diff options
| -rw-r--r-- | libavcodec/mips/h264qpel_msa.c | 746 | ||||
| -rw-r--r-- | libavcodec/mips/hevc_mc_uniw_msa.c | 67 | ||||
| -rw-r--r-- | libavcodec/mips/hevcdsp_msa.c | 50 |
3 files changed, 0 insertions, 863 deletions
diff --git a/libavcodec/mips/h264qpel_msa.c b/libavcodec/mips/h264qpel_msa.c index 0b42bc4..afc0183 100644 --- a/libavcodec/mips/h264qpel_msa.c +++ b/libavcodec/mips/h264qpel_msa.c @@ -159,300 +159,6 @@ static const uint8_t luma_mask_arr[16 * 8] = { out0_m; \ } ) -static void avc_luma_hz_4w_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height) -{ - uint32_t loop_cnt; - v16i8 src0, src1, src2, src3; - v8i16 res0, res1; - v16u8 out; - v16i8 mask0, mask1, mask2; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v16i8 minus5b = __msa_ldi_b(-5); - v16i8 plus20b = __msa_ldi_b(20); - - LD_SB3(&luma_mask_arr[48], 16, mask0, mask1, mask2); - for (loop_cnt = (height >> 2); loop_cnt--;) { - LD_SB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - - XORI_B4_128_SB(src0, src1, src2, src3); - VSHF_B2_SB(src0, src1, src2, src3, mask0, mask0, vec0, vec1); - HADD_SB2_SH(vec0, vec1, res0, res1); - VSHF_B2_SB(src0, src1, src2, src3, mask1, mask1, vec2, vec3); - DPADD_SB2_SH(vec2, vec3, minus5b, minus5b, res0, res1); - VSHF_B2_SB(src0, src1, src2, src3, mask2, mask2, vec4, vec5); - DPADD_SB2_SH(vec4, vec5, plus20b, plus20b, res0, res1); - SRARI_H2_SH(res0, res1, 5); - SAT_SH2_SH(res0, res1, 7); - out = PCKEV_XORI128_UB(res0, res1); - ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride); - dst += (4 * dst_stride); - } -} - -static void avc_luma_hz_8w_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height) -{ - uint32_t loop_cnt; - v16i8 src0, src1, src2, src3; - v8i16 res0, res1, res2, res3; - v16i8 mask0, mask1, mask2; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v16i8 vec6, vec7, vec8, vec9, vec10, vec11; - v16i8 minus5b = __msa_ldi_b(-5); - v16i8 plus20b = __msa_ldi_b(20); - v16u8 out0, out1; - - LD_SB3(&luma_mask_arr[0], 16, mask0, mask1, mask2); - - for (loop_cnt = (height >> 2); loop_cnt--;) { - LD_SB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - - XORI_B4_128_SB(src0, src1, src2, src3); - VSHF_B2_SB(src0, src0, src1, src1, mask0, mask0, vec0, vec1); - VSHF_B2_SB(src2, src2, src3, src3, mask0, mask0, vec2, vec3); - HADD_SB4_SH(vec0, vec1, vec2, vec3, res0, res1, res2, res3); - VSHF_B2_SB(src0, src0, src1, src1, mask1, mask1, vec4, vec5); - VSHF_B2_SB(src2, src2, src3, src3, mask1, mask1, vec6, vec7); - DPADD_SB4_SH(vec4, vec5, vec6, vec7, minus5b, minus5b, minus5b, minus5b, - res0, res1, res2, res3); - VSHF_B2_SB(src0, src0, src1, src1, mask2, mask2, vec8, vec9); - VSHF_B2_SB(src2, src2, src3, src3, mask2, mask2, vec10, vec11); - DPADD_SB4_SH(vec8, vec9, vec10, vec11, plus20b, plus20b, plus20b, - plus20b, res0, res1, res2, res3); - SRARI_H4_SH(res0, res1, res2, res3, 5); - SAT_SH4_SH(res0, res1, res2, res3, 7); - out0 = PCKEV_XORI128_UB(res0, res1); - out1 = PCKEV_XORI128_UB(res2, res3); - ST8x4_UB(out0, out1, dst, dst_stride); - dst += (4 * dst_stride); - } -} - -static void avc_luma_hz_16w_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height) -{ - uint32_t loop_cnt; - v16i8 src0, src1, src2, src3, src4, src5, src6, src7; - v8i16 res0, res1, res2, res3, res4, res5, res6, res7; - v16i8 mask0, mask1, mask2; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v16i8 vec6, vec7, vec8, vec9, vec10, vec11; - v16i8 minus5b = __msa_ldi_b(-5); - v16i8 plus20b = __msa_ldi_b(20); - - LD_SB3(&luma_mask_arr[0], 16, mask0, mask1, mask2); - - for (loop_cnt = (height >> 2); loop_cnt--;) { - LD_SB2(src, 8, src0, src1); - src += src_stride; - LD_SB2(src, 8, src2, src3); - src += src_stride; - - XORI_B4_128_SB(src0, src1, src2, src3); - VSHF_B2_SB(src0, src0, src1, src1, mask0, mask0, vec0, vec3); - VSHF_B2_SB(src2, src2, src3, src3, mask0, mask0, vec6, vec9); - VSHF_B2_SB(src0, src0, src1, src1, mask1, mask1, vec1, vec4); - VSHF_B2_SB(src2, src2, src3, src3, mask1, mask1, vec7, vec10); - VSHF_B2_SB(src0, src0, src1, src1, mask2, mask2, vec2, vec5); - VSHF_B2_SB(src2, src2, src3, src3, mask2, mask2, vec8, vec11); - HADD_SB4_SH(vec0, vec3, vec6, vec9, res0, res1, res2, res3); - DPADD_SB4_SH(vec1, vec4, vec7, vec10, minus5b, minus5b, minus5b, - minus5b, res0, res1, res2, res3); - DPADD_SB4_SH(vec2, vec5, vec8, vec11, plus20b, plus20b, plus20b, - plus20b, res0, res1, res2, res3); - - LD_SB2(src, 8, src4, src5); - src += src_stride; - LD_SB2(src, 8, src6, src7); - src += src_stride; - - XORI_B4_128_SB(src4, src5, src6, src7); - VSHF_B2_SB(src4, src4, src5, src5, mask0, mask0, vec0, vec3); - VSHF_B2_SB(src6, src6, src7, src7, mask0, mask0, vec6, vec9); - VSHF_B2_SB(src4, src4, src5, src5, mask1, mask1, vec1, vec4); - VSHF_B2_SB(src6, src6, src7, src7, mask1, mask1, vec7, vec10); - VSHF_B2_SB(src4, src4, src5, src5, mask2, mask2, vec2, vec5); - VSHF_B2_SB(src6, src6, src7, src7, mask2, mask2, vec8, vec11); - HADD_SB4_SH(vec0, vec3, vec6, vec9, res4, res5, res6, res7); - DPADD_SB4_SH(vec1, vec4, vec7, vec10, minus5b, minus5b, minus5b, - minus5b, res4, res5, res6, res7); - DPADD_SB4_SH(vec2, vec5, vec8, vec11, plus20b, plus20b, plus20b, - plus20b, res4, res5, res6, res7); - SRARI_H4_SH(res0, res1, res2, res3, 5); - SRARI_H4_SH(res4, res5, res6, res7, 5); - SAT_SH4_SH(res0, res1, res2, res3, 7); - SAT_SH4_SH(res4, res5, res6, res7, 7); - PCKEV_B4_SB(res1, res0, res3, res2, res5, res4, res7, res6, - vec0, vec1, vec2, vec3); - XORI_B4_128_SB(vec0, vec1, vec2, vec3); - - ST_SB4(vec0, vec1, vec2, vec3, dst, dst_stride); - dst += (4 * dst_stride); - } -} - -static void avc_luma_hz_qrt_4w_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height, uint8_t hor_offset) -{ - uint8_t slide; - uint32_t loop_cnt; - v16i8 src0, src1, src2, src3; - v8i16 res0, res1; - v16i8 res, mask0, mask1, mask2; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v16i8 minus5b = __msa_ldi_b(-5); - v16i8 plus20b = __msa_ldi_b(20); - - LD_SB3(&luma_mask_arr[48], 16, mask0, mask1, mask2); - slide = 2 + hor_offset; - - for (loop_cnt = (height >> 2); loop_cnt--;) { - LD_SB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - - XORI_B4_128_SB(src0, src1, src2, src3); - VSHF_B2_SB(src0, src1, src2, src3, mask0, mask0, vec0, vec1); - HADD_SB2_SH(vec0, vec1, res0, res1); - VSHF_B2_SB(src0, src1, src2, src3, mask1, mask1, vec2, vec3); - DPADD_SB2_SH(vec2, vec3, minus5b, minus5b, res0, res1); - VSHF_B2_SB(src0, src1, src2, src3, mask2, mask2, vec4, vec5); - DPADD_SB2_SH(vec4, vec5, plus20b, plus20b, res0, res1); - SRARI_H2_SH(res0, res1, 5); - SAT_SH2_SH(res0, res1, 7); - - res = __msa_pckev_b((v16i8) res1, (v16i8) res0); - src0 = __msa_sld_b(src0, src0, slide); - src1 = __msa_sld_b(src1, src1, slide); - src2 = __msa_sld_b(src2, src2, slide); - src3 = __msa_sld_b(src3, src3, slide); - src0 = (v16i8) __msa_insve_w((v4i32) src0, 1, (v4i32) src1); - src1 = (v16i8) __msa_insve_w((v4i32) src2, 1, (v4i32) src3); - src0 = (v16i8) __msa_insve_d((v2i64) src0, 1, (v2i64) src1); - res = __msa_aver_s_b(res, src0); - res = (v16i8) __msa_xori_b((v16u8) res, 128); - - ST4x4_UB(res, res, 0, 1, 2, 3, dst, dst_stride); - dst += (4 * dst_stride); - } -} - -static void avc_luma_hz_qrt_8w_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height, uint8_t hor_offset) -{ - uint8_t slide; - uint32_t loop_cnt; - v16i8 src0, src1, src2, src3; - v16i8 tmp0, tmp1; - v8i16 res0, res1, res2, res3; - v16i8 mask0, mask1, mask2; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v16i8 vec6, vec7, vec8, vec9, vec10, vec11; - v16i8 minus5b = __msa_ldi_b(-5); - v16i8 plus20b = __msa_ldi_b(20); - - LD_SB3(&luma_mask_arr[0], 16, mask0, mask1, mask2); - slide = 2 + hor_offset; - - for (loop_cnt = height >> 2; loop_cnt--;) { - LD_SB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - - XORI_B4_128_SB(src0, src1, src2, src3); - VSHF_B2_SB(src0, src0, src1, src1, mask0, mask0, vec0, vec1); - VSHF_B2_SB(src2, src2, src3, src3, mask0, mask0, vec2, vec3); - HADD_SB4_SH(vec0, vec1, vec2, vec3, res0, res1, res2, res3); - VSHF_B2_SB(src0, src0, src1, src1, mask1, mask1, vec4, vec5); - VSHF_B2_SB(src2, src2, src3, src3, mask1, mask1, vec6, vec7); - DPADD_SB4_SH(vec4, vec5, vec6, vec7, minus5b, minus5b, minus5b, minus5b, - res0, res1, res2, res3); - VSHF_B2_SB(src0, src0, src1, src1, mask2, mask2, vec8, vec9); - VSHF_B2_SB(src2, src2, src3, src3, mask2, mask2, vec10, vec11); - DPADD_SB4_SH(vec8, vec9, vec10, vec11, plus20b, plus20b, plus20b, - plus20b, res0, res1, res2, res3); - - src0 = __msa_sld_b(src0, src0, slide); - src1 = __msa_sld_b(src1, src1, slide); - src2 = __msa_sld_b(src2, src2, slide); - src3 = __msa_sld_b(src3, src3, slide); - - SRARI_H4_SH(res0, res1, res2, res3, 5); - SAT_SH4_SH(res0, res1, res2, res3, 7); - PCKEV_B2_SB(res1, res0, res3, res2, tmp0, tmp1); - PCKEV_D2_SB(src1, src0, src3, src2, src0, src1); - - tmp0 = __msa_aver_s_b(tmp0, src0); - tmp1 = __msa_aver_s_b(tmp1, src1); - - XORI_B2_128_SB(tmp0, tmp1); - ST8x4_UB(tmp0, tmp1, dst, dst_stride); - - dst += (4 * dst_stride); - } -} - -static void avc_luma_hz_qrt_16w_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height, uint8_t hor_offset) -{ - uint32_t loop_cnt; - v16i8 dst0, dst1; - v16i8 src0, src1, src2, src3; - v16i8 mask0, mask1, mask2, vshf; - v8i16 res0, res1, res2, res3; - v16i8 vec0, vec1, vec2, vec3, vec4, vec5; - v16i8 vec6, vec7, vec8, vec9, vec10, vec11; - v16i8 minus5b = __msa_ldi_b(-5); - v16i8 plus20b = __msa_ldi_b(20); - - LD_SB3(&luma_mask_arr[0], 16, mask0, mask1, mask2); - - if (hor_offset) { - vshf = LD_SB(&luma_mask_arr[16 + 96]); - } else { - vshf = LD_SB(&luma_mask_arr[96]); - } - - for (loop_cnt = height >> 1; loop_cnt--;) { - LD_SB2(src, 8, src0, src1); - src += src_stride; - LD_SB2(src, 8, src2, src3); - src += src_stride; - - XORI_B4_128_SB(src0, src1, src2, src3); - VSHF_B2_SB(src0, src0, src1, src1, mask0, mask0, vec0, vec3); - VSHF_B2_SB(src2, src2, src3, src3, mask0, mask0, vec6, vec9); - VSHF_B2_SB(src0, src0, src1, src1, mask1, mask1, vec1, vec4); - VSHF_B2_SB(src2, src2, src3, src3, mask1, mask1, vec7, vec10); - VSHF_B2_SB(src0, src0, src1, src1, mask2, mask2, vec2, vec5); - VSHF_B2_SB(src2, src2, src3, src3, mask2, mask2, vec8, vec11); - HADD_SB4_SH(vec0, vec3, vec6, vec9, res0, res1, res2, res3); - DPADD_SB4_SH(vec1, vec4, vec7, vec10, minus5b, minus5b, minus5b, - minus5b, res0, res1, res2, res3); - DPADD_SB4_SH(vec2, vec5, vec8, vec11, plus20b, plus20b, plus20b, - plus20b, res0, res1, res2, res3); - VSHF_B2_SB(src0, src1, src2, src3, vshf, vshf, src0, src2); - SRARI_H4_SH(res0, res1, res2, res3, 5); - SAT_SH4_SH(res0, res1, res2, res3, 7); - PCKEV_B2_SB(res1, res0, res3, res2, dst0, dst1); - - dst0 = __msa_aver_s_b(dst0, src0); - dst1 = __msa_aver_s_b(dst1, src2); - - XORI_B2_128_SB(dst0, dst1); - - ST_SB2(dst0, dst1, dst, dst_stride); - dst += (2 * dst_stride); - } -} - static void avc_luma_vt_4w_msa(const uint8_t *src, int32_t src_stride, uint8_t *dst, int32_t dst_stride, int32_t height) @@ -623,217 +329,6 @@ static void avc_luma_vt_16w_msa(const uint8_t *src, int32_t src_stride, } } -static void avc_luma_vt_qrt_4w_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height, uint8_t ver_offset) -{ - int32_t loop_cnt; - int16_t filt_const0 = 0xfb01; - int16_t filt_const1 = 0x1414; - int16_t filt_const2 = 0x1fb; - v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8; - v16i8 src10_r, src32_r, src54_r, src76_r, src21_r, src43_r, src65_r; - v16i8 src87_r, src2110, src4332, src6554, src8776; - v8i16 out10, out32; - v16i8 filt0, filt1, filt2; - v16u8 out; - - filt0 = (v16i8) __msa_fill_h(filt_const0); - filt1 = (v16i8) __msa_fill_h(filt_const1); - filt2 = (v16i8) __msa_fill_h(filt_const2); - - LD_SB5(src, src_stride, src0, src1, src2, src3, src4); - src += (5 * src_stride); - - ILVR_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3, - src10_r, src21_r, src32_r, src43_r); - ILVR_D2_SB(src21_r, src10_r, src43_r, src32_r, src2110, src4332); - XORI_B2_128_SB(src2110, src4332); - - for (loop_cnt = (height >> 2); loop_cnt--;) { - LD_SB4(src, src_stride, src5, src6, src7, src8); - src += (4 * src_stride); - - ILVR_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, - src54_r, src65_r, src76_r, src87_r); - ILVR_D2_SB(src65_r, src54_r, src87_r, src76_r, src6554, src8776); - XORI_B2_128_SB(src6554, src8776); - out10 = DPADD_SH3_SH(src2110, src4332, src6554, filt0, filt1, filt2); - out32 = DPADD_SH3_SH(src4332, src6554, src8776, filt0, filt1, filt2); - SRARI_H2_SH(out10, out32, 5); - SAT_SH2_SH(out10, out32, 7); - - out = PCKEV_XORI128_UB(out10, out32); - - if (ver_offset) { - src32_r = (v16i8) __msa_insve_w((v4i32) src3, 1, (v4i32) src4); - src54_r = (v16i8) __msa_insve_w((v4i32) src5, 1, (v4i32) src6); - } else { - src32_r = (v16i8) __msa_insve_w((v4i32) src2, 1, (v4i32) src3); - src54_r = (v16i8) __msa_insve_w((v4i32) src4, 1, (v4i32) src5); - } - - src32_r = (v16i8) __msa_insve_d((v2i64) src32_r, 1, (v2i64) src54_r); - out = __msa_aver_u_b(out, (v16u8) src32_r); - - ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride); - dst += (4 * dst_stride); - src2110 = src6554; - src4332 = src8776; - src2 = src6; - src3 = src7; - src4 = src8; - } -} - -static void avc_luma_vt_qrt_8w_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height, uint8_t ver_offset) -{ - int32_t loop_cnt; - int16_t filt_const0 = 0xfb01; - int16_t filt_const1 = 0x1414; - int16_t filt_const2 = 0x1fb; - v16i8 src0, src1, src2, src3, src4, src7, src8, src9, src10; - v16i8 src10_r, src32_r, src76_r, src98_r; - v16i8 src21_r, src43_r, src87_r, src109_r; - v8i16 out0_r, out1_r, out2_r, out3_r; - v16i8 res0, res1; - v16i8 filt0, filt1, filt2; - - filt0 = (v16i8) __msa_fill_h(filt_const0); - filt1 = (v16i8) __msa_fill_h(filt_const1); - filt2 = (v16i8) __msa_fill_h(filt_const2); - - LD_SB5(src, src_stride, src0, src1, src2, src3, src4); - src += (5 * src_stride); - - XORI_B5_128_SB(src0, src1, src2, src3, src4); - ILVR_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3, - src10_r, src21_r, src32_r, src43_r); - - 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); - ILVR_B4_SB(src7, src4, src8, src7, src9, src8, src10, src9, - src76_r, src87_r, src98_r, src109_r); - out0_r = DPADD_SH3_SH(src10_r, src32_r, src76_r, filt0, filt1, filt2); - out1_r = DPADD_SH3_SH(src21_r, src43_r, src87_r, filt0, filt1, filt2); - out2_r = DPADD_SH3_SH(src32_r, src76_r, src98_r, filt0, filt1, filt2); - out3_r = DPADD_SH3_SH(src43_r, src87_r, src109_r, filt0, filt1, filt2); - SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, 5); - SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7); - PCKEV_B2_SB(out1_r, out0_r, out3_r, out2_r, res0, res1); - - if (ver_offset) { - PCKEV_D2_SB(src4, src3, src8, src7, src10_r, src32_r); - } else { - PCKEV_D2_SB(src3, src2, src7, src4, src10_r, src32_r); - } - - res0 = __msa_aver_s_b(res0, (v16i8) src10_r); - res1 = __msa_aver_s_b(res1, (v16i8) src32_r); - - XORI_B2_128_SB(res0, res1); - ST8x4_UB(res0, res1, dst, dst_stride); - - dst += (4 * dst_stride); - src10_r = src76_r; - src32_r = src98_r; - src21_r = src87_r; - src43_r = src109_r; - src2 = src8; - src3 = src9; - src4 = src10; - } -} - -static void avc_luma_vt_qrt_16w_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height, uint8_t ver_offset) -{ - int32_t loop_cnt; - int16_t filt_const0 = 0xfb01; - int16_t filt_const1 = 0x1414; - int16_t filt_const2 = 0x1fb; - v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8; - v16i8 src10_r, src32_r, src54_r, src76_r, src21_r, src43_r, src65_r; - v16i8 src87_r, src10_l, src32_l, src54_l, src76_l, src21_l, src43_l; - v16i8 src65_l, src87_l; - v8i16 out0_r, out1_r, out2_r, out3_r, out0_l, out1_l, out2_l, out3_l; - v16u8 res0, res1, res2, res3; - v16i8 filt0, filt1, filt2; - - filt0 = (v16i8) __msa_fill_h(filt_const0); - filt1 = (v16i8) __msa_fill_h(filt_const1); - filt2 = (v16i8) __msa_fill_h(filt_const2); - - LD_SB5(src, src_stride, src0, src1, src2, src3, src4); - src += (5 * src_stride); - - XORI_B5_128_SB(src0, src1, src2, src3, src4); - ILVR_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3, - src10_r, src21_r, src32_r, src43_r); - ILVL_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3, - src10_l, src21_l, src32_l, src43_l); - - for (loop_cnt = (height >> 2); loop_cnt--;) { - LD_SB4(src, src_stride, src5, src6, src7, src8); - src += (4 * src_stride); - - XORI_B4_128_SB(src5, src6, src7, src8); - ILVR_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, - src54_r, src65_r, src76_r, src87_r); - ILVL_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, - src54_l, src65_l, src76_l, src87_l); - out0_r = DPADD_SH3_SH(src10_r, src32_r, src54_r, filt0, filt1, filt2); - out1_r = DPADD_SH3_SH(src21_r, src43_r, src65_r, filt0, filt1, filt2); - out2_r = DPADD_SH3_SH(src32_r, src54_r, src76_r, filt0, filt1, filt2); - out3_r = DPADD_SH3_SH(src43_r, src65_r, src87_r, filt0, filt1, filt2); - out0_l = DPADD_SH3_SH(src10_l, src32_l, src54_l, filt0, filt1, filt2); - out1_l = DPADD_SH3_SH(src21_l, src43_l, src65_l, filt0, filt1, filt2); - out2_l = DPADD_SH3_SH(src32_l, src54_l, src76_l, filt0, filt1, filt2); - out3_l = DPADD_SH3_SH(src43_l, src65_l, src87_l, filt0, filt1, filt2); - SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, 5); - SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7); - SRARI_H4_SH(out0_l, out1_l, out2_l, out3_l, 5); - SAT_SH4_SH(out0_l, out1_l, out2_l, out3_l, 7); - PCKEV_B4_UB(out0_l, out0_r, out1_l, out1_r, out2_l, out2_r, out3_l, - out3_r, res0, res1, res2, res3); - - if (ver_offset) { - res0 = (v16u8) __msa_aver_s_b((v16i8) res0, src3); - res1 = (v16u8) __msa_aver_s_b((v16i8) res1, src4); - res2 = (v16u8) __msa_aver_s_b((v16i8) res2, src5); - res3 = (v16u8) __msa_aver_s_b((v16i8) res3, src6); - } else { - res0 = (v16u8) __msa_aver_s_b((v16i8) res0, src2); - res1 = (v16u8) __msa_aver_s_b((v16i8) res1, src3); - res2 = (v16u8) __msa_aver_s_b((v16i8) res2, src4); - res3 = (v16u8) __msa_aver_s_b((v16i8) res3, src5); - } - - XORI_B4_128_UB(res0, res1, res2, res3); - ST_UB4(res0, res1, res2, res3, dst, dst_stride); - - dst += (4 * dst_stride); - - src10_r = src54_r; - src32_r = src76_r; - src21_r = src65_r; - src43_r = src87_r; - src10_l = src54_l; - src32_l = src76_l; - src21_l = src65_l; - src43_l = src87_l; - src2 = src6; - src3 = src7; - src4 = src8; - } -} - static void avc_luma_mid_4w_msa(const uint8_t *src, int32_t src_stride, uint8_t *dst, int32_t dst_stride, int32_t height) @@ -2733,247 +2228,6 @@ static void avc_luma_hv_qrt_and_aver_dst_16x16_msa(const uint8_t *src_x, } } -static void copy_width8_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height) -{ - int32_t cnt; - uint64_t out0, out1, out2, out3, out4, out5, out6, out7; - v16u8 src0, src1, src2, src3, src4, src5, src6, src7; - - if (0 == height % 12) { - for (cnt = (height / 12); cnt--;) { - LD_UB8(src, src_stride, - src0, src1, src2, src3, src4, src5, src6, src7); - src += (8 * src_stride); - - out0 = __msa_copy_u_d((v2i64) src0, 0); - out1 = __msa_copy_u_d((v2i64) src1, 0); - out2 = __msa_copy_u_d((v2i64) src2, 0); - out3 = __msa_copy_u_d((v2i64) src3, 0); - out4 = __msa_copy_u_d((v2i64) src4, 0); - out5 = __msa_copy_u_d((v2i64) src5, 0); - out6 = __msa_copy_u_d((v2i64) src6, 0); - out7 = __msa_copy_u_d((v2i64) src7, 0); - - SD4(out0, out1, out2, out3, dst, dst_stride); - dst += (4 * dst_stride); - SD4(out4, out5, out6, out7, dst, dst_stride); - dst += (4 * dst_stride); - - LD_UB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - - out0 = __msa_copy_u_d((v2i64) src0, 0); - out1 = __msa_copy_u_d((v2i64) src1, 0); - out2 = __msa_copy_u_d((v2i64) src2, 0); - out3 = __msa_copy_u_d((v2i64) src3, 0); - - SD4(out0, out1, out2, out3, dst, dst_stride); - dst += (4 * dst_stride); - } - } else if (0 == height % 8) { - for (cnt = height >> 3; cnt--;) { - LD_UB8(src, src_stride, - src0, src1, src2, src3, src4, src5, src6, src7); - src += (8 * src_stride); - - out0 = __msa_copy_u_d((v2i64) src0, 0); - out1 = __msa_copy_u_d((v2i64) src1, 0); - out2 = __msa_copy_u_d((v2i64) src2, 0); - out3 = __msa_copy_u_d((v2i64) src3, 0); - out4 = __msa_copy_u_d((v2i64) src4, 0); - out5 = __msa_copy_u_d((v2i64) src5, 0); - out6 = __msa_copy_u_d((v2i64) src6, 0); - out7 = __msa_copy_u_d((v2i64) src7, 0); - - SD4(out0, out1, out2, out3, dst, dst_stride); - dst += (4 * dst_stride); - SD4(out4, out5, out6, out7, dst, dst_stride); - dst += (4 * dst_stride); - } - } else if (0 == height % 4) { - for (cnt = (height / 4); cnt--;) { - LD_UB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - out0 = __msa_copy_u_d((v2i64) src0, 0); - out1 = __msa_copy_u_d((v2i64) src1, 0); - out2 = __msa_copy_u_d((v2i64) src2, 0); - out3 = __msa_copy_u_d((v2i64) src3, 0); - - SD4(out0, out1, out2, out3, dst, dst_stride); - dst += (4 * dst_stride); - } - } else if (0 == height % 2) { - for (cnt = (height / 2); cnt--;) { - LD_UB2(src, src_stride, src0, src1); - src += (2 * src_stride); - out0 = __msa_copy_u_d((v2i64) src0, 0); - out1 = __msa_copy_u_d((v2i64) src1, 0); - - SD(out0, dst); - dst += dst_stride; - SD(out1, dst); - dst += dst_stride; - } - } -} - -static void copy_16multx8mult_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height, int32_t width) -{ - int32_t cnt, loop_cnt; - const uint8_t *src_tmp; - uint8_t *dst_tmp; - v16u8 src0, src1, src2, src3, src4, src5, src6, src7; - - for (cnt = (width >> 4); cnt--;) { - src_tmp = src; - dst_tmp = dst; - - for (loop_cnt = (height >> 3); loop_cnt--;) { - LD_UB8(src_tmp, src_stride, - src0, src1, src2, src3, src4, src5, src6, src7); - src_tmp += (8 * src_stride); - - ST_UB8(src0, src1, src2, src3, src4, src5, src6, src7, - dst_tmp, dst_stride); - dst_tmp += (8 * dst_stride); - } - - src += 16; - dst += 16; - } -} - -static void copy_width16_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height) -{ - int32_t cnt; - v16u8 src0, src1, src2, src3, src4, src5, src6, src7; - - if (0 == height % 12) { - for (cnt = (height / 12); cnt--;) { - LD_UB8(src, src_stride, - src0, src1, src2, src3, src4, src5, src6, src7); - src += (8 * src_stride); - ST_UB8(src0, src1, src2, src3, src4, src5, src6, src7, - dst, dst_stride); - dst += (8 * dst_stride); - - LD_UB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - ST_UB4(src0, src1, src2, src3, dst, dst_stride); - dst += (4 * dst_stride); - } - } else if (0 == height % 8) { - copy_16multx8mult_msa(src, src_stride, dst, dst_stride, height, 16); - } else if (0 == height % 4) { - for (cnt = (height >> 2); cnt--;) { - LD_UB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - - ST_UB4(src0, src1, src2, src3, dst, dst_stride); - dst += (4 * dst_stride); - } - } -} - -static void avg_width4_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height) -{ - int32_t cnt; - uint32_t out0, out1, out2, out3; - v16u8 src0, src1, src2, src3; - v16u8 dst0, dst1, dst2, dst3; - - if (0 == (height % 4)) { - for (cnt = (height / 4); cnt--;) { - LD_UB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - - LD_UB4(dst, dst_stride, dst0, dst1, dst2, dst3); - - AVER_UB4_UB(src0, dst0, src1, dst1, src2, dst2, src3, dst3, - dst0, dst1, dst2, dst3); - - out0 = __msa_copy_u_w((v4i32) dst0, 0); - out1 = __msa_copy_u_w((v4i32) dst1, 0); - out2 = __msa_copy_u_w((v4i32) dst2, 0); - out3 = __msa_copy_u_w((v4i32) dst3, 0); - SW4(out0, out1, out2, out3, dst, dst_stride); - dst += (4 * dst_stride); - } - } else if (0 == (height % 2)) { - for (cnt = (height / 2); cnt--;) { - LD_UB2(src, src_stride, src0, src1); - src += (2 * src_stride); - - LD_UB2(dst, dst_stride, dst0, dst1); - - AVER_UB2_UB(src0, dst0, src1, dst1, dst0, dst1); - - out0 = __msa_copy_u_w((v4i32) dst0, 0); - out1 = __msa_copy_u_w((v4i32) dst1, 0); - SW(out0, dst); - dst += dst_stride; - SW(out1, dst); - dst += dst_stride; - } - } -} - -static void avg_width8_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height) -{ - int32_t cnt; - uint64_t out0, out1, out2, out3; - v16u8 src0, src1, src2, src3; - v16u8 dst0, dst1, dst2, dst3; - - for (cnt = (height / 4); cnt--;) { - LD_UB4(src, src_stride, src0, src1, src2, src3); - src += (4 * src_stride); - LD_UB4(dst, dst_stride, dst0, dst1, dst2, dst3); - - AVER_UB4_UB(src0, dst0, src1, dst1, src2, dst2, src3, dst3, - dst0, dst1, dst2, dst3); - - out0 = __msa_copy_u_d((v2i64) dst0, 0); - out1 = __msa_copy_u_d((v2i64) dst1, 0); - out2 = __msa_copy_u_d((v2i64) dst2, 0); - out3 = __msa_copy_u_d((v2i64) dst3, 0); - SD4(out0, out1, out2, out3, dst, dst_stride); - dst += (4 * dst_stride); - } -} - -static void avg_width16_msa(const uint8_t *src, int32_t src_stride, - uint8_t *dst, int32_t dst_stride, - int32_t height) -{ - int32_t cnt; - v16u8 src0, src1, src2, src3, src4, src5, src6, src7; - v16u8 dst0, dst1, dst2, dst3, dst4, dst5, dst6, dst7; - - for (cnt = (height / 8); cnt--;) { - LD_UB8(src, src_stride, src0, src1, src2, src3, src4, src5, src6, src7); - src += (8 * src_stride); - LD_UB8(dst, dst_stride, dst0, dst1, dst2, dst3, dst4, dst5, dst6, dst7); - - AVER_UB4_UB(src0, dst0, src1, dst1, src2, dst2, src3, dst3, - dst0, dst1, dst2, dst3); - AVER_UB4_UB(src4, dst4, src5, dst5, src6, dst6, src7, dst7, - dst4, dst5, dst6, dst7); - ST_UB8(dst0, dst1, dst2, dst3, dst4, dst5, dst6, dst7, dst, dst_stride); - dst += (8 * dst_stride); - } -} - void ff_put_h264_qpel16_mc00_msa(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) { diff --git a/libavcodec/mips/hevc_mc_uniw_msa.c b/libavcodec/mips/hevc_mc_uniw_msa.c index d184419..38a8844 100644 --- a/libavcodec/mips/hevc_mc_uniw_msa.c +++ b/libavcodec/mips/hevc_mc_uniw_msa.c @@ -351,73 +351,6 @@ static void hevc_uniwgt_copy_12w_msa(uint8_t *src, } } -static void hevc_uniwgt_copy_16multx4mult_msa(uint8_t *src, - int32_t src_stride, - uint8_t *dst, - int32_t dst_stride, - int32_t height, - int32_t weight, - int32_t offset, - int32_t rnd_val, - int32_t width) -{ - uint32_t loop_cnt, cnt; - uint8_t *src_tmp; - uint8_t *dst_tmp; - v16i8 src0, src1, src2, src3; - v8i16 tmp0, tmp1, tmp2, tmp3; - v4i32 dst0_r, dst1_r, dst2_r, dst3_r, dst0_l, dst1_l, dst2_l, dst3_l; - v16i8 zero = { 0 }; - v4i32 weight_vec, offset_vec, rnd_vec; - - weight = weight & 0x0000FFFF; - weight_vec = __msa_fill_w(weight); - offset_vec = __msa_fill_w(offset); - rnd_vec = __msa_fill_w(rnd_val); - - for (cnt = width >> 4; cnt--;) { - src_tmp = src; - dst_tmp = dst; - - for (loop_cnt = height >> 2; loop_cnt--;) { - LD_SB4(src_tmp, src_stride, src0, src1, src2, src3); - src_tmp += (4 * src_stride); - ILVR_B2_SH(zero, src0, zero, src1, tmp0, tmp1); - ILVL_B2_SH(zero, src0, zero, src1, tmp2, tmp3); - - SLLI_4V(tmp0, tmp1, tmp2, tmp3, 6); - HEVC_UNIW_RND_CLIP4(tmp0, tmp1, tmp2, tmp3, - weight_vec, offset_vec, rnd_vec, - dst0_r, dst1_r, dst2_r, dst3_r, - dst0_l, dst1_l, dst2_l, dst3_l); - - HEVC_PCK_SW_SB8(dst0_l, dst0_r, dst2_l, dst2_r, - dst1_l, dst1_r, dst3_l, dst3_r, dst0_r, dst1_r); - - ST_SW2(dst0_r, dst1_r, dst_tmp, dst_stride); - dst_tmp += (2 * dst_stride); - - ILVR_B2_SH(zero, src2, zero, src3, tmp0, tmp1); - ILVL_B2_SH(zero, src2, zero, src3, tmp2, tmp3); - - SLLI_4V(tmp0, tmp1, tmp2, tmp3, 6); - HEVC_UNIW_RND_CLIP4(tmp0, tmp1, tmp2, tmp3, - weight_vec, offset_vec, rnd_vec, - dst0_r, dst1_r, dst2_r, dst3_r, - dst0_l, dst1_l, dst2_l, dst3_l); - - HEVC_PCK_SW_SB8(dst0_l, dst0_r, dst2_l, dst2_r, - dst1_l, dst1_r, dst3_l, dst3_r, dst0_r, dst1_r); - - ST_SW2(dst0_r, dst1_r, dst_tmp, dst_stride); - dst_tmp += (2 * dst_stride); - } - - src += 16; - dst += 16; - } -} - static void hevc_uniwgt_copy_16w_msa(uint8_t *src, int32_t src_stride, uint8_t *dst, diff --git a/libavcodec/mips/hevcdsp_msa.c b/libavcodec/mips/hevcdsp_msa.c index 1a854b2..73cc3ea 100644 --- a/libavcodec/mips/hevcdsp_msa.c +++ b/libavcodec/mips/hevcdsp_msa.c @@ -192,56 +192,6 @@ static void hevc_copy_12w_msa(uint8_t *src, int32_t src_stride, } } -static void hevc_copy_16multx8mult_msa(uint8_t *src, - int32_t src_stride, - int16_t *dst, - int32_t dst_stride, - int32_t height, - int32_t width) -{ - uint8_t *src_tmp; - int16_t *dst_tmp; - uint32_t loop_cnt, cnt; - v16i8 zero = { 0 }; - v16i8 src0, src1, src2, src3, src4, src5, src6, src7; - v8i16 in0_r, in1_r, in2_r, in3_r; - v8i16 in0_l, in1_l, in2_l, in3_l; - - for (cnt = (width >> 4); cnt--;) { - src_tmp = src; - dst_tmp = dst; - - for (loop_cnt = (height >> 3); loop_cnt--;) { - LD_SB8(src_tmp, src_stride, - src0, src1, src2, src3, src4, src5, src6, src7); - src_tmp += (8 * src_stride); - - ILVR_B4_SH(zero, src0, zero, src1, zero, src2, zero, src3, - in0_r, in1_r, in2_r, in3_r); - ILVL_B4_SH(zero, src0, zero, src1, zero, src2, zero, src3, - in0_l, in1_l, in2_l, in3_l); - SLLI_4V(in0_r, in1_r, in2_r, in3_r, 6); - SLLI_4V(in0_l, in1_l, in2_l, in3_l, 6); - ST_SH4(in0_r, in1_r, in2_r, in3_r, dst_tmp, dst_stride); - ST_SH4(in0_l, in1_l, in2_l, in3_l, (dst_tmp + 8), dst_stride); - dst_tmp += (4 * dst_stride); - - ILVR_B4_SH(zero, src4, zero, src5, zero, src6, zero, src7, - in0_r, in1_r, in2_r, in3_r); - ILVL_B4_SH(zero, src4, zero, src5, zero, src6, zero, src7, - in0_l, in1_l, in2_l, in3_l); - SLLI_4V(in0_r, in1_r, in2_r, in3_r, 6); - SLLI_4V(in0_l, in1_l, in2_l, in3_l, 6); - ST_SH4(in0_r, in1_r, in2_r, in3_r, dst_tmp, dst_stride); - ST_SH4(in0_l, in1_l, in2_l, in3_l, (dst_tmp + 8), dst_stride); - dst_tmp += (4 * dst_stride); - } - - src += 16; - dst += 16; - } -} - static void hevc_copy_16w_msa(uint8_t *src, int32_t src_stride, int16_t *dst, int32_t dst_stride, int32_t height) |
