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/vp9_idct_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/vp9_idct_msa.c')
| -rw-r--r-- | libavcodec/mips/vp9_idct_msa.c | 10 |
1 files changed, 4 insertions, 6 deletions
diff --git a/libavcodec/mips/vp9_idct_msa.c b/libavcodec/mips/vp9_idct_msa.c index 1f32770139..a3f5270b73 100644 --- a/libavcodec/mips/vp9_idct_msa.c +++ b/libavcodec/mips/vp9_idct_msa.c @@ -763,13 +763,13 @@ static void vp9_iadst8x8_colcol_addblk_msa(int16_t *input, uint8_t *dst, res0 = (v8i16) __msa_ilvr_b((v16i8) zero, (v16i8) dst0); res0 += out0; - res0 = CLIP_SH_0_255(res0); + CLIP_SH_0_255(res0); res0 = (v8i16) __msa_pckev_b((v16i8) res0, (v16i8) res0); ST_D1(res0, 0, dst); res7 = (v8i16) __msa_ilvr_b((v16i8) zero, (v16i8) dst7); res7 += out7; - res7 = CLIP_SH_0_255(res7); + CLIP_SH_0_255(res7); res7 = (v8i16) __msa_pckev_b((v16i8) res7, (v16i8) res7); ST_D1(res7, 0, dst + 7 * dst_stride); @@ -1192,8 +1192,7 @@ static void vp9_idct16x16_1_add_msa(int16_t *input, uint8_t *dst, res3); ADD4(res4, vec, res5, vec, res6, vec, res7, vec, 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_UB(res4, res0, res5, res1, res6, res2, res7, res3, tmp0, tmp1, tmp2, tmp3); ST_UB4(tmp0, tmp1, tmp2, tmp3, dst, dst_stride); @@ -1981,8 +1980,7 @@ static void vp9_idct32x32_1_add_msa(int16_t *input, uint8_t *dst, res3); ADD4(res4, vec, res5, vec, res6, vec, res7, vec, 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_UB(res4, res0, res5, res1, res6, res2, res7, res3, tmp0, tmp1, tmp2, tmp3); |