc_addblk_msa(uint8_t *dst, int16_t *src, dst0_r, dst1_r, dst2_r, dst3_r); ADD4(dst4_r, dc, dst5_r, dc, dst6_r, dc, dst7_r, dc, dst4_r, dst5_r, dst6_r, dst7_r); - CLIP_SH4_0_255(dst0_r, dst1_r, dst2_r, dst3_r); - CLIP_SH4_0_255(dst4_r, dst5_r, dst6_r, dst7_r); + CLIP_SH8_0_255(dst0_r, dst1_r, dst2_r, dst3_r, + dst4_r, dst5_r, dst6_r, dst7_r); PCKEV_B4_SB(dst1_r, dst0_r, dst3_r, dst2_r, dst5_r, dst4_r, dst7_r, dst6_r, dst0, dst1, dst2, dst3); ST_D8(dst0, dst1, dst2, dst3, 0, 1, 0, 1, 0, 1, 0, 1, dst, dst_stride) |