[FFmpeg-devel] avcodec/mips: Improve hevc uni weighted hv mc msa functions

Message ID	1509454841-9470-1-git-send-email-kaustubh.raste@imgtec.com
State	Accepted
Commit	48d77d5cd4ac0858e1e9abf387f739a5c773200f
Headers	show Delivered-To: ffmpegpatchwork@gmail.com Received-SPF: pass (google.com: domain of ffmpeg-devel-bounces@ffmpeg.org designates 79.124.17.100 as permitted sender) client-ip=79.124.17.100; From: <kaustubh.raste@imgtec.com> To: <ffmpeg-devel@ffmpeg.org> Date: Tue, 31 Oct 2017 18:30:40 +0530 Message-ID: <1509454841-9470-1-git-send-email-kaustubh.raste@imgtec.com> MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH] avcodec/mips: Improve hevc uni weighted hv mc msa functions Precedence: list Reply-To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org> Cc: Kaustubh Raste <kaustubh.raste@imgtec.com> Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: base64 Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" <ffmpeg-devel-bounces@ffmpeg.org>

diff --git a/libavcodec/mips/hevc_macros_msa.h b/libavcodec/mips/hevc_macros_msa.h index 7dcfea0..27c69ff 100644 --- a/libavcodec/mips/hevc_macros_msa.h +++ b/libavcodec/mips/hevc_macros_msa.h @@ -80,6 +80,15 @@ out_m; \ } ) +#define HEVC_FILT_4TAP_SH(in0, in1, filt0, filt1) \ +( { \ + v8i16 out_m; \ + \ + out_m = __msa_dotp_s_h((v16i8) in0, (v16i8) filt0); \ + out_m = __msa_dpadd_s_h(out_m, (v16i8) in1, (v16i8) filt1); \ + out_m; \ +} ) + #define HEVC_FILT_4TAP(in0, in1, filt0, filt1) \ ( { \ v4i32 out_m; \ diff --git a/libavcodec/mips/hevc_mc_uniw_msa.c b/libavcodec/mips/hevc_mc_uniw_msa.c index 28c7062f..0796b0a 100644 --- a/libavcodec/mips/hevc_mc_uniw_msa.c +++ b/libavcodec/mips/hevc_mc_uniw_msa.c @@ -1801,40 +1801,42 @@ static void hevc_hv_uniwgt_8t_4w_msa(uint8_t *src, int32_t rnd_val) { uint32_t loop_cnt; - v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8; + v16u8 out; + v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8, src9, src10; v8i16 filt0, filt1, filt2, filt3; - v4i32 filt_h0, filt_h1, filt_h2, filt_h3; + v8i16 filt_h0, filt_h1, filt_h2, filt_h3; v16i8 mask1, mask2, mask3; - v8i16 filter_vec, const_vec; + v8i16 filter_vec; v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7; v16i8 vec8, vec9, vec10, vec11, vec12, vec13, vec14, vec15; - v8i16 dst30, dst41, dst52, dst63, dst66, dst87; - v4i32 dst0_r, dst1_r, weight_vec, offset_vec, rnd_vec; - v8i16 dst10_r, dst32_r, dst54_r, dst76_r; - v8i16 dst21_r, dst43_r, dst65_r, dst87_r; - v16i8 mask0 = { 0, 1, 1, 2, 2, 3, 3, 4, 16, 17, 17, 18, 18, 19, 19, 20 }; - v8u16 mask4 = { 0, 4, 1, 5, 2, 6, 3, 7 }; + v8i16 dst30, dst41, dst52, dst63, dst66, dst97, dst108; + v8i16 dst10_r, dst32_r, dst54_r, dst76_r, dst98_r; + v8i16 dst21_r, dst43_r, dst65_r, dst87_r, dst109_r; + v4i32 dst0_r, dst1_r, dst2_r, dst3_r; + v4i32 weight_vec, offset_vec, rnd_vec, const_128, denom_vec; + v16i8 mask0 = LD_SB(ff_hevc_mask_arr + 16); src -= ((3 * src_stride) + 3); filter_vec = LD_SH(filter_x); SPLATI_H4_SH(filter_vec, 0, 1, 2, 3, filt0, filt1, filt2, filt3); filter_vec = LD_SH(filter_y); - vec0 = __msa_clti_s_b((v16i8) filter_vec, 0); - filter_vec = (v8i16) __msa_ilvr_b(vec0, (v16i8) filter_vec); + UNPCK_R_SB_SH(filter_vec, filter_vec); - SPLATI_W4_SW(filter_vec, filt_h0, filt_h1, filt_h2, filt_h3); + SPLATI_W4_SH(filter_vec, filt_h0, filt_h1, filt_h2, filt_h3); mask1 = mask0 + 2; mask2 = mask0 + 4; mask3 = mask0 + 6; - const_vec = __msa_ldi_h(128); - const_vec <<= 6; - weight_vec = __msa_fill_w(weight); offset_vec = __msa_fill_w(offset); rnd_vec = __msa_fill_w(rnd_val); + denom_vec = rnd_vec - 6; + + const_128 = __msa_ldi_w(128); + const_128 *= weight_vec; + offset_vec += __msa_srar_w(const_128, denom_vec); LD_SB7(src, src_stride, src0, src1, src2, src3, src4, src5, src6); src += (7 * src_stride); @@ -1847,64 +1849,68 @@ static void hevc_hv_uniwgt_8t_4w_msa(uint8_t *src, vec8, vec9, vec10, vec11); VSHF_B4_SB(src3, src6, mask0, mask1, mask2, mask3, vec12, vec13, vec14, vec15); - dst30 = const_vec; - DPADD_SB4_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, filt3, - dst30, dst30, dst30, dst30); - dst41 = const_vec; - DPADD_SB4_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, filt3, - dst41, dst41, dst41, dst41); - dst52 = const_vec; - DPADD_SB4_SH(vec8, vec9, vec10, vec11, filt0, filt1, filt2, filt3, - dst52, dst52, dst52, dst52); - dst63 = const_vec; - DPADD_SB4_SH(vec12, vec13, vec14, vec15, filt0, filt1, filt2, filt3, - dst63, dst63, dst63, dst63); - - ILVR_H3_SH(dst41, dst30, dst52, dst41, dst63, dst52, - dst10_r, dst21_r, dst32_r); - - dst43_r = __msa_ilvl_h(dst41, dst30); - dst54_r = __msa_ilvl_h(dst52, dst41); - dst65_r = __msa_ilvl_h(dst63, dst52); + dst30 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, + filt3); + dst41 = HEVC_FILT_8TAP_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, + filt3); + dst52 = HEVC_FILT_8TAP_SH(vec8, vec9, vec10, vec11, filt0, filt1, filt2, + filt3); + dst63 = HEVC_FILT_8TAP_SH(vec12, vec13, vec14, vec15, filt0, filt1, filt2, + filt3); + + ILVRL_H2_SH(dst41, dst30, dst10_r, dst43_r); + ILVRL_H2_SH(dst52, dst41, dst21_r, dst54_r); + ILVRL_H2_SH(dst63, dst52, dst32_r, dst65_r); dst66 = (v8i16) __msa_splati_d((v2i64) dst63, 1); - for (loop_cnt = height >> 1; loop_cnt--;) { - LD_SB2(src, src_stride, src7, src8); - src += (2 * src_stride); - XORI_B2_128_SB(src7, src8); + for (loop_cnt = height >> 2; loop_cnt--;) { + LD_SB4(src, src_stride, src7, src8, src9, src10); + src += (4 * src_stride); + XORI_B4_128_SB(src7, src8, src9, src10); - VSHF_B4_SB(src7, src8, mask0, mask1, mask2, mask3, + VSHF_B4_SB(src7, src9, mask0, mask1, mask2, mask3, vec0, vec1, vec2, vec3); - dst87 = const_vec; - DPADD_SB4_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, filt3, - dst87, dst87, dst87, dst87); - dst76_r = __msa_ilvr_h(dst87, dst66); - dst0_r = HEVC_FILT_8TAP(dst10_r, dst32_r, dst54_r, dst76_r, - filt_h0, filt_h1, filt_h2, filt_h3); - dst87_r = __msa_vshf_h((v8i16) mask4, dst87, dst87); - dst1_r = HEVC_FILT_8TAP(dst21_r, dst43_r, dst65_r, dst87_r, - filt_h0, filt_h1, filt_h2, filt_h3); - - dst0_r >>= 6; - dst1_r >>= 6; + VSHF_B4_SB(src8, src10, mask0, mask1, mask2, mask3, + vec4, vec5, vec6, vec7); + dst97 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, + filt3); + dst108 = HEVC_FILT_8TAP_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, + filt3); + + dst76_r = __msa_ilvr_h(dst97, dst66); + ILVRL_H2_SH(dst108, dst97, dst87_r, dst109_r); + dst66 = (v8i16) __msa_splati_d((v2i64) dst97, 1); + dst98_r = __msa_ilvr_h(dst66, dst108); + + dst0_r = HEVC_FILT_8TAP(dst10_r, dst32_r, dst54_r, dst76_r, filt_h0, + filt_h1, filt_h2, filt_h3); + dst1_r = HEVC_FILT_8TAP(dst21_r, dst43_r, dst65_r, dst87_r, filt_h0, + filt_h1, filt_h2, filt_h3); + dst2_r = HEVC_FILT_8TAP(dst32_r, dst54_r, dst76_r, dst98_r, filt_h0, + filt_h1, filt_h2, filt_h3); + dst3_r = HEVC_FILT_8TAP(dst43_r, dst65_r, dst87_r, dst109_r, filt_h0, + filt_h1, filt_h2, filt_h3); + + SRA_4V(dst0_r, dst1_r, dst2_r, dst3_r, 6); MUL2(dst0_r, weight_vec, dst1_r, weight_vec, dst0_r, dst1_r); - SRAR_W2_SW(dst0_r, dst1_r, rnd_vec); + MUL2(dst2_r, weight_vec, dst3_r, weight_vec, dst2_r, dst3_r); + SRAR_W4_SW(dst0_r, dst1_r, dst2_r, dst3_r, rnd_vec); ADD2(dst0_r, offset_vec, dst1_r, offset_vec, dst0_r, dst1_r); - dst0_r = CLIP_SW_0_255(dst0_r); - dst1_r = CLIP_SW_0_255(dst1_r); - - HEVC_PCK_SW_SB2(dst1_r, dst0_r, dst0_r); - ST4x2_UB(dst0_r, dst, dst_stride); - dst += (2 * dst_stride); + ADD2(dst2_r, offset_vec, dst3_r, offset_vec, dst2_r, dst3_r); + CLIP_SW4_0_255_MAX_SATU(dst0_r, dst1_r, dst2_r, dst3_r); + PCKEV_H2_SW(dst1_r, dst0_r, dst3_r, dst2_r, dst0_r, dst1_r); + out = (v16u8) __msa_pckev_b((v16i8) dst1_r, (v16i8) dst0_r); + ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride); + dst += (4 * dst_stride); - dst10_r = dst32_r; - dst32_r = dst54_r; - dst54_r = dst76_r; - dst21_r = dst43_r; - dst43_r = dst65_r; - dst65_r = dst87_r; - dst66 = (v8i16) __msa_splati_d((v2i64) dst87, 1); + dst10_r = dst54_r; + dst32_r = dst76_r; + dst54_r = dst98_r; + dst21_r = dst65_r; + dst43_r = dst87_r; + dst65_r = dst109_r; + dst66 = (v8i16) __msa_splati_d((v2i64) dst108, 1); } } @@ -1927,7 +1933,7 @@ static void hevc_hv_uniwgt_8t_8multx2mult_msa(uint8_t *src, v8i16 filt0, filt1, filt2, filt3; v4i32 filt_h0, filt_h1, filt_h2, filt_h3; v16i8 mask1, mask2, mask3; - v8i16 filter_vec, const_vec; + v8i16 filter_vec; v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7; v16i8 vec8, vec9, vec10, vec11, vec12, vec13, vec14, vec15; v8i16 dst0, dst1, dst2, dst3, dst4, dst5, dst6, dst7, dst8; @@ -1936,23 +1942,25 @@ static void hevc_hv_uniwgt_8t_8multx2mult_msa(uint8_t *src, v8i16 dst10_l, dst32_l, dst54_l, dst76_l; v8i16 dst21_r, dst43_r, dst65_r, dst87_r; v8i16 dst21_l, dst43_l, dst65_l, dst87_l; - v4i32 weight_vec, offset_vec, rnd_vec; - v16i8 mask0 = { 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8 }; + v4i32 weight_vec, offset_vec, rnd_vec, const_128, denom_vec; + v16i8 mask0 = LD_SB(ff_hevc_mask_arr); src -= ((3 * src_stride) + 3); - const_vec = __msa_ldi_h(128); - const_vec <<= 6; weight_vec = __msa_fill_w(weight); offset_vec = __msa_fill_w(offset); rnd_vec = __msa_fill_w(rnd_val); + denom_vec = rnd_vec - 6; + + const_128 = __msa_ldi_w(128); + const_128 *= weight_vec; + offset_vec += __msa_srar_w(const_128, denom_vec); filter_vec = LD_SH(filter_x); SPLATI_H4_SH(filter_vec, 0, 1, 2, 3, filt0, filt1, filt2, filt3); filter_vec = LD_SH(filter_y); - vec0 = __msa_clti_s_b((v16i8) filter_vec, 0); - filter_vec = (v8i16) __msa_ilvr_b(vec0, (v16i8) filter_vec); + UNPCK_R_SB_SH(filter_vec, filter_vec); SPLATI_W4_SW(filter_vec, filt_h0, filt_h1, filt_h2, filt_h3); mask1 = mask0 + 2; @@ -1975,18 +1983,14 @@ static void hevc_hv_uniwgt_8t_8multx2mult_msa(uint8_t *src, vec8, vec9, vec10, vec11); VSHF_B4_SB(src3, src3, mask0, mask1, mask2, mask3, vec12, vec13, vec14, vec15); - dst0 = const_vec; - DPADD_SB4_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, filt3, - dst0, dst0, dst0, dst0); - dst1 = const_vec; - DPADD_SB4_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, filt3, - dst1, dst1, dst1, dst1); - dst2 = const_vec; - DPADD_SB4_SH(vec8, vec9, vec10, vec11, filt0, filt1, filt2, filt3, - dst2, dst2, dst2, dst2); - dst3 = const_vec; - DPADD_SB4_SH(vec12, vec13, vec14, vec15, filt0, filt1, filt2, filt3, - dst3, dst3, dst3, dst3); + dst0 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, + filt3); + dst1 = HEVC_FILT_8TAP_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, + filt3); + dst2 = HEVC_FILT_8TAP_SH(vec8, vec9, vec10, vec11, filt0, filt1, filt2, + filt3); + dst3 = HEVC_FILT_8TAP_SH(vec12, vec13, vec14, vec15, filt0, filt1, + filt2, filt3); VSHF_B4_SB(src4, src4, mask0, mask1, mask2, mask3, vec0, vec1, vec2, vec3); @@ -1994,15 +1998,12 @@ static void hevc_hv_uniwgt_8t_8multx2mult_msa(uint8_t *src, vec4, vec5, vec6, vec7); VSHF_B4_SB(src6, src6, mask0, mask1, mask2, mask3, vec8, vec9, vec10, vec11); - dst4 = const_vec; - DPADD_SB4_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, filt3, - dst4, dst4, dst4, dst4); - dst5 = const_vec; - DPADD_SB4_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, filt3, - dst5, dst5, dst5, dst5); - dst6 = const_vec; - DPADD_SB4_SH(vec8, vec9, vec10, vec11, filt0, filt1, filt2, filt3, - dst6, dst6, dst6, dst6); + dst4 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, + filt3); + dst5 = HEVC_FILT_8TAP_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, + filt3); + dst6 = HEVC_FILT_8TAP_SH(vec8, vec9, vec10, vec11, filt0, filt1, filt2, + filt3); ILVR_H4_SH(dst1, dst0, dst3, dst2, dst5, dst4, dst2, dst1, dst10_r, dst32_r, dst54_r, dst21_r); @@ -2018,9 +2019,8 @@ static void hevc_hv_uniwgt_8t_8multx2mult_msa(uint8_t *src, VSHF_B4_SB(src7, src7, mask0, mask1, mask2, mask3, vec0, vec1, vec2, vec3); - dst7 = const_vec; - DPADD_SB4_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, filt3, - dst7, dst7, dst7, dst7); + dst7 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, + filt2, filt3); ILVRL_H2_SH(dst7, dst6, dst76_r, dst76_l); dst0_r = HEVC_FILT_8TAP(dst10_r, dst32_r, dst54_r, dst76_r, @@ -2033,9 +2033,8 @@ static void hevc_hv_uniwgt_8t_8multx2mult_msa(uint8_t *src, /* row 8 */ VSHF_B4_SB(src8, src8, mask0, mask1, mask2, mask3, vec0, vec1, vec2, vec3); - dst8 = const_vec; - DPADD_SB4_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, filt3, - dst8, dst8, dst8, dst8); + dst8 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, + filt2, filt3); ILVRL_H2_SH(dst8, dst7, dst87_r, dst87_l); dst1_r = HEVC_FILT_8TAP(dst21_r, dst43_r, dst65_r, dst87_r, @@ -2045,11 +2044,15 @@ static void hevc_hv_uniwgt_8t_8multx2mult_msa(uint8_t *src, dst1_r >>= 6; dst1_l >>= 6; - HEVC_HV_UNIW_RND_CLIP4(dst0_r, dst1_r, dst0_l, dst1_l, - weight_vec, offset_vec, rnd_vec, - dst0_r, dst1_r, dst0_l, dst1_l); + MUL2(dst0_r, weight_vec, dst0_l, weight_vec, dst0_r, dst0_l); + MUL2(dst1_r, weight_vec, dst1_l, weight_vec, dst1_r, dst1_l); + SRAR_W4_SW(dst0_r, dst1_r, dst0_l, dst1_l, rnd_vec); + ADD2(dst0_r, offset_vec, dst0_l, offset_vec, dst0_r, dst0_l); + ADD2(dst1_r, offset_vec, dst1_l, offset_vec, dst1_r, dst1_l); + CLIP_SW4_0_255_MAX_SATU(dst0_r, dst1_r, dst0_l, dst1_l); - HEVC_PCK_SW_SB4(dst0_l, dst0_r, dst1_l, dst1_r, dst0_r); + PCKEV_H2_SW(dst0_l, dst0_r, dst1_l, dst1_r, dst0_r, dst1_r); + dst0_r = (v4i32) __msa_pckev_b((v16i8) dst1_r, (v16i8) dst0_r); ST8x2_UB(dst0_r, dst_tmp, dst_stride); dst_tmp += (2 * dst_stride); @@ -2100,12 +2103,198 @@ static void hevc_hv_uniwgt_8t_12w_msa(uint8_t *src, int32_t offset, int32_t rnd_val) { - hevc_hv_uniwgt_8t_8multx2mult_msa(src, src_stride, dst, dst_stride, - filter_x, filter_y, height, weight, - offset, rnd_val, 8); - hevc_hv_uniwgt_8t_4w_msa(src + 8, src_stride, dst + 8, dst_stride, - filter_x, filter_y, height, weight, offset, - rnd_val); + uint32_t loop_cnt; + uint8_t *src_tmp, *dst_tmp; + v16u8 out; + v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8, src9, src10; + v16i8 mask0, mask1, mask2, mask3, mask4, mask5, mask6, mask7; + v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7; + v16i8 vec8, vec9, vec10, vec11, vec12, vec13, vec14, vec15; + v8i16 dst0, dst1, dst2, dst3, dst4, dst5, dst6, dst7; + v8i16 dst30, dst41, dst52, dst63, dst66, dst97, dst108; + v8i16 filt0, filt1, filt2, filt3, filt_h0, filt_h1, filt_h2, filt_h3; + v8i16 dst10_r, dst32_r, dst54_r, dst76_r, dst10_l, dst32_l, dst54_l; + v8i16 dst98_r, dst21_r, dst43_r, dst65_r, dst87_r, dst109_r; + v8i16 dst76_l, filter_vec; + v4i32 dst0_r, dst0_l, dst1_r, dst2_r, dst3_r; + v4i32 weight_vec, offset_vec, rnd_vec, const_128, denom_vec; + + src -= ((3 * src_stride) + 3); + + filter_vec = LD_SH(filter_x); + SPLATI_H4_SH(filter_vec, 0, 1, 2, 3, filt0, filt1, filt2, filt3); + + filter_vec = LD_SH(filter_y); + UNPCK_R_SB_SH(filter_vec, filter_vec); + + SPLATI_W4_SH(filter_vec, filt_h0, filt_h1, filt_h2, filt_h3); + + weight_vec = __msa_fill_w(weight); + offset_vec = __msa_fill_w(offset); + rnd_vec = __msa_fill_w(rnd_val); + denom_vec = rnd_vec - 6; + + const_128 = __msa_ldi_w(128); + const_128 *= weight_vec; + offset_vec += __msa_srar_w(const_128, denom_vec); + + mask0 = LD_SB(ff_hevc_mask_arr); + mask1 = mask0 + 2; + mask2 = mask0 + 4; + mask3 = mask0 + 6; + + src_tmp = src; + dst_tmp = dst; + + LD_SB7(src_tmp, src_stride, src0, src1, src2, src3, src4, src5, src6); + src_tmp += (7 * src_stride); + XORI_B7_128_SB(src0, src1, src2, src3, src4, src5, src6); + + /* row 0 row 1 row 2 row 3 */ + VSHF_B4_SB(src0, src0, mask0, mask1, mask2, mask3, vec0, vec1, vec2, vec3); + VSHF_B4_SB(src1, src1, mask0, mask1, mask2, mask3, vec4, vec5, vec6, vec7); + VSHF_B4_SB(src2, src2, mask0, mask1, mask2, mask3, vec8, vec9, vec10, + vec11); + VSHF_B4_SB(src3, src3, mask0, mask1, mask2, mask3, vec12, vec13, vec14, + vec15); + dst0 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, + filt3); + dst1 = HEVC_FILT_8TAP_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, + filt3); + dst2 = HEVC_FILT_8TAP_SH(vec8, vec9, vec10, vec11, filt0, filt1, filt2, + filt3); + dst3 = HEVC_FILT_8TAP_SH(vec12, vec13, vec14, vec15, filt0, filt1, + filt2, filt3); + VSHF_B4_SB(src4, src4, mask0, mask1, mask2, mask3, vec0, vec1, vec2, vec3); + VSHF_B4_SB(src5, src5, mask0, mask1, mask2, mask3, vec4, vec5, vec6, vec7); + VSHF_B4_SB(src6, src6, mask0, mask1, mask2, mask3, vec8, vec9, vec10, + vec11); + dst4 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, + filt3); + dst5 = HEVC_FILT_8TAP_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, + filt3); + dst6 = HEVC_FILT_8TAP_SH(vec8, vec9, vec10, vec11, filt0, filt1, filt2, + filt3); + + for (loop_cnt = 16; loop_cnt--;) { + src7 = LD_SB(src_tmp); + src7 = (v16i8) __msa_xori_b((v16u8) src7, 128); + src_tmp += src_stride; + + VSHF_B4_SB(src7, src7, mask0, mask1, mask2, mask3, vec0, vec1, vec2, + vec3); + dst7 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, + filt3); + ILVRL_H2_SH(dst1, dst0, dst10_r, dst10_l); + ILVRL_H2_SH(dst3, dst2, dst32_r, dst32_l); + ILVRL_H2_SH(dst5, dst4, dst54_r, dst54_l); + ILVRL_H2_SH(dst7, dst6, dst76_r, dst76_l); + + dst0_r = HEVC_FILT_8TAP(dst10_r, dst32_r, dst54_r, dst76_r, + filt_h0, filt_h1, filt_h2, filt_h3); + dst0_l = HEVC_FILT_8TAP(dst10_l, dst32_l, dst54_l, dst76_l, + filt_h0, filt_h1, filt_h2, filt_h3); + dst0_r >>= 6; + dst0_l >>= 6; + + MUL2(dst0_r, weight_vec, dst0_l, weight_vec, dst0_r, dst0_l); + SRAR_W2_SW(dst0_r, dst0_l, rnd_vec); + ADD2(dst0_r, offset_vec, dst0_l, offset_vec, dst0_r, dst0_l); + CLIP_SW2_0_255_MAX_SATU(dst0_r, dst0_l); + dst0_r = (v4i32) __msa_pckev_h((v8i16) dst0_l, (v8i16) dst0_r); + out = (v16u8) __msa_pckev_b((v16i8) dst0_r, (v16i8) dst0_r); + ST8x1_UB(out, dst_tmp); + dst_tmp += dst_stride; + + dst0 = dst1; + dst1 = dst2; + dst2 = dst3; + dst3 = dst4; + dst4 = dst5; + dst5 = dst6; + dst6 = dst7; + } + + src += 8; + dst += 8; + + mask4 = LD_SB(ff_hevc_mask_arr + 16); + mask5 = mask4 + 2; + mask6 = mask4 + 4; + mask7 = mask4 + 6; + + LD_SB7(src, src_stride, src0, src1, src2, src3, src4, src5, src6); + src += (7 * src_stride); + XORI_B7_128_SB(src0, src1, src2, src3, src4, src5, src6); + + VSHF_B4_SB(src0, src3, mask4, mask5, mask6, mask7, vec0, vec1, vec2, vec3); + VSHF_B4_SB(src1, src4, mask4, mask5, mask6, mask7, vec4, vec5, vec6, vec7); + VSHF_B4_SB(src2, src5, mask4, mask5, mask6, mask7, vec8, vec9, vec10, + vec11); + VSHF_B4_SB(src3, src6, mask4, mask5, mask6, mask7, vec12, vec13, vec14, + vec15); + dst30 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, + filt3); + dst41 = HEVC_FILT_8TAP_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, + filt3); + dst52 = HEVC_FILT_8TAP_SH(vec8, vec9, vec10, vec11, filt0, filt1, filt2, + filt3); + dst63 = HEVC_FILT_8TAP_SH(vec12, vec13, vec14, vec15, filt0, filt1, filt2, + filt3); + ILVRL_H2_SH(dst41, dst30, dst10_r, dst43_r); + ILVRL_H2_SH(dst52, dst41, dst21_r, dst54_r); + ILVRL_H2_SH(dst63, dst52, dst32_r, dst65_r); + + dst66 = (v8i16) __msa_splati_d((v2i64) dst63, 1); + + for (loop_cnt = 4; loop_cnt--;) { + LD_SB4(src, src_stride, src7, src8, src9, src10); + src += (4 * src_stride); + XORI_B4_128_SB(src7, src8, src9, src10); + + VSHF_B4_SB(src7, src9, mask4, mask5, mask6, mask7, vec0, vec1, vec2, + vec3); + VSHF_B4_SB(src8, src10, mask4, mask5, mask6, mask7, vec4, vec5, vec6, + vec7); + dst97 = HEVC_FILT_8TAP_SH(vec0, vec1, vec2, vec3, filt0, filt1, filt2, + filt3); + dst108 = HEVC_FILT_8TAP_SH(vec4, vec5, vec6, vec7, filt0, filt1, filt2, + filt3); + + dst76_r = __msa_ilvr_h(dst97, dst66); + ILVRL_H2_SH(dst108, dst97, dst87_r, dst109_r); + dst66 = (v8i16) __msa_splati_d((v2i64) dst97, 1); + dst98_r = __msa_ilvr_h(dst66, dst108); + + dst0_r = HEVC_FILT_8TAP(dst10_r, dst32_r, dst54_r, dst76_r, filt_h0, + filt_h1, filt_h2, filt_h3); + dst1_r = HEVC_FILT_8TAP(dst21_r, dst43_r, dst65_r, dst87_r, filt_h0, + filt_h1, filt_h2, filt_h3); + dst2_r = HEVC_FILT_8TAP(dst32_r, dst54_r, dst76_r, dst98_r, filt_h0, + filt_h1, filt_h2, filt_h3); + dst3_r = HEVC_FILT_8TAP(dst43_r, dst65_r, dst87_r, dst109_r, filt_h0, + filt_h1, filt_h2, filt_h3); + + SRA_4V(dst0_r, dst1_r, dst2_r, dst3_r, 6); + MUL2(dst0_r, weight_vec, dst1_r, weight_vec, dst0_r, dst1_r); + MUL2(dst2_r, weight_vec, dst3_r, weight_vec, dst2_r, dst3_r); + SRAR_W4_SW(dst0_r, dst1_r, dst2_r, dst3_r, rnd_vec); + ADD2(dst0_r, offset_vec, dst1_r, offset_vec, dst0_r, dst1_r); + ADD2(dst2_r, offset_vec, dst3_r, offset_vec, dst2_r, dst3_r); + CLIP_SW4_0_255_MAX_SATU(dst0_r, dst1_r, dst2_r, dst3_r); + PCKEV_H2_SW(dst1_r, dst0_r, dst3_r, dst2_r, dst0_r, dst1_r); + out = (v16u8) __msa_pckev_b((v16i8) dst1_r, (v16i8) dst0_r); + ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride); + dst += (4 * dst_stride); + + dst10_r = dst54_r; + dst32_r = dst76_r; + dst54_r = dst98_r; + dst21_r = dst65_r; + dst43_r = dst87_r; + dst65_r = dst109_r; + dst66 = (v8i16) __msa_splati_d((v2i64) dst108, 1); + } } static void hevc_hv_uniwgt_8t_16w_msa(uint8_t *src, @@ -3977,22 +4166,20 @@ static void hevc_hv_uniwgt_4t_4x2_msa(uint8_t *src, int32_t dst_stride, const int8_t *filter_x, const int8_t *filter_y, - int32_t height, int32_t weight, int32_t offset, int32_t rnd_val) { + v16u8 out; v16i8 src0, src1, src2, src3, src4; v8i16 filt0, filt1; - v4i32 filt_h0, filt_h1; - v16i8 mask0 = { 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8 }; + v16i8 mask0 = LD_SB(ff_hevc_mask_arr + 16); v16i8 mask1; - v8i16 filter_vec, const_vec; + v8i16 filt_h0, filt_h1, filter_vec, tmp; v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v8i16 dst0, dst1, dst2, dst3, dst4; - v8i16 dst10_r, dst32_r, dst21_r, dst43_r; - v4i32 dst0_r, dst1_r; - v4i32 weight_vec, offset_vec, rnd_vec; + v8i16 dst20, dst31, dst42, dst10, dst32, dst21, dst43; + v8i16 offset_vec, const_128, denom_vec; + v4i32 dst0, dst1, weight_vec, rnd_vec; src -= (src_stride + 1); @@ -4000,63 +4187,41 @@ static void hevc_hv_uniwgt_4t_4x2_msa(uint8_t *src, SPLATI_H2_SH(filter_vec, 0, 1, filt0, filt1); filter_vec = LD_SH(filter_y); - vec0 = __msa_clti_s_b((v16i8) filter_vec, 0); - filter_vec = (v8i16) __msa_ilvr_b(vec0, (v16i8) filter_vec); + UNPCK_R_SB_SH(filter_vec, filter_vec); - SPLATI_W2_SW(filter_vec, 0, filt_h0, filt_h1); + SPLATI_W2_SH(filter_vec, 0, filt_h0, filt_h1); mask1 = mask0 + 2; - const_vec = __msa_ldi_h(128); - const_vec <<= 6; - weight_vec = __msa_fill_w(weight); - offset_vec = __msa_fill_w(offset); rnd_vec = __msa_fill_w(rnd_val); - LD_SB3(src, src_stride, src0, src1, src2); - src += (3 * src_stride); - XORI_B3_128_SB(src0, src1, src2); - - VSHF_B2_SB(src0, src0, src0, src0, mask0, mask1, vec0, vec1); - VSHF_B2_SB(src1, src1, src1, src1, mask0, mask1, vec2, vec3); - VSHF_B2_SB(src2, src2, src2, src2, mask0, mask1, vec4, vec5); - dst0 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst0, dst0); - dst1 = const_vec; - DPADD_SB2_SH(vec2, vec3, filt0, filt1, dst1, dst1); - dst2 = const_vec; - DPADD_SB2_SH(vec4, vec5, filt0, filt1, dst2, dst2); - - ILVR_H2_SH(dst1, dst0, dst2, dst1, dst10_r, dst21_r); - LD_SB2(src, src_stride, src3, src4); - XORI_B2_128_SB(src3, src4); - - /* row 3 */ - VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); - dst3 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst3, dst3); - - dst32_r = __msa_ilvr_h(dst3, dst2); - dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); - dst0_r >>= 6; - - VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec0, vec1); - dst4 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst4, dst4); - - dst43_r = __msa_ilvr_h(dst4, dst3); - dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); - dst1_r >>= 6; - - MUL2(dst0_r, weight_vec, dst1_r, weight_vec, dst0_r, dst1_r); - SRAR_W2_SW(dst0_r, dst1_r, rnd_vec); - ADD2(dst0_r, offset_vec, dst1_r, offset_vec, dst0_r, dst1_r); - dst0_r = CLIP_SW_0_255(dst0_r); - dst1_r = CLIP_SW_0_255(dst1_r); - - HEVC_PCK_SW_SB2(dst1_r, dst0_r, dst0_r); - ST4x2_UB(dst0_r, dst, dst_stride); + offset_vec = __msa_fill_h(offset); + denom_vec = __msa_fill_h(rnd_val - 6); + const_128 = __msa_fill_h((128 * weight)); + offset_vec += __msa_srar_h(const_128, denom_vec); + + LD_SB5(src, src_stride, src0, src1, src2, src3, src4); + XORI_B5_128_SB(src0, src1, src2, src3, src4); + VSHF_B2_SB(src0, src2, src0, src2, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src1, src3, src1, src3, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src2, src4, src2, src4, mask0, mask1, vec4, vec5); + dst20 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst31 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst42 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + ILVRL_H2_SH(dst31, dst20, dst10, dst32); + ILVRL_H2_SH(dst42, dst31, dst21, dst43); + dst0 = HEVC_FILT_4TAP(dst10, dst32, filt_h0, filt_h1); + dst1 = HEVC_FILT_4TAP(dst21, dst43, filt_h0, filt_h1); + dst0 >>= 6; + dst1 >>= 6; + MUL2(dst0, weight_vec, dst1, weight_vec, dst0, dst1); + SRAR_W2_SW(dst0, dst1, rnd_vec); + tmp = __msa_pckev_h((v8i16) dst1, (v8i16) dst0); + tmp += offset_vec; + tmp = CLIP_SH_0_255_MAX_SATU(tmp); + out = (v16u8) __msa_pckev_b((v16i8) tmp, (v16i8) tmp); + ST4x2_UB(out, dst, dst_stride); } static void hevc_hv_uniwgt_4t_4x4_msa(uint8_t *src, @@ -4065,22 +4230,20 @@ static void hevc_hv_uniwgt_4t_4x4_msa(uint8_t *src, int32_t dst_stride, const int8_t *filter_x, const int8_t *filter_y, - int32_t height, int32_t weight, int32_t offset, int32_t rnd_val) { + v16u8 out; v16i8 src0, src1, src2, src3, src4, src5, src6; v8i16 filt0, filt1; - v4i32 filt_h0, filt_h1; - v16i8 mask0 = { 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8 }; + v8i16 filt_h0, filt_h1, filter_vec, tmp0, tmp1; + v16i8 mask0 = LD_SB(ff_hevc_mask_arr + 16); v16i8 mask1; - v8i16 filter_vec, const_vec; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v8i16 dst0, dst1, dst2, dst3, dst4, dst5; - v4i32 dst0_r, dst1_r, dst2_r, dst3_r; - v8i16 dst10_r, dst32_r, dst21_r, dst43_r; - v4i32 weight_vec, offset_vec, rnd_vec; + v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7; + v8i16 dst30, dst41, dst52, dst63, dst10, dst32, dst54, dst21, dst43, dst65; + v8i16 offset_vec, const_128, denom_vec; + v4i32 dst0, dst1, dst2, dst3, weight_vec, rnd_vec; src -= (src_stride + 1); @@ -4088,76 +4251,46 @@ static void hevc_hv_uniwgt_4t_4x4_msa(uint8_t *src, SPLATI_H2_SH(filter_vec, 0, 1, filt0, filt1); filter_vec = LD_SH(filter_y); - vec0 = __msa_clti_s_b((v16i8) filter_vec, 0); - filter_vec = (v8i16) __msa_ilvr_b(vec0, (v16i8) filter_vec); + UNPCK_R_SB_SH(filter_vec, filter_vec); - SPLATI_W2_SW(filter_vec, 0, filt_h0, filt_h1); + SPLATI_W2_SH(filter_vec, 0, filt_h0, filt_h1); mask1 = mask0 + 2; - const_vec = __msa_ldi_h(128); - const_vec <<= 6; - weight_vec = __msa_fill_w(weight); - offset_vec = __msa_fill_w(offset); rnd_vec = __msa_fill_w(rnd_val); - LD_SB3(src, src_stride, src0, src1, src2); - src += (3 * src_stride); - XORI_B3_128_SB(src0, src1, src2); - - VSHF_B2_SB(src0, src0, src0, src0, mask0, mask1, vec0, vec1); - VSHF_B2_SB(src1, src1, src1, src1, mask0, mask1, vec2, vec3); - VSHF_B2_SB(src2, src2, src2, src2, mask0, mask1, vec4, vec5); - dst0 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst0, dst0); - dst1 = const_vec; - DPADD_SB2_SH(vec2, vec3, filt0, filt1, dst1, dst1); - dst2 = const_vec; - DPADD_SB2_SH(vec4, vec5, filt0, filt1, dst2, dst2); - - ILVR_H2_SH(dst1, dst0, dst2, dst1, dst10_r, dst21_r); - - LD_SB4(src, src_stride, src3, src4, src5, src6); - XORI_B4_128_SB(src3, src4, src5, src6); - - /* row 3 */ - VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); - dst3 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst3, dst3); - dst32_r = __msa_ilvr_h(dst3, dst2); - dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); - dst0_r >>= 6; - - /* row 4 */ - VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec0, vec1); - dst4 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst4, dst4); - dst43_r = __msa_ilvr_h(dst4, dst3); - dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); - dst1_r >>= 6; - - /* row 5 */ - VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec0, vec1); - dst5 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst5, dst5); - dst10_r = __msa_ilvr_h(dst5, dst4); - dst2_r = HEVC_FILT_4TAP(dst32_r, dst10_r, filt_h0, filt_h1); - dst2_r >>= 6; + offset_vec = __msa_fill_h(offset); + denom_vec = __msa_fill_h(rnd_val - 6); + const_128 = __msa_fill_h((128 * weight)); + offset_vec += __msa_srar_h(const_128, denom_vec); - /* row 6 */ - VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec0, vec1); - dst2 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst2, dst2); - dst21_r = __msa_ilvr_h(dst2, dst5); - dst3_r = HEVC_FILT_4TAP(dst43_r, dst21_r, filt_h0, filt_h1); - dst3_r >>= 6; - - HEVC_HV_UNIW_RND_CLIP4(dst0_r, dst1_r, dst2_r, dst3_r, - weight_vec, offset_vec, rnd_vec, - dst0_r, dst1_r, dst2_r, dst3_r); - HEVC_PCK_SW_SB4(dst1_r, dst0_r, dst3_r, dst2_r, dst0_r); - ST4x4_UB(dst0_r, dst0_r, 0, 1, 2, 3, dst, dst_stride); + LD_SB7(src, src_stride, src0, src1, src2, src3, src4, src5, src6); + XORI_B7_128_SB(src0, src1, src2, src3, src4, src5, src6); + VSHF_B2_SB(src0, src3, src0, src3, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src1, src4, src1, src4, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src2, src5, src2, src5, mask0, mask1, vec4, vec5); + VSHF_B2_SB(src3, src6, src3, src6, mask0, mask1, vec6, vec7); + dst30 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst41 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst52 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dst63 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); + ILVRL_H2_SH(dst41, dst30, dst10, dst43); + ILVRL_H2_SH(dst52, dst41, dst21, dst54); + ILVRL_H2_SH(dst63, dst52, dst32, dst65); + dst0 = HEVC_FILT_4TAP(dst10, dst32, filt_h0, filt_h1); + dst1 = HEVC_FILT_4TAP(dst21, dst43, filt_h0, filt_h1); + dst2 = HEVC_FILT_4TAP(dst32, dst54, filt_h0, filt_h1); + dst3 = HEVC_FILT_4TAP(dst43, dst65, filt_h0, filt_h1); + SRA_4V(dst0, dst1, dst2, dst3, 6); + MUL2(dst0, weight_vec, dst1, weight_vec, dst0, dst1); + MUL2(dst2, weight_vec, dst3, weight_vec, dst2, dst3); + SRAR_W4_SW(dst0, dst1, dst2, dst3, rnd_vec); + PCKEV_H2_SH(dst1, dst0, dst3, dst2, tmp0, tmp1); + ADD2(tmp0, offset_vec, tmp1, offset_vec, tmp0, tmp1); + CLIP_SH2_0_255_MAX_SATU(tmp0, tmp1); + out = (v16u8) __msa_pckev_b((v16i8) tmp1, (v16i8) tmp0); + ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride); } static void hevc_hv_uniwgt_4t_4multx8mult_msa(uint8_t *src, @@ -4172,18 +4305,18 @@ static void hevc_hv_uniwgt_4t_4multx8mult_msa(uint8_t *src, int32_t rnd_val) { uint32_t loop_cnt; + v16u8 out0, out1; v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8, src9, src10; v8i16 filt0, filt1; - v4i32 filt_h0, filt_h1; - v16i8 mask0 = { 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8 }; + v16i8 mask0 = LD_SB(ff_hevc_mask_arr + 16); v16i8 mask1; - v8i16 filter_vec, const_vec; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v8i16 dst0, dst1, dst2, dst3, dst4, dst5, dst6, dst7, dst8, dst9; - v4i32 dst0_r, dst1_r, dst2_r, dst3_r, dst4_r, dst5_r, dst6_r, dst7_r; + v8i16 filt_h0, filt_h1, filter_vec, tmp0, tmp1, tmp2, tmp3; + v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7; + v8i16 dst10, dst21, dst22, dst73, dst84, dst95, dst106; v8i16 dst10_r, dst32_r, dst54_r, dst76_r; v8i16 dst21_r, dst43_r, dst65_r, dst87_r; - v4i32 weight_vec, offset_vec, rnd_vec; + v8i16 dst98_r, dst109_r, offset_vec, const_128, denom_vec; + v4i32 dst0, dst1, dst2, dst3, dst4, dst5, dst6, dst7, weight_vec, rnd_vec; src -= (src_stride + 1); @@ -4191,34 +4324,30 @@ static void hevc_hv_uniwgt_4t_4multx8mult_msa(uint8_t *src, SPLATI_H2_SH(filter_vec, 0, 1, filt0, filt1); filter_vec = LD_SH(filter_y); - vec0 = __msa_clti_s_b((v16i8) filter_vec, 0); - filter_vec = (v8i16) __msa_ilvr_b(vec0, (v16i8) filter_vec); + UNPCK_R_SB_SH(filter_vec, filter_vec); - SPLATI_W2_SW(filter_vec, 0, filt_h0, filt_h1); + SPLATI_W2_SH(filter_vec, 0, filt_h0, filt_h1); mask1 = mask0 + 2; - const_vec = __msa_ldi_h(128); - const_vec <<= 6; - weight_vec = __msa_fill_w(weight); - offset_vec = __msa_fill_w(offset); rnd_vec = __msa_fill_w(rnd_val); + offset_vec = __msa_fill_h(offset); + denom_vec = __msa_fill_h(rnd_val - 6); + const_128 = __msa_fill_h((128 * weight)); + offset_vec += __msa_srar_h(const_128, denom_vec); + LD_SB3(src, src_stride, src0, src1, src2); src += (3 * src_stride); XORI_B3_128_SB(src0, src1, src2); - VSHF_B2_SB(src0, src0, src0, src0, mask0, mask1, vec0, vec1); - VSHF_B2_SB(src1, src1, src1, src1, mask0, mask1, vec2, vec3); - VSHF_B2_SB(src2, src2, src2, src2, mask0, mask1, vec4, vec5); - dst0 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst0, dst0); - dst1 = const_vec; - DPADD_SB2_SH(vec2, vec3, filt0, filt1, dst1, dst1); - dst2 = const_vec; - DPADD_SB2_SH(vec4, vec5, filt0, filt1, dst2, dst2); - ILVR_H2_SH(dst1, dst0, dst2, dst1, dst10_r, dst21_r); + VSHF_B2_SB(src0, src1, src0, src1, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src1, src2, src1, src2, mask0, mask1, vec2, vec3); + dst10 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst21 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + ILVRL_H2_SH(dst21, dst10, dst10_r, dst21_r); + dst22 = (v8i16) __msa_splati_d((v2i64) dst21, 1); for (loop_cnt = height >> 3; loop_cnt--;) { LD_SB8(src, src_stride, @@ -4226,77 +4355,50 @@ static void hevc_hv_uniwgt_4t_4multx8mult_msa(uint8_t *src, src += (8 * src_stride); XORI_B8_128_SB(src3, src4, src5, src6, src7, src8, src9, src10); - VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); - dst3 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst3, dst3); - dst32_r = __msa_ilvr_h(dst3, dst2); - dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); - dst0_r >>= 6; + VSHF_B2_SB(src3, src7, src3, src7, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src4, src8, src4, src8, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src5, src9, src5, src9, mask0, mask1, vec4, vec5); + VSHF_B2_SB(src6, src10, src6, src10, mask0, mask1, vec6, vec7); + dst73 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst84 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst95 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dst106 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); + dst32_r = __msa_ilvr_h(dst73, dst22); + ILVRL_H2_SH(dst84, dst73, dst43_r, dst87_r); + ILVRL_H2_SH(dst95, dst84, dst54_r, dst98_r); + ILVRL_H2_SH(dst106, dst95, dst65_r, dst109_r); + dst22 = (v8i16) __msa_splati_d((v2i64) dst73, 1); + dst76_r = __msa_ilvr_h(dst22, dst106); + dst0 = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); + dst1 = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); + dst2 = HEVC_FILT_4TAP(dst32_r, dst54_r, filt_h0, filt_h1); + dst3 = HEVC_FILT_4TAP(dst43_r, dst65_r, filt_h0, filt_h1); + dst4 = HEVC_FILT_4TAP(dst54_r, dst76_r, filt_h0, filt_h1); + dst5 = HEVC_FILT_4TAP(dst65_r, dst87_r, filt_h0, filt_h1); + dst6 = HEVC_FILT_4TAP(dst76_r, dst98_r, filt_h0, filt_h1); + dst7 = HEVC_FILT_4TAP(dst87_r, dst109_r, filt_h0, filt_h1); + SRA_4V(dst0, dst1, dst2, dst3, 6); + SRA_4V(dst4, dst5, dst6, dst7, 6); + MUL2(dst0, weight_vec, dst1, weight_vec, dst0, dst1); + MUL2(dst2, weight_vec, dst3, weight_vec, dst2, dst3); + MUL2(dst4, weight_vec, dst5, weight_vec, dst4, dst5); + MUL2(dst6, weight_vec, dst7, weight_vec, dst6, dst7); + SRAR_W4_SW(dst0, dst1, dst2, dst3, rnd_vec); + SRAR_W4_SW(dst4, dst5, dst6, dst7, rnd_vec); + PCKEV_H4_SH(dst1, dst0, dst3, dst2, dst5, dst4, dst7, dst6, tmp0, tmp1, + tmp2, tmp3); + ADD2(tmp0, offset_vec, tmp1, offset_vec, tmp0, tmp1); + ADD2(tmp2, offset_vec, tmp3, offset_vec, tmp2, tmp3); + CLIP_SH4_0_255_MAX_SATU(tmp0, tmp1, tmp2, tmp3); + PCKEV_B2_UB(tmp1, tmp0, tmp3, tmp2, out0, out1); + ST4x8_UB(out0, out1, dst, dst_stride); + dst += (8 * dst_stride); - VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec0, vec1); - dst4 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst4, dst4); - dst43_r = __msa_ilvr_h(dst4, dst3); - dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); - dst1_r >>= 6; - - VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec0, vec1); - dst5 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst5, dst5); - dst54_r = __msa_ilvr_h(dst5, dst4); - dst2_r = HEVC_FILT_4TAP(dst32_r, dst54_r, filt_h0, filt_h1); - dst2_r >>= 6; - - VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec0, vec1); - dst6 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst6, dst6); - dst65_r = __msa_ilvr_h(dst6, dst5); - dst3_r = HEVC_FILT_4TAP(dst43_r, dst65_r, filt_h0, filt_h1); - dst3_r >>= 6; - - VSHF_B2_SB(src7, src7, src7, src7, mask0, mask1, vec0, vec1); - dst7 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst7, dst7); - dst76_r = __msa_ilvr_h(dst7, dst6); - dst4_r = HEVC_FILT_4TAP(dst54_r, dst76_r, filt_h0, filt_h1); - dst4_r >>= 6; - - VSHF_B2_SB(src8, src8, src8, src8, mask0, mask1, vec0, vec1); - dst8 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst8, dst8); - dst87_r = __msa_ilvr_h(dst8, dst7); - dst5_r = HEVC_FILT_4TAP(dst65_r, dst87_r, filt_h0, filt_h1); - dst5_r >>= 6; - - VSHF_B2_SB(src9, src9, src9, src9, mask0, mask1, vec0, vec1); - dst9 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst9, dst9); - dst10_r = __msa_ilvr_h(dst9, dst8); - dst6_r = HEVC_FILT_4TAP(dst76_r, dst10_r, filt_h0, filt_h1); - dst6_r >>= 6; - - VSHF_B2_SB(src10, src10, src10, src10, mask0, mask1, vec0, vec1); - dst2 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst2, dst2); - dst21_r = __msa_ilvr_h(dst2, dst9); - dst7_r = HEVC_FILT_4TAP(dst87_r, dst21_r, filt_h0, filt_h1); - dst7_r >>= 6; - - HEVC_HV_UNIW_RND_CLIP4(dst0_r, dst1_r, dst2_r, dst3_r, - weight_vec, offset_vec, rnd_vec, - dst0_r, dst1_r, dst2_r, dst3_r); - HEVC_PCK_SW_SB4(dst1_r, dst0_r, dst3_r, dst2_r, dst0_r); - ST4x4_UB(dst0_r, dst0_r, 0, 1, 2, 3, dst, dst_stride); - dst += (4 * dst_stride); - - HEVC_HV_UNIW_RND_CLIP4(dst4_r, dst5_r, dst6_r, dst7_r, - weight_vec, offset_vec, rnd_vec, - dst4_r, dst5_r, dst6_r, dst7_r); - HEVC_PCK_SW_SB4(dst5_r, dst4_r, dst7_r, dst6_r, dst0_r); - ST4x4_UB(dst0_r, dst0_r, 0, 1, 2, 3, dst, dst_stride); - dst += (4 * dst_stride); - } -} + dst10_r = dst98_r; + dst21_r = dst109_r; + dst22 = (v8i16) __msa_splati_d((v2i64) dst106, 1); + } +} static void hevc_hv_uniwgt_4t_4w_msa(uint8_t *src, int32_t src_stride, @@ -4311,11 +4413,11 @@ static void hevc_hv_uniwgt_4t_4w_msa(uint8_t *src, { if (2 == height) { hevc_hv_uniwgt_4t_4x2_msa(src, src_stride, dst, dst_stride, - filter_x, filter_y, height, weight, + filter_x, filter_y, weight, offset, rnd_val); } else if (4 == height) { hevc_hv_uniwgt_4t_4x4_msa(src, src_stride, dst, dst_stride, - filter_x, filter_y, height, weight, + filter_x,filter_y, weight, offset, rnd_val); } else if (0 == (height % 8)) { hevc_hv_uniwgt_4t_4multx8mult_msa(src, src_stride, dst, dst_stride, @@ -4335,20 +4437,22 @@ static void hevc_hv_uniwgt_4t_6w_msa(uint8_t *src, int32_t offset, int32_t rnd_val) { - uint32_t loop_cnt; - v16i8 src0, src1, src2, src3, src4, src5, src6; + v16u8 out0, out1, out2; + v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8, src9, src10; v8i16 filt0, filt1; - v4i32 filt_h0, filt_h1; - v16i8 mask0 = { 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8 }; + v16i8 mask0 = LD_SB(ff_hevc_mask_arr); v16i8 mask1; - v8i16 filter_vec, const_vec; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v8i16 dst0, dst1, dst2, dst3, dst4, dst5; - v4i32 dst0_r, dst0_l, dst1_r, dst1_l; - v4i32 weight_vec, offset_vec, rnd_vec; - v4i32 dst2_r, dst2_l, dst3_r, dst3_l; - v8i16 dst10_r, dst32_r, dst21_r, dst43_r; - v8i16 dst10_l, dst32_l, dst21_l, dst43_l; + v8i16 filt_h0, filt_h1, filter_vec; + v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7; + v8i16 dsth0, dsth1, dsth2, dsth3, dsth4, dsth5, dsth6, dsth7, dsth8, dsth9; + v8i16 dsth10, tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; + v8i16 dst10_r, dst32_r, dst54_r, dst76_r, dst98_r, dst21_r, dst43_r; + v8i16 dst65_r, dst87_r, dst109_r, dst10_l, dst32_l, dst54_l, dst76_l; + v8i16 dst98_l, dst21_l, dst43_l, dst65_l, dst87_l, dst109_l; + v8i16 dst1021_l, dst3243_l, dst5465_l, dst7687_l, dst98109_l; + v8i16 offset_vec, const_128, denom_vec; + v4i32 dst0_r, dst1_r, dst2_r, dst3_r, dst4_r, dst5_r, dst6_r, dst7_r; + v4i32 dst0_l, dst1_l, dst2_l, dst3_l, weight_vec, rnd_vec; src -= (src_stride + 1); @@ -4356,20 +4460,20 @@ static void hevc_hv_uniwgt_4t_6w_msa(uint8_t *src, SPLATI_H2_SH(filter_vec, 0, 1, filt0, filt1); filter_vec = LD_SH(filter_y); - vec0 = __msa_clti_s_b((v16i8) filter_vec, 0); - filter_vec = (v8i16) __msa_ilvr_b(vec0, (v16i8) filter_vec); + UNPCK_R_SB_SH(filter_vec, filter_vec); - SPLATI_W2_SW(filter_vec, 0, filt_h0, filt_h1); + SPLATI_W2_SH(filter_vec, 0, filt_h0, filt_h1); mask1 = mask0 + 2; - const_vec = __msa_ldi_h(128); - const_vec <<= 6; - weight_vec = __msa_fill_w(weight); - offset_vec = __msa_fill_w(offset); rnd_vec = __msa_fill_w(rnd_val); + offset_vec = __msa_fill_h(offset); + denom_vec = __msa_fill_h(rnd_val - 6); + const_128 = __msa_fill_h((128 * weight)); + offset_vec += __msa_srar_h(const_128, denom_vec); + LD_SB3(src, src_stride, src0, src1, src2); src += (3 * src_stride); XORI_B3_128_SB(src0, src1, src2); @@ -4377,72 +4481,78 @@ static void hevc_hv_uniwgt_4t_6w_msa(uint8_t *src, VSHF_B2_SB(src0, src0, src0, src0, mask0, mask1, vec0, vec1); VSHF_B2_SB(src1, src1, src1, src1, mask0, mask1, vec2, vec3); VSHF_B2_SB(src2, src2, src2, src2, mask0, mask1, vec4, vec5); - dst0 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst0, dst0); - dst1 = const_vec; - DPADD_SB2_SH(vec2, vec3, filt0, filt1, dst1, dst1); - dst2 = const_vec; - DPADD_SB2_SH(vec4, vec5, filt0, filt1, dst2, dst2); - - ILVRL_H2_SH(dst1, dst0, dst10_r, dst10_l); - ILVRL_H2_SH(dst2, dst1, dst21_r, dst21_l); - - for (loop_cnt = height >> 2; loop_cnt--;) { - LD_SB4(src, src_stride, src3, src4, src5, src6); - src += (4 * src_stride); - XORI_B4_128_SB(src3, src4, src5, src6); - - /* row 3 */ - VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); - dst3 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst3, dst3); - ILVRL_H2_SH(dst3, dst2, dst32_r, dst32_l); - dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); - dst0_l = HEVC_FILT_4TAP(dst10_l, dst32_l, filt_h0, filt_h1); - dst0_r >>= 6; - dst0_l >>= 6; - - /* row 4 */ - VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec0, vec1); - dst4 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst4, dst4); - ILVRL_H2_SH(dst4, dst3, dst43_r, dst43_l); - dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); - dst1_l = HEVC_FILT_4TAP(dst21_l, dst43_l, filt_h0, filt_h1); - dst1_r >>= 6; - dst1_l >>= 6; - - /* row 5 */ - VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec0, vec1); - dst5 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst5, dst5); - ILVRL_H2_SH(dst5, dst4, dst10_r, dst10_l); - dst2_r = HEVC_FILT_4TAP(dst32_r, dst10_r, filt_h0, filt_h1); - dst2_l = HEVC_FILT_4TAP(dst32_l, dst10_l, filt_h0, filt_h1); - dst2_r >>= 6; - dst2_l >>= 6; - - /* row 6 */ - VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec0, vec1); - dst2 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst2, dst2); - ILVRL_H2_SH(dst2, dst5, dst21_r, dst21_l); - dst3_r = HEVC_FILT_4TAP(dst43_r, dst21_r, filt_h0, filt_h1); - dst3_l = HEVC_FILT_4TAP(dst43_l, dst21_l, filt_h0, filt_h1); - dst3_r >>= 6; - dst3_l >>= 6; - - HEVC_HV_UNIW_RND_CLIP4(dst0_r, dst1_r, dst0_l, dst1_l, - weight_vec, offset_vec, rnd_vec, - dst0_r, dst1_r, dst0_l, dst1_l); - HEVC_HV_UNIW_RND_CLIP4(dst2_r, dst3_r, dst2_l, dst3_l, - weight_vec, offset_vec, rnd_vec, - dst2_r, dst3_r, dst2_l, dst3_l); - HEVC_PCK_SW_SB8(dst0_l, dst0_r, dst1_l, dst1_r, - dst2_l, dst2_r, dst3_l, dst3_r, dst0_r, dst1_r); - ST6x4_UB(dst0_r, dst1_r, dst, dst_stride); - dst += (4 * dst_stride); - } + dsth0 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dsth1 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dsth2 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + ILVRL_H2_SH(dsth1, dsth0, dst10_r, dst10_l); + ILVRL_H2_SH(dsth2, dsth1, dst21_r, dst21_l); + + LD_SB8(src, src_stride, src3, src4, src5, src6, src7, src8, src9, src10); + XORI_B8_128_SB(src3, src4, src5, src6, src7, src8, src9, src10); + VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec4, vec5); + VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec6, vec7); + dsth3 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dsth4 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dsth5 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dsth6 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); + VSHF_B2_SB(src7, src7, src7, src7, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src8, src8, src8, src8, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src9, src9, src9, src9, mask0, mask1, vec4, vec5); + VSHF_B2_SB(src10, src10, src10, src10, mask0, mask1, vec6, vec7); + dsth7 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dsth8 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dsth9 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dsth10 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); + ILVRL_H2_SH(dsth3, dsth2, dst32_r, dst32_l); + ILVRL_H2_SH(dsth4, dsth3, dst43_r, dst43_l); + ILVRL_H2_SH(dsth5, dsth4, dst54_r, dst54_l); + ILVRL_H2_SH(dsth6, dsth5, dst65_r, dst65_l); + ILVRL_H2_SH(dsth7, dsth6, dst76_r, dst76_l); + ILVRL_H2_SH(dsth8, dsth7, dst87_r, dst87_l); + ILVRL_H2_SH(dsth9, dsth8, dst98_r, dst98_l); + ILVRL_H2_SH(dsth10, dsth9, dst109_r, dst109_l); + PCKEV_D2_SH(dst21_l, dst10_l, dst43_l, dst32_l, dst1021_l, dst3243_l); + PCKEV_D2_SH(dst65_l, dst54_l, dst87_l, dst76_l, dst5465_l, dst7687_l); + dst98109_l = (v8i16) __msa_pckev_d((v2i64) dst109_l, (v2i64) dst98_l); + dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); + dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); + dst2_r = HEVC_FILT_4TAP(dst32_r, dst54_r, filt_h0, filt_h1); + dst3_r = HEVC_FILT_4TAP(dst43_r, dst65_r, filt_h0, filt_h1); + dst4_r = HEVC_FILT_4TAP(dst54_r, dst76_r, filt_h0, filt_h1); + dst5_r = HEVC_FILT_4TAP(dst65_r, dst87_r, filt_h0, filt_h1); + dst6_r = HEVC_FILT_4TAP(dst76_r, dst98_r, filt_h0, filt_h1); + dst7_r = HEVC_FILT_4TAP(dst87_r, dst109_r, filt_h0, filt_h1); + dst0_l = HEVC_FILT_4TAP(dst1021_l, dst3243_l, filt_h0, filt_h1); + dst1_l = HEVC_FILT_4TAP(dst3243_l, dst5465_l, filt_h0, filt_h1); + dst2_l = HEVC_FILT_4TAP(dst5465_l, dst7687_l, filt_h0, filt_h1); + dst3_l = HEVC_FILT_4TAP(dst7687_l, dst98109_l, filt_h0, filt_h1); + SRA_4V(dst0_r, dst1_r, dst2_r, dst3_r, 6); + SRA_4V(dst4_r, dst5_r, dst6_r, dst7_r, 6); + SRA_4V(dst0_l, dst1_l, dst2_l, dst3_l, 6); + MUL2(dst0_r, weight_vec, dst1_r, weight_vec, dst0_r, dst1_r); + MUL2(dst2_r, weight_vec, dst3_r, weight_vec, dst2_r, dst3_r); + MUL2(dst4_r, weight_vec, dst5_r, weight_vec, dst4_r, dst5_r); + MUL2(dst6_r, weight_vec, dst7_r, weight_vec, dst6_r, dst7_r); + MUL2(dst0_l, weight_vec, dst1_l, weight_vec, dst0_l, dst1_l); + MUL2(dst2_l, weight_vec, dst3_l, weight_vec, dst2_l, dst3_l); + SRAR_W4_SW(dst0_r, dst1_r, dst2_r, dst3_r, rnd_vec); + SRAR_W4_SW(dst4_r, dst5_r, dst6_r, dst7_r, rnd_vec); + SRAR_W4_SW(dst0_l, dst1_l, dst2_l, dst3_l, rnd_vec); + PCKEV_H2_SH(dst1_r, dst0_r, dst3_r, dst2_r, tmp0, tmp1); + PCKEV_H2_SH(dst5_r, dst4_r, dst7_r, dst6_r, tmp2, tmp3); + PCKEV_H2_SH(dst1_l, dst0_l, dst3_l, dst2_l, tmp4, tmp5); + ADD2(tmp0, offset_vec, tmp1, offset_vec, tmp0, tmp1); + ADD2(tmp2, offset_vec, tmp3, offset_vec, tmp2, tmp3); + ADD2(tmp4, offset_vec, tmp5, offset_vec, tmp4, tmp5); + CLIP_SH4_0_255_MAX_SATU(tmp0, tmp1, tmp2, tmp3); + CLIP_SH2_0_255_MAX_SATU(tmp4, tmp5); + PCKEV_B3_UB(tmp1, tmp0, tmp3, tmp2, tmp5, tmp4, out0, out1, out2); + ST4x8_UB(out0, out1, dst, dst_stride); + ST2x4_UB(out2, 0, dst + 4, dst_stride); + dst += 4 * dst_stride; + ST2x4_UB(out2, 4, dst + 4, dst_stride); } static void hevc_hv_uniwgt_4t_8x2_msa(uint8_t *src, @@ -4451,23 +4561,24 @@ static void hevc_hv_uniwgt_4t_8x2_msa(uint8_t *src, int32_t dst_stride, const int8_t *filter_x, const int8_t *filter_y, - int32_t height, int32_t weight, int32_t offset, int32_t rnd_val) { + v16u8 out; v16i8 src0, src1, src2, src3, src4; v8i16 filt0, filt1; - v4i32 filt_h0, filt_h1; - v16i8 mask0 = { 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8 }; + v8i16 filt_h0, filt_h1, filter_vec; + v16i8 mask0 = LD_SB(ff_hevc_mask_arr); v16i8 mask1; - v8i16 filter_vec, const_vec; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; + v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7, vec8, vec9; v8i16 dst0, dst1, dst2, dst3, dst4; v4i32 dst0_r, dst0_l, dst1_r, dst1_l; v8i16 dst10_r, dst32_r, dst21_r, dst43_r; v8i16 dst10_l, dst32_l, dst21_l, dst43_l; - v4i32 weight_vec, offset_vec, rnd_vec; + v8i16 tmp0, tmp1; + v8i16 offset_vec, const_128, denom_vec; + v4i32 weight_vec, rnd_vec; src -= (src_stride + 1); @@ -4475,65 +4586,144 @@ static void hevc_hv_uniwgt_4t_8x2_msa(uint8_t *src, SPLATI_H2_SH(filter_vec, 0, 1, filt0, filt1); filter_vec = LD_SH(filter_y); - vec0 = __msa_clti_s_b((v16i8) filter_vec, 0); - filter_vec = (v8i16) __msa_ilvr_b(vec0, (v16i8) filter_vec); + UNPCK_R_SB_SH(filter_vec, filter_vec); - SPLATI_W2_SW(filter_vec, 0, filt_h0, filt_h1); + SPLATI_W2_SH(filter_vec, 0, filt_h0, filt_h1); mask1 = mask0 + 2; - const_vec = __msa_ldi_h(128); - const_vec <<= 6; - weight_vec = __msa_fill_w(weight); - offset_vec = __msa_fill_w(offset); rnd_vec = __msa_fill_w(rnd_val); - LD_SB3(src, src_stride, src0, src1, src2); - src += (3 * src_stride); - XORI_B3_128_SB(src0, src1, src2); + offset_vec = __msa_fill_h(offset); + denom_vec = __msa_fill_h(rnd_val - 6); + const_128 = __msa_fill_h((128 * weight)); + offset_vec += __msa_srar_h(const_128, denom_vec); + LD_SB5(src, src_stride, src0, src1, src2, src3, src4); + XORI_B5_128_SB(src0, src1, src2, src3, src4); VSHF_B2_SB(src0, src0, src0, src0, mask0, mask1, vec0, vec1); VSHF_B2_SB(src1, src1, src1, src1, mask0, mask1, vec2, vec3); VSHF_B2_SB(src2, src2, src2, src2, mask0, mask1, vec4, vec5); - dst0 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst0, dst0); - dst1 = const_vec; - DPADD_SB2_SH(vec2, vec3, filt0, filt1, dst1, dst1); - dst2 = const_vec; - DPADD_SB2_SH(vec4, vec5, filt0, filt1, dst2, dst2); - + VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec6, vec7); + VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec8, vec9); + dst0 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst1 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst2 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dst3 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); + dst4 = HEVC_FILT_4TAP_SH(vec8, vec9, filt0, filt1); ILVRL_H2_SH(dst1, dst0, dst10_r, dst10_l); ILVRL_H2_SH(dst2, dst1, dst21_r, dst21_l); - - LD_SB2(src, src_stride, src3, src4); - src += (2 * src_stride); - XORI_B2_128_SB(src3, src4); - - VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); - dst3 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst3, dst3); ILVRL_H2_SH(dst3, dst2, dst32_r, dst32_l); + ILVRL_H2_SH(dst4, dst3, dst43_r, dst43_l); dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); dst0_l = HEVC_FILT_4TAP(dst10_l, dst32_l, filt_h0, filt_h1); - dst0_r >>= 6; - dst0_l >>= 6; - - VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec0, vec1); - dst4 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst4, dst4); - ILVRL_H2_SH(dst4, dst3, dst43_r, dst43_l); dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); dst1_l = HEVC_FILT_4TAP(dst21_l, dst43_l, filt_h0, filt_h1); - dst1_r >>= 6; - dst1_l >>= 6; + SRA_4V(dst0_r, dst0_l, dst1_r, dst1_l, 6); + MUL2(dst0_r, weight_vec, dst1_r, weight_vec, dst0_r, dst1_r); + MUL2(dst0_l, weight_vec, dst1_l, weight_vec, dst0_l, dst1_l); + SRAR_W4_SW(dst0_r, dst0_l, dst1_r, dst1_l, rnd_vec); + PCKEV_H2_SH(dst0_l, dst0_r, dst1_l, dst1_r, tmp0, tmp1); + ADD2(tmp0, offset_vec, tmp1, offset_vec, tmp0, tmp1); + CLIP_SH2_0_255_MAX_SATU(tmp0, tmp1); + out = (v16u8) __msa_pckev_b((v16i8) tmp1, (v16i8) tmp0); + ST8x2_UB(out, dst, dst_stride); +} - HEVC_HV_UNIW_RND_CLIP4(dst0_r, dst1_r, dst0_l, dst1_l, - weight_vec, offset_vec, rnd_vec, - dst0_r, dst1_r, dst0_l, dst1_l); - HEVC_PCK_SW_SB4(dst0_l, dst0_r, dst1_l, dst1_r, dst0_r); - ST8x2_UB(dst0_r, dst, dst_stride); - dst += (2 * dst_stride); +static void hevc_hv_uniwgt_4t_8multx4_msa(uint8_t *src, + int32_t src_stride, + uint8_t *dst, + int32_t dst_stride, + const int8_t *filter_x, + const int8_t *filter_y, + int32_t width8mult, + int32_t weight, + int32_t offset, + int32_t rnd_val) +{ + uint32_t cnt; + v16u8 out0, out1; + v16i8 src0, src1, src2, src3, src4, src5, src6, mask0, mask1; + v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7; + v8i16 filt0, filt1, filt_h0, filt_h1, filter_vec; + v8i16 dst0, dst1, dst2, dst3, dst4, dst5, dst6, tmp0, tmp1, tmp2, tmp3; + v8i16 dst10_r, dst32_r, dst54_r, dst21_r, dst43_r, dst65_r; + v8i16 dst10_l, dst32_l, dst54_l, dst21_l, dst43_l, dst65_l; + v8i16 offset_vec, const_128, denom_vec; + v4i32 dst0_r, dst0_l, dst1_r, dst1_l, dst2_r, dst2_l, dst3_r, dst3_l; + v4i32 weight_vec, rnd_vec; + + src -= (src_stride + 1); + + filter_vec = LD_SH(filter_x); + SPLATI_H2_SH(filter_vec, 0, 1, filt0, filt1); + + filter_vec = LD_SH(filter_y); + UNPCK_R_SB_SH(filter_vec, filter_vec); + + SPLATI_W2_SH(filter_vec, 0, filt_h0, filt_h1); + + mask0 = LD_SB(ff_hevc_mask_arr); + mask1 = mask0 + 2; + + weight_vec = __msa_fill_w(weight); + rnd_vec = __msa_fill_w(rnd_val); + + offset_vec = __msa_fill_h(offset); + denom_vec = __msa_fill_h(rnd_val - 6); + const_128 = __msa_fill_h((128 * weight)); + offset_vec += __msa_srar_h(const_128, denom_vec); + + for (cnt = width8mult; cnt--;) { + LD_SB7(src, src_stride, src0, src1, src2, src3, src4, src5, src6); + src += 8; + XORI_B7_128_SB(src0, src1, src2, src3, src4, src5, src6); + VSHF_B2_SB(src0, src0, src0, src0, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src1, src1, src1, src1, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src2, src2, src2, src2, mask0, mask1, vec4, vec5); + dst0 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst1 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst2 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + ILVRL_H2_SH(dst1, dst0, dst10_r, dst10_l); + ILVRL_H2_SH(dst2, dst1, dst21_r, dst21_l); + VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec4, vec5); + VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec6, vec7); + dst3 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst4 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst5 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dst6 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); + ILVRL_H2_SH(dst3, dst2, dst32_r, dst32_l); + ILVRL_H2_SH(dst4, dst3, dst43_r, dst43_l); + ILVRL_H2_SH(dst5, dst4, dst54_r, dst54_l); + ILVRL_H2_SH(dst6, dst5, dst65_r, dst65_l); + dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); + dst0_l = HEVC_FILT_4TAP(dst10_l, dst32_l, filt_h0, filt_h1); + dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); + dst1_l = HEVC_FILT_4TAP(dst21_l, dst43_l, filt_h0, filt_h1); + dst2_r = HEVC_FILT_4TAP(dst32_r, dst54_r, filt_h0, filt_h1); + dst2_l = HEVC_FILT_4TAP(dst32_l, dst54_l, filt_h0, filt_h1); + dst3_r = HEVC_FILT_4TAP(dst43_r, dst65_r, filt_h0, filt_h1); + dst3_l = HEVC_FILT_4TAP(dst43_l, dst65_l, filt_h0, filt_h1); + SRA_4V(dst0_r, dst0_l, dst1_r, dst1_l, 6); + SRA_4V(dst2_r, dst2_l, dst3_r, dst3_l, 6); + MUL2(dst0_r, weight_vec, dst1_r, weight_vec, dst0_r, dst1_r); + MUL2(dst2_r, weight_vec, dst3_r, weight_vec, dst2_r, dst3_r); + MUL2(dst0_l, weight_vec, dst1_l, weight_vec, dst0_l, dst1_l); + MUL2(dst2_l, weight_vec, dst3_l, weight_vec, dst2_l, dst3_l); + SRAR_W4_SW(dst0_r, dst0_l, dst1_r, dst1_l, rnd_vec); + SRAR_W4_SW(dst2_r, dst2_l, dst3_r, dst3_l, rnd_vec); + PCKEV_H4_SH(dst0_l, dst0_r, dst1_l, dst1_r, dst2_l, dst2_r, dst3_l, + dst3_r, tmp0, tmp1, tmp2, tmp3); + ADD2(tmp0, offset_vec, tmp1, offset_vec, tmp0, tmp1); + ADD2(tmp2, offset_vec, tmp3, offset_vec, tmp2, tmp3); + CLIP_SH4_0_255_MAX_SATU(tmp0, tmp1, tmp2, tmp3); + PCKEV_B2_UB(tmp1, tmp0, tmp3, tmp2, out0, out1); + ST8x4_UB(out0, out1, dst, dst_stride); + dst += 8; + } } static void hevc_hv_uniwgt_4t_8x6_msa(uint8_t *src, @@ -4542,26 +4732,27 @@ static void hevc_hv_uniwgt_4t_8x6_msa(uint8_t *src, int32_t dst_stride, const int8_t *filter_x, const int8_t *filter_y, - int32_t height, int32_t weight, int32_t offset, int32_t rnd_val) { + v16u8 out0, out1, out2; v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8; v8i16 filt0, filt1; - v4i32 filt_h0, filt_h1; - v16i8 mask0 = { 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8 }; + v8i16 filt_h0, filt_h1, filter_vec; + v16i8 mask0 = LD_SB(ff_hevc_mask_arr); v16i8 mask1; - v8i16 filter_vec, const_vec; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; + v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7, vec8, vec9; + v16i8 vec10, vec11, vec12, vec13, vec14, vec15, vec16, vec17; v8i16 dst0, dst1, dst2, dst3, dst4, dst5, dst6, dst7, dst8; v4i32 dst0_r, dst0_l, dst1_r, dst1_l, dst2_r, dst2_l, dst3_r, dst3_l; - v4i32 dst4_r, dst4_l, dst5_r, dst5_l; + v4i32 dst4_r, dst4_l, dst5_r, dst5_l, weight_vec, rnd_vec; v8i16 dst10_r, dst32_r, dst10_l, dst32_l; v8i16 dst21_r, dst43_r, dst21_l, dst43_l; v8i16 dst54_r, dst54_l, dst65_r, dst65_l; v8i16 dst76_r, dst76_l, dst87_r, dst87_l; - v4i32 weight_vec, offset_vec, rnd_vec; + v8i16 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; + v8i16 offset_vec, const_128, denom_vec; src -= (src_stride + 1); @@ -4569,126 +4760,87 @@ static void hevc_hv_uniwgt_4t_8x6_msa(uint8_t *src, SPLATI_H2_SH(filter_vec, 0, 1, filt0, filt1); filter_vec = LD_SH(filter_y); - vec0 = __msa_clti_s_b((v16i8) filter_vec, 0); - filter_vec = (v8i16) __msa_ilvr_b(vec0, (v16i8) filter_vec); + UNPCK_R_SB_SH(filter_vec, filter_vec); - SPLATI_W2_SW(filter_vec, 0, filt_h0, filt_h1); + SPLATI_W2_SH(filter_vec, 0, filt_h0, filt_h1); mask1 = mask0 + 2; - const_vec = __msa_ldi_h(128); - const_vec <<= 6; - weight_vec = __msa_fill_w(weight); - offset_vec = __msa_fill_w(offset); rnd_vec = __msa_fill_w(rnd_val); - LD_SB3(src, src_stride, src0, src1, src2); - src += (3 * src_stride); - - XORI_B3_128_SB(src0, src1, src2); - + offset_vec = __msa_fill_h(offset); + denom_vec = __msa_fill_h(rnd_val - 6); + const_128 = __msa_fill_h((128 * weight)); + offset_vec += __msa_srar_h(const_128, denom_vec); + + LD_SB5(src, src_stride, src0, src1, src2, src3, src4); + src += (5 * src_stride); + LD_SB4(src, src_stride, src5, src6, src7, src8); + XORI_B5_128_SB(src0, src1, src2, src3, src4); + XORI_B4_128_SB(src5, src6, src7, src8); VSHF_B2_SB(src0, src0, src0, src0, mask0, mask1, vec0, vec1); VSHF_B2_SB(src1, src1, src1, src1, mask0, mask1, vec2, vec3); VSHF_B2_SB(src2, src2, src2, src2, mask0, mask1, vec4, vec5); - dst0 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst0, dst0); - dst1 = const_vec; - DPADD_SB2_SH(vec2, vec3, filt0, filt1, dst1, dst1); - dst2 = const_vec; - DPADD_SB2_SH(vec4, vec5, filt0, filt1, dst2, dst2); - + VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec6, vec7); + VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec8, vec9); + VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec10, vec11); + VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec12, vec13); + VSHF_B2_SB(src7, src7, src7, src7, mask0, mask1, vec14, vec15); + VSHF_B2_SB(src8, src8, src8, src8, mask0, mask1, vec16, vec17); + dst0 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst1 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst2 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dst3 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); + dst4 = HEVC_FILT_4TAP_SH(vec8, vec9, filt0, filt1); + dst5 = HEVC_FILT_4TAP_SH(vec10, vec11, filt0, filt1); + dst6 = HEVC_FILT_4TAP_SH(vec12, vec13, filt0, filt1); + dst7 = HEVC_FILT_4TAP_SH(vec14, vec15, filt0, filt1); + dst8 = HEVC_FILT_4TAP_SH(vec16, vec17, filt0, filt1); ILVRL_H2_SH(dst1, dst0, dst10_r, dst10_l); ILVRL_H2_SH(dst2, dst1, dst21_r, dst21_l); - - LD_SB2(src, src_stride, src3, src4); - src += (2 * src_stride); - XORI_B2_128_SB(src3, src4); - - /* row 3 */ - VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); - dst3 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst3, dst3); ILVRL_H2_SH(dst3, dst2, dst32_r, dst32_l); + ILVRL_H2_SH(dst4, dst3, dst43_r, dst43_l); + ILVRL_H2_SH(dst5, dst4, dst54_r, dst54_l); + ILVRL_H2_SH(dst6, dst5, dst65_r, dst65_l); + ILVRL_H2_SH(dst7, dst6, dst76_r, dst76_l); + ILVRL_H2_SH(dst8, dst7, dst87_r, dst87_l); dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); dst0_l = HEVC_FILT_4TAP(dst10_l, dst32_l, filt_h0, filt_h1); - dst0_r >>= 6; - dst0_l >>= 6; - - /* row 4 */ - VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec0, vec1); - dst4 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst4, dst4); - ILVRL_H2_SH(dst4, dst3, dst43_r, dst43_l); dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); dst1_l = HEVC_FILT_4TAP(dst21_l, dst43_l, filt_h0, filt_h1); - dst1_r >>= 6; - dst1_l >>= 6; - - LD_SB2(src, src_stride, src5, src6); - src += (2 * src_stride); - XORI_B2_128_SB(src5, src6); - - /* row 5 */ - VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec0, vec1); - dst5 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst5, dst5); - ILVRL_H2_SH(dst5, dst4, dst54_r, dst54_l); dst2_r = HEVC_FILT_4TAP(dst32_r, dst54_r, filt_h0, filt_h1); dst2_l = HEVC_FILT_4TAP(dst32_l, dst54_l, filt_h0, filt_h1); - dst2_r >>= 6; - dst2_l >>= 6; - - /* row 6 */ - VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec0, vec1); - dst6 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst6, dst6); - ILVRL_H2_SH(dst6, dst5, dst65_r, dst65_l); dst3_r = HEVC_FILT_4TAP(dst43_r, dst65_r, filt_h0, filt_h1); dst3_l = HEVC_FILT_4TAP(dst43_l, dst65_l, filt_h0, filt_h1); - dst3_r >>= 6; - dst3_l >>= 6; - - LD_SB2(src, src_stride, src7, src8); - src += (2 * src_stride); - XORI_B2_128_SB(src7, src8); - - /* row 7 */ - VSHF_B2_SB(src7, src7, src7, src7, mask0, mask1, vec0, vec1); - dst7 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst7, dst7); - ILVRL_H2_SH(dst7, dst6, dst76_r, dst76_l); dst4_r = HEVC_FILT_4TAP(dst54_r, dst76_r, filt_h0, filt_h1); dst4_l = HEVC_FILT_4TAP(dst54_l, dst76_l, filt_h0, filt_h1); - - dst4_r >>= 6; - dst4_l >>= 6; - - /* row 8 */ - VSHF_B2_SB(src8, src8, src8, src8, mask0, mask1, vec0, vec1); - dst8 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst8, dst8); - ILVRL_H2_SH(dst8, dst7, dst87_r, dst87_l); dst5_r = HEVC_FILT_4TAP(dst65_r, dst87_r, filt_h0, filt_h1); dst5_l = HEVC_FILT_4TAP(dst65_l, dst87_l, filt_h0, filt_h1); - dst5_r >>= 6; - dst5_l >>= 6; - - HEVC_HV_UNIW_RND_CLIP4(dst0_r, dst1_r, dst0_l, dst1_l, - weight_vec, offset_vec, rnd_vec, - dst0_r, dst1_r, dst0_l, dst1_l); - HEVC_HV_UNIW_RND_CLIP4(dst2_r, dst3_r, dst2_l, dst3_l, - weight_vec, offset_vec, rnd_vec, - dst2_r, dst3_r, dst2_l, dst3_l); - HEVC_HV_UNIW_RND_CLIP4(dst4_r, dst5_r, dst4_l, dst5_l, - weight_vec, offset_vec, rnd_vec, - dst4_r, dst5_r, dst4_l, dst5_l); - HEVC_PCK_SW_SB12(dst0_l, dst0_r, dst1_l, dst1_r, - dst2_l, dst2_r, dst3_l, dst3_r, - dst4_l, dst4_r, dst5_l, dst5_r, dst0_r, dst1_r, dst2_r); - ST8x4_UB(dst0_r, dst1_r, dst, dst_stride); + SRA_4V(dst0_r, dst0_l, dst1_r, dst1_l, 6); + SRA_4V(dst2_r, dst2_l, dst3_r, dst3_l, 6); + SRA_4V(dst4_r, dst4_l, dst5_r, dst5_l, 6); + MUL2(dst0_r, weight_vec, dst1_r, weight_vec, dst0_r, dst1_r); + MUL2(dst2_r, weight_vec, dst3_r, weight_vec, dst2_r, dst3_r); + MUL2(dst4_r, weight_vec, dst5_r, weight_vec, dst4_r, dst5_r); + MUL2(dst0_l, weight_vec, dst1_l, weight_vec, dst0_l, dst1_l); + MUL2(dst2_l, weight_vec, dst3_l, weight_vec, dst2_l, dst3_l); + MUL2(dst4_l, weight_vec, dst5_l, weight_vec, dst4_l, dst5_l); + SRAR_W4_SW(dst0_r, dst0_l, dst1_r, dst1_l, rnd_vec); + SRAR_W4_SW(dst2_r, dst2_l, dst3_r, dst3_l, rnd_vec); + SRAR_W4_SW(dst4_r, dst4_l, dst5_r, dst5_l, rnd_vec); + PCKEV_H4_SH(dst0_l, dst0_r, dst1_l, dst1_r, dst2_l, dst2_r, dst3_l, dst3_r, + tmp0, tmp1, tmp2, tmp3); + PCKEV_H2_SH(dst4_l, dst4_r, dst5_l, dst5_r, tmp4, tmp5); + ADD2(tmp0, offset_vec, tmp1, offset_vec, tmp0, tmp1); + ADD2(tmp2, offset_vec, tmp3, offset_vec, tmp2, tmp3); + ADD2(tmp4, offset_vec, tmp5, offset_vec, tmp4, tmp5); + CLIP_SH4_0_255_MAX_SATU(tmp0, tmp1, tmp2, tmp3); + CLIP_SH2_0_255_MAX_SATU(tmp4, tmp5); + PCKEV_B3_UB(tmp1, tmp0, tmp3, tmp2, tmp5, tmp4, out0, out1, out2); + ST8x4_UB(out0, out1, dst, dst_stride); dst += (4 * dst_stride); - ST8x2_UB(dst2_r, dst, dst_stride); + ST8x2_UB(out2, dst, dst_stride); } static void hevc_hv_uniwgt_4t_8multx4mult_msa(uint8_t *src, @@ -4701,24 +4853,25 @@ static void hevc_hv_uniwgt_4t_8multx4mult_msa(uint8_t *src, int32_t weight, int32_t offset, int32_t rnd_val, - int32_t width) + int32_t width8mult) { uint32_t loop_cnt, cnt; uint8_t *src_tmp; uint8_t *dst_tmp; + v16u8 out0, out1; v16i8 src0, src1, src2, src3, src4, src5, src6; v8i16 filt0, filt1; - v4i32 filt_h0, filt_h1; - v16i8 mask0 = { 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8 }; + v8i16 filt_h0, filt_h1, filter_vec; + v16i8 mask0 = LD_SB(ff_hevc_mask_arr); v16i8 mask1; - v8i16 filter_vec, const_vec; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v8i16 dst0, dst1, dst2, dst3, dst4, dst5; + v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7; + v8i16 dst0, dst1, dst2, dst3, dst4, dst5, dst6, tmp0, tmp1, tmp2, tmp3; + v8i16 dst10_r, dst32_r, dst54_r, dst21_r, dst43_r, dst65_r; + v8i16 dst10_l, dst32_l, dst54_l, dst21_l, dst43_l, dst65_l; v4i32 dst0_r, dst0_l, dst1_r, dst1_l; - v4i32 weight_vec, offset_vec, rnd_vec; + v8i16 offset_vec, const_128, denom_vec; v4i32 dst2_r, dst2_l, dst3_r, dst3_l; - v8i16 dst10_r, dst32_r, dst21_r, dst43_r; - v8i16 dst10_l, dst32_l, dst21_l, dst43_l; + v4i32 weight_vec, rnd_vec; src -= (src_stride + 1); @@ -4726,21 +4879,21 @@ static void hevc_hv_uniwgt_4t_8multx4mult_msa(uint8_t *src, SPLATI_H2_SH(filter_vec, 0, 1, filt0, filt1); filter_vec = LD_SH(filter_y); - vec0 = __msa_clti_s_b((v16i8) filter_vec, 0); - filter_vec = (v8i16) __msa_ilvr_b(vec0, (v16i8) filter_vec); + UNPCK_R_SB_SH(filter_vec, filter_vec); - SPLATI_W2_SW(filter_vec, 0, filt_h0, filt_h1); + SPLATI_W2_SH(filter_vec, 0, filt_h0, filt_h1); mask1 = mask0 + 2; - const_vec = __msa_ldi_h(128); - const_vec <<= 6; - weight_vec = __msa_fill_w(weight); - offset_vec = __msa_fill_w(offset); rnd_vec = __msa_fill_w(rnd_val); - for (cnt = width >> 3; cnt--;) { + offset_vec = __msa_fill_h(offset); + denom_vec = __msa_fill_h(rnd_val - 6); + const_128 = __msa_fill_h((128 * weight)); + offset_vec += __msa_srar_h(const_128, denom_vec); + + for (cnt = width8mult; cnt--;) { src_tmp = src; dst_tmp = dst; @@ -4751,12 +4904,9 @@ static void hevc_hv_uniwgt_4t_8multx4mult_msa(uint8_t *src, VSHF_B2_SB(src0, src0, src0, src0, mask0, mask1, vec0, vec1); VSHF_B2_SB(src1, src1, src1, src1, mask0, mask1, vec2, vec3); VSHF_B2_SB(src2, src2, src2, src2, mask0, mask1, vec4, vec5); - dst0 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst0, dst0); - dst1 = const_vec; - DPADD_SB2_SH(vec2, vec3, filt0, filt1, dst1, dst1); - dst2 = const_vec; - DPADD_SB2_SH(vec4, vec5, filt0, filt1, dst2, dst2); + dst0 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst1 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst2 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); ILVRL_H2_SH(dst1, dst0, dst10_r, dst10_l); ILVRL_H2_SH(dst2, dst1, dst21_r, dst21_l); @@ -4767,51 +4917,47 @@ static void hevc_hv_uniwgt_4t_8multx4mult_msa(uint8_t *src, XORI_B4_128_SB(src3, src4, src5, src6); VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); - dst3 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst3, dst3); + VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec4, vec5); + VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec6, vec7); + dst3 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst4 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst5 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dst6 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); ILVRL_H2_SH(dst3, dst2, dst32_r, dst32_l); + ILVRL_H2_SH(dst4, dst3, dst43_r, dst43_l); + ILVRL_H2_SH(dst5, dst4, dst54_r, dst54_l); + ILVRL_H2_SH(dst6, dst5, dst65_r, dst65_l); dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); dst0_l = HEVC_FILT_4TAP(dst10_l, dst32_l, filt_h0, filt_h1); - dst0_r >>= 6; - dst0_l >>= 6; - - VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec0, vec1); - dst4 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst4, dst4); - ILVRL_H2_SH(dst4, dst3, dst43_r, dst43_l); dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); dst1_l = HEVC_FILT_4TAP(dst21_l, dst43_l, filt_h0, filt_h1); - dst1_r >>= 6; - dst1_l >>= 6; - - VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec0, vec1); - dst5 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst5, dst5); - ILVRL_H2_SH(dst5, dst4, dst10_r, dst10_l); - dst2_r = HEVC_FILT_4TAP(dst32_r, dst10_r, filt_h0, filt_h1); - dst2_l = HEVC_FILT_4TAP(dst32_l, dst10_l, filt_h0, filt_h1); - dst2_r >>= 6; - dst2_l >>= 6; - - VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec0, vec1); - dst2 = const_vec; - DPADD_SB2_SH(vec0, vec1, filt0, filt1, dst2, dst2); - ILVRL_H2_SH(dst2, dst5, dst21_r, dst21_l); - dst3_r = HEVC_FILT_4TAP(dst43_r, dst21_r, filt_h0, filt_h1); - dst3_l = HEVC_FILT_4TAP(dst43_l, dst21_l, filt_h0, filt_h1); - dst3_r >>= 6; - dst3_l >>= 6; - - HEVC_HV_UNIW_RND_CLIP4(dst0_r, dst1_r, dst0_l, dst1_l, - weight_vec, offset_vec, rnd_vec, - dst0_r, dst1_r, dst0_l, dst1_l); - HEVC_HV_UNIW_RND_CLIP4(dst2_r, dst3_r, dst2_l, dst3_l, - weight_vec, offset_vec, rnd_vec, - dst2_r, dst3_r, dst2_l, dst3_l); - HEVC_PCK_SW_SB8(dst0_l, dst0_r, dst1_l, dst1_r, - dst2_l, dst2_r, dst3_l, dst3_r, dst0_r, dst1_r); - ST8x4_UB(dst0_r, dst1_r, dst_tmp, dst_stride); + dst2_r = HEVC_FILT_4TAP(dst32_r, dst54_r, filt_h0, filt_h1); + dst2_l = HEVC_FILT_4TAP(dst32_l, dst54_l, filt_h0, filt_h1); + dst3_r = HEVC_FILT_4TAP(dst43_r, dst65_r, filt_h0, filt_h1); + dst3_l = HEVC_FILT_4TAP(dst43_l, dst65_l, filt_h0, filt_h1); + SRA_4V(dst0_r, dst0_l, dst1_r, dst1_l, 6); + SRA_4V(dst2_r, dst2_l, dst3_r, dst3_l, 6); + MUL2(dst0_r, weight_vec, dst1_r, weight_vec, dst0_r, dst1_r); + MUL2(dst2_r, weight_vec, dst3_r, weight_vec, dst2_r, dst3_r); + MUL2(dst0_l, weight_vec, dst1_l, weight_vec, dst0_l, dst1_l); + MUL2(dst2_l, weight_vec, dst3_l, weight_vec, dst2_l, dst3_l); + SRAR_W4_SW(dst0_r, dst0_l, dst1_r, dst1_l, rnd_vec); + SRAR_W4_SW(dst2_r, dst2_l, dst3_r, dst3_l, rnd_vec); + PCKEV_H4_SH(dst0_l, dst0_r, dst1_l, dst1_r, dst2_l, dst2_r, dst3_l, + dst3_r, tmp0, tmp1, tmp2, tmp3); + ADD2(tmp0, offset_vec, tmp1, offset_vec, tmp0, tmp1); + ADD2(tmp2, offset_vec, tmp3, offset_vec, tmp2, tmp3); + CLIP_SH4_0_255_MAX_SATU(tmp0, tmp1, tmp2, tmp3); + PCKEV_B2_UB(tmp1, tmp0, tmp3, tmp2, out0, out1); + ST8x4_UB(out0, out1, dst_tmp, dst_stride); dst_tmp += (4 * dst_stride); + + dst10_r = dst54_r; + dst10_l = dst54_l; + dst21_r = dst65_r; + dst21_l = dst65_l; + dst2 = dst6; } src += 8; @@ -4833,16 +4979,20 @@ static void hevc_hv_uniwgt_4t_8w_msa(uint8_t *src, if (2 == height) { hevc_hv_uniwgt_4t_8x2_msa(src, src_stride, dst, dst_stride, - filter_x, filter_y, height, weight, + filter_x, filter_y, weight, offset, rnd_val); + } else if (4 == height) { + hevc_hv_uniwgt_4t_8multx4_msa(src, src_stride, dst, dst_stride, + filter_x, filter_y, 1, weight, + offset, rnd_val); } else if (6 == height) { hevc_hv_uniwgt_4t_8x6_msa(src, src_stride, dst, dst_stride, - filter_x, filter_y, height, weight, + filter_x, filter_y, weight, offset, rnd_val); } else if (0 == (height % 4)) { hevc_hv_uniwgt_4t_8multx4mult_msa(src, src_stride, dst, dst_stride, filter_x, filter_y, height, weight, - offset, rnd_val, 8); + offset, rnd_val, 1); } } @@ -4857,12 +5007,170 @@ static void hevc_hv_uniwgt_4t_12w_msa(uint8_t *src, int32_t offset, int32_t rnd_val) { - hevc_hv_uniwgt_4t_8multx4mult_msa(src, src_stride, dst, dst_stride, - filter_x, filter_y, height, weight, - offset, rnd_val, 8); - hevc_hv_uniwgt_4t_4w_msa(src + 8, src_stride, dst + 8, dst_stride, - filter_x, filter_y, height, weight, - offset, rnd_val); + uint32_t loop_cnt; + uint8_t *src_tmp, *dst_tmp; + v16u8 out0, out1; + v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8, src9, src10; + v16i8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7; + v16i8 mask0, mask1, mask2, mask3; + v8i16 filt0, filt1, filt_h0, filt_h1, filter_vec, tmp0, tmp1, tmp2, tmp3; + v8i16 dsth0, dsth1, dsth2, dsth3, dsth4, dsth5, dsth6; + v8i16 dst10, dst21, dst22, dst73, dst84, dst95, dst106; + v8i16 dst76_r, dst98_r, dst87_r, dst109_r; + v8i16 dst10_r, dst32_r, dst54_r, dst21_r, dst43_r, dst65_r; + v8i16 dst10_l, dst32_l, dst54_l, dst21_l, dst43_l, dst65_l; + v8i16 offset_vec, const_128, denom_vec; + v4i32 dst0_r, dst0_l, dst1_r, dst1_l, dst2_r, dst2_l, dst3_r, dst3_l; + v4i32 dst0, dst1, dst2, dst3, dst4, dst5, dst6, dst7, weight_vec, rnd_vec; + + src -= (src_stride + 1); + + filter_vec = LD_SH(filter_x); + SPLATI_H2_SH(filter_vec, 0, 1, filt0, filt1); + + filter_vec = LD_SH(filter_y); + UNPCK_R_SB_SH(filter_vec, filter_vec); + + SPLATI_W2_SH(filter_vec, 0, filt_h0, filt_h1); + + mask0 = LD_SB(ff_hevc_mask_arr); + mask1 = mask0 + 2; + + weight_vec = __msa_fill_w(weight); + rnd_vec = __msa_fill_w(rnd_val); + + offset_vec = __msa_fill_h(offset); + denom_vec = __msa_fill_h(rnd_val - 6); + const_128 = __msa_fill_h((128 * weight)); + offset_vec += __msa_srar_h(const_128, denom_vec); + + src_tmp = src; + dst_tmp = dst; + + LD_SB3(src_tmp, src_stride, src0, src1, src2); + src_tmp += (3 * src_stride); + XORI_B3_128_SB(src0, src1, src2); + VSHF_B2_SB(src0, src0, src0, src0, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src1, src1, src1, src1, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src2, src2, src2, src2, mask0, mask1, vec4, vec5); + dsth0 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dsth1 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dsth2 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + ILVRL_H2_SH(dsth1, dsth0, dst10_r, dst10_l); + ILVRL_H2_SH(dsth2, dsth1, dst21_r, dst21_l); + + for (loop_cnt = 4; loop_cnt--;) { + LD_SB4(src_tmp, src_stride, src3, src4, src5, src6); + src_tmp += (4 * src_stride); + XORI_B4_128_SB(src3, src4, src5, src6); + VSHF_B2_SB(src3, src3, src3, src3, mask0, mask1, vec0, vec1); + VSHF_B2_SB(src4, src4, src4, src4, mask0, mask1, vec2, vec3); + VSHF_B2_SB(src5, src5, src5, src5, mask0, mask1, vec4, vec5); + VSHF_B2_SB(src6, src6, src6, src6, mask0, mask1, vec6, vec7); + dsth3 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dsth4 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dsth5 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dsth6 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); + ILVRL_H2_SH(dsth3, dsth2, dst32_r, dst32_l); + ILVRL_H2_SH(dsth4, dsth3, dst43_r, dst43_l); + ILVRL_H2_SH(dsth5, dsth4, dst54_r, dst54_l); + ILVRL_H2_SH(dsth6, dsth5, dst65_r, dst65_l); + dst0_r = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); + dst0_l = HEVC_FILT_4TAP(dst10_l, dst32_l, filt_h0, filt_h1); + dst1_r = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); + dst1_l = HEVC_FILT_4TAP(dst21_l, dst43_l, filt_h0, filt_h1); + dst2_r = HEVC_FILT_4TAP(dst32_r, dst54_r, filt_h0, filt_h1); + dst2_l = HEVC_FILT_4TAP(dst32_l, dst54_l, filt_h0, filt_h1); + dst3_r = HEVC_FILT_4TAP(dst43_r, dst65_r, filt_h0, filt_h1); + dst3_l = HEVC_FILT_4TAP(dst43_l, dst65_l, filt_h0, filt_h1); + SRA_4V(dst0_r, dst0_l, dst1_r, dst1_l, 6); + SRA_4V(dst2_r, dst2_l, dst3_r, dst3_l, 6); + MUL2(dst0_r, weight_vec, dst1_r, weight_vec, dst0_r, dst1_r); + MUL2(dst2_r, weight_vec, dst3_r, weight_vec, dst2_r, dst3_r); + MUL2(dst0_l, weight_vec, dst1_l, weight_vec, dst0_l, dst1_l); + MUL2(dst2_l, weight_vec, dst3_l, weight_vec, dst2_l, dst3_l); + SRAR_W4_SW(dst0_r, dst0_l, dst1_r, dst1_l, rnd_vec); + SRAR_W4_SW(dst2_r, dst2_l, dst3_r, dst3_l, rnd_vec); + PCKEV_H4_SH(dst0_l, dst0_r, dst1_l, dst1_r, dst2_l, dst2_r, dst3_l, + dst3_r, tmp0, tmp1, tmp2, tmp3); + ADD2(tmp0, offset_vec, tmp1, offset_vec, tmp0, tmp1); + ADD2(tmp2, offset_vec, tmp3, offset_vec, tmp2, tmp3); + CLIP_SH4_0_255_MAX_SATU(tmp0, tmp1, tmp2, tmp3); + PCKEV_B2_UB(tmp1, tmp0, tmp3, tmp2, out0, out1); + ST8x4_UB(out0, out1, dst_tmp, dst_stride); + dst_tmp += (4 * dst_stride); + + dst10_r = dst54_r; + dst10_l = dst54_l; + dst21_r = dst65_r; + dst21_l = dst65_l; + dsth2 = dsth6; + } + + src += 8; + dst += 8; + + mask2 = LD_SB(ff_hevc_mask_arr + 16); + mask3 = mask2 + 2; + + LD_SB3(src, src_stride, src0, src1, src2); + src += (3 * src_stride); + XORI_B3_128_SB(src0, src1, src2); + VSHF_B2_SB(src0, src1, src0, src1, mask2, mask3, vec0, vec1); + VSHF_B2_SB(src1, src2, src1, src2, mask2, mask3, vec2, vec3); + dst10 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst21 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + ILVRL_H2_SH(dst21, dst10, dst10_r, dst21_r); + dst22 = (v8i16) __msa_splati_d((v2i64) dst21, 1); + + for (loop_cnt = 2; loop_cnt--;) { + LD_SB8(src, src_stride, src3, src4, src5, src6, src7, src8, src9, + src10); + src += (8 * src_stride); + XORI_B8_128_SB(src3, src4, src5, src6, src7, src8, src9, src10); + VSHF_B2_SB(src3, src7, src3, src7, mask2, mask3, vec0, vec1); + VSHF_B2_SB(src4, src8, src4, src8, mask2, mask3, vec2, vec3); + VSHF_B2_SB(src5, src9, src5, src9, mask2, mask3, vec4, vec5); + VSHF_B2_SB(src6, src10, src6, src10, mask2, mask3, vec6, vec7); + dst73 = HEVC_FILT_4TAP_SH(vec0, vec1, filt0, filt1); + dst84 = HEVC_FILT_4TAP_SH(vec2, vec3, filt0, filt1); + dst95 = HEVC_FILT_4TAP_SH(vec4, vec5, filt0, filt1); + dst106 = HEVC_FILT_4TAP_SH(vec6, vec7, filt0, filt1); + dst32_r = __msa_ilvr_h(dst73, dst22); + ILVRL_H2_SH(dst84, dst73, dst43_r, dst87_r); + ILVRL_H2_SH(dst95, dst84, dst54_r, dst98_r); + ILVRL_H2_SH(dst106, dst95, dst65_r, dst109_r); + dst22 = (v8i16) __msa_splati_d((v2i64) dst73, 1); + dst76_r = __msa_ilvr_h(dst22, dst106); + dst0 = HEVC_FILT_4TAP(dst10_r, dst32_r, filt_h0, filt_h1); + dst1 = HEVC_FILT_4TAP(dst21_r, dst43_r, filt_h0, filt_h1); + dst2 = HEVC_FILT_4TAP(dst32_r, dst54_r, filt_h0, filt_h1); + dst3 = HEVC_FILT_4TAP(dst43_r, dst65_r, filt_h0, filt_h1); + dst4 = HEVC_FILT_4TAP(dst54_r, dst76_r, filt_h0, filt_h1); + dst5 = HEVC_FILT_4TAP(dst65_r, dst87_r, filt_h0, filt_h1); + dst6 = HEVC_FILT_4TAP(dst76_r, dst98_r, filt_h0, filt_h1); + dst7 = HEVC_FILT_4TAP(dst87_r, dst109_r, filt_h0, filt_h1); + SRA_4V(dst0, dst1, dst2, dst3, 6); + SRA_4V(dst4, dst5, dst6, dst7, 6); + MUL2(dst0, weight_vec, dst1, weight_vec, dst0, dst1); + MUL2(dst2, weight_vec, dst3, weight_vec, dst2, dst3); + MUL2(dst4, weight_vec, dst5, weight_vec, dst4, dst5); + MUL2(dst6, weight_vec, dst7, weight_vec, dst6, dst7); + SRAR_W4_SW(dst0, dst1, dst2, dst3, rnd_vec); + SRAR_W4_SW(dst4, dst5, dst6, dst7, rnd_vec); + PCKEV_H4_SH(dst1, dst0, dst3, dst2, dst5, dst4, dst7, dst6, tmp0, tmp1, + tmp2, tmp3); + ADD2(tmp0, offset_vec, tmp1, offset_vec, tmp0, tmp1); + ADD2(tmp2, offset_vec, tmp3, offset_vec, tmp2, tmp3); + CLIP_SH4_0_255_MAX_SATU(tmp0, tmp1, tmp2, tmp3); + PCKEV_B2_UB(tmp1, tmp0, tmp3, tmp2, out0, out1); + ST4x8_UB(out0, out1, dst, dst_stride); + dst += (8 * dst_stride); + + dst10_r = dst98_r; + dst21_r = dst109_r; + dst22 = (v8i16) __msa_splati_d((v2i64) dst106, 1); + } } static void hevc_hv_uniwgt_4t_16w_msa(uint8_t *src, @@ -4876,9 +5184,15 @@ static void hevc_hv_uniwgt_4t_16w_msa(uint8_t *src, int32_t offset, int32_t rnd_val) { - hevc_hv_uniwgt_4t_8multx4mult_msa(src, src_stride, dst, dst_stride, - filter_x, filter_y, height, weight, - offset, rnd_val, 16); + if (4 == height) { + hevc_hv_uniwgt_4t_8multx4_msa(src, src_stride, dst, dst_stride, + filter_x, filter_y, 2, weight, offset, + rnd_val); + } else { + hevc_hv_uniwgt_4t_8multx4mult_msa(src, src_stride, dst, dst_stride, + filter_x, filter_y, height, weight, + offset, rnd_val, 2); + } } static void hevc_hv_uniwgt_4t_24w_msa(uint8_t *src, @@ -4894,7 +5208,7 @@ static void hevc_hv_uniwgt_4t_24w_msa(uint8_t *src, { hevc_hv_uniwgt_4t_8multx4mult_msa(src, src_stride, dst, dst_stride, filter_x, filter_y, height, weight, - offset, rnd_val, 24); + offset, rnd_val, 3); } static void hevc_hv_uniwgt_4t_32w_msa(uint8_t *src, @@ -4910,7 +5224,7 @@ static void hevc_hv_uniwgt_4t_32w_msa(uint8_t *src, { hevc_hv_uniwgt_4t_8multx4mult_msa(src, src_stride, dst, dst_stride, filter_x, filter_y, height, weight, - offset, rnd_val, 32); + offset, rnd_val, 4); } #define UNIWGT_MC_COPY(WIDTH) \

[FFmpeg-devel] avcodec/mips: Improve hevc uni weighted hv mc msa functions

Commit Message

Patch