/third_party/ffmpeg/libavcodec/mips/ |
H A D | vp3dsp_idct_msa.c | 29 v4i32 r0_r, r0_l, r1_r, r1_l, r2_r, r2_l, r3_r, r3_l, in idct_msa() 31 v4i32 A, B, C, D, Ad, Bd, Cd, Dd, E, F, G, H; in idct_msa() 32 v4i32 Ed, Gd, Add, Bdd, Fd, Hd; in idct_msa() 35 v4i32 c0, c1, c2, c3, c4, c5, c6, c7; in idct_msa() 36 v4i32 f0, f1, f2, f3, f4, f5, f6, f7; in idct_msa() 37 v4i32 sign_t; in idct_msa() 40 v4i32 cnst64277w = {64277, 64277, 64277, 64277}; in idct_msa() 41 v4i32 cnst60547w = {60547, 60547, 60547, 60547}; in idct_msa() 42 v4i32 cnst54491w = {54491, 54491, 54491, 54491}; in idct_msa() 43 v4i32 cnst46341 in idct_msa() [all...] |
H A D | vc1dsp_msa.c | 31 v4i32 in_r0, in_r1, in_r2, in_r3, in_r4, in_r5, in_r6, in_r7; in ff_vc1_inv_trans_8x8_msa() 32 v4i32 in_l0, in_l1, in_l2, in_l3, in_l4, in_l5, in_l6, in_l7; in ff_vc1_inv_trans_8x8_msa() 33 v4i32 t_r1, t_r2, t_r3, t_r4, t_r5, t_r6, t_r7, t_r8; in ff_vc1_inv_trans_8x8_msa() 34 v4i32 t_l1, t_l2, t_l3, t_l4, t_l5, t_l6, t_l7, t_l8; in ff_vc1_inv_trans_8x8_msa() 35 v4i32 cnst_12 = {12, 12, 12, 12}; in ff_vc1_inv_trans_8x8_msa() 36 v4i32 cnst_4 = {4, 4, 4, 4}; in ff_vc1_inv_trans_8x8_msa() 37 v4i32 cnst_16 = {16, 16, 16, 16}; in ff_vc1_inv_trans_8x8_msa() 38 v4i32 cnst_6 = {6, 6, 6, 6}; in ff_vc1_inv_trans_8x8_msa() 39 v4i32 cnst_15 = {15, 15, 15, 15}; in ff_vc1_inv_trans_8x8_msa() 40 v4i32 cnst_ in ff_vc1_inv_trans_8x8_msa() [all...] |
H A D | simple_idct_msa.c | 31 v4i32 temp0_r, temp1_r, temp2_r, temp3_r; in simple_idct_msa() 32 v4i32 temp0_l, temp1_l, temp2_l, temp3_l; in simple_idct_msa() 33 v4i32 a0_r, a1_r, a2_r, a3_r, a0_l, a1_l, a2_l, a3_l; in simple_idct_msa() 34 v4i32 b0_r, b1_r, b2_r, b3_r, b0_l, b1_l, b2_l, b3_l; in simple_idct_msa() 35 v4i32 w2, w4, w6; in simple_idct_msa() 38 v4i32 const_val0 = __msa_ldi_w(1); in simple_idct_msa() 39 v4i32 const_val1 = __msa_ldi_w(1); in simple_idct_msa() 53 w2 = (v4i32) __msa_splati_h(weights, 2); in simple_idct_msa() 54 w2 = (v4i32) __msa_ilvr_h(zero, (v8i16) w2); in simple_idct_msa() 55 w4 = (v4i32) __msa_splati_ in simple_idct_msa() [all...] |
H A D | vp9_intra_msa.c | 123 sum_w = (v4u32) __msa_srari_w((v4i32) sum_d, 3); in ff_dc_4x4_msa() 125 val0 = __msa_copy_u_w((v4i32) store, 0); in ff_dc_4x4_msa() 141 data = (v16i8) __msa_insert_w((v4i32) data, 0, val0); \ 144 sum_w = (v4u32) __msa_srari_w((v4i32) sum_w, 2); \ 146 val0 = __msa_copy_u_w((v4i32) store, 0); \ 169 sum_w = (v4u32) __msa_pckev_w((v4i32) sum_d, (v4i32) sum_d); in ff_dc_8x8_msa() 171 sum_w = (v4u32) __msa_srari_w((v4i32) sum_d, 4); in ff_dc_8x8_msa() 197 sum_w = (v4u32) __msa_srari_w((v4i32) sum_d, 3); \ 223 sum_w = (v4u32) __msa_pckev_w((v4i32) sum_ in ff_dc_16x16_msa() [all...] |
H A D | h264_deblock_msa.c | 52 ref_2 = (v16i8)__msa_ilvr_w((v4i32)ref_3, (v4i32)ref_2); \ 53 ref_0 = (v16i8)__msa_ilvr_w((v4i32)ref_0, (v4i32)ref_0); \ 54 ref_1 = (v16i8)__msa_ilvr_w((v4i32)ref_1, (v4i32)ref_1); \
|
H A D | vp8_idct_msa.c | 31 v4i32 a1_m, b1_m, c1_m, d1_m; \ 32 v4i32 c_tmp1_m, c_tmp2_m, d_tmp1_m, d_tmp2_m; \ 33 v4i32 const_cospi8sqrt2minus1_m, sinpi8_sqrt2_m; \ 51 v4i32 in0, in1, in2, in3, hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3; in ff_vp8_idct_add_msa() 52 v4i32 res0, res1, res2, res3; in ff_vp8_idct_add_msa() 108 v4i32 in0, in1, in2, in3, a1, b1, c1, d1; in ff_vp8_luma_dc_wht_msa() 109 v4i32 hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3; in ff_vp8_luma_dc_wht_msa()
|
H A D | h264pred_msa.c | 144 v4i32 int_multiplier = { 0, 1, 2, 3 }; in intra_predict_plane_8x8_msa() 147 v4i32 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7, vec8; in intra_predict_plane_8x8_msa() 158 res0 = __msa_copy_s_w((v4i32) sum, 0); in intra_predict_plane_8x8_msa() 212 v4i32 int_multiplier = { 0, 1, 2, 3 }; in intra_predict_plane_16x16_msa() 216 v4i32 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7, vec8, res_add; in intra_predict_plane_16x16_msa() 217 v4i32 reg0, reg1, reg2, reg3; in intra_predict_plane_16x16_msa() 229 res_add = (v4i32) __msa_hadd_s_d(vec8, vec8); in intra_predict_plane_16x16_msa() 296 src0 = __msa_copy_u_w((v4i32) sum, 0); in intra_predict_dc_4blk_8x8_msa() 297 src1 = __msa_copy_u_w((v4i32) sum, 1); in intra_predict_dc_4blk_8x8_msa() 356 sum = (v4u32) __msa_srari_w((v4i32) su in intra_predict_vert_dc_8x8_msa() [all...] |
H A D | h264idct_msa.c | 49 v4i32 vres0_r, vres1_r, vres2_r, vres3_r; in avc_deq_idct_luma_dc_msa() 50 const v4i32 de_q_vec = __msa_fill_w(de_q_val); in avc_deq_idct_luma_dc_msa() 117 v4i32 tmp0_r, tmp1_r, tmp2_r, tmp3_r, tmp4_r, tmp5_r, tmp6_r, tmp7_r; in avc_idct8_addblk_msa() 118 v4i32 tmp0_l, tmp1_l, tmp2_l, tmp3_l, tmp4_l, tmp5_l, tmp6_l, tmp7_l; in avc_idct8_addblk_msa() 119 v4i32 vec0_r, vec1_r, vec2_r, vec3_r, vec0_l, vec1_l, vec2_l, vec3_l; in avc_idct8_addblk_msa() 120 v4i32 res0_r, res1_r, res2_r, res3_r, res4_r, res5_r, res6_r, res7_r; in avc_idct8_addblk_msa() 121 v4i32 res0_l, res1_l, res2_l, res3_l, res4_l, res5_l, res6_l, res7_l; in avc_idct8_addblk_msa() 303 out0_m = __msa_copy_u_w((v4i32) dst0_m, 0); in ff_h264_idct_add_msa() 304 out1_m = __msa_copy_u_w((v4i32) dst0_m, 1); in ff_h264_idct_add_msa() 305 out2_m = __msa_copy_u_w((v4i32) dst1_ in ff_h264_idct_add_msa() [all...] |
H A D | hevc_idct_msa.c | 68 v4i32 vec0, vec1, vec2, vec3, vec4, vec5; \ 69 v4i32 cnst64 = __msa_ldi_w(64); \ 70 v4i32 cnst83 = __msa_ldi_w(83); \ 71 v4i32 cnst36 = __msa_ldi_w(36); \ 99 v4i32 temp0_r, temp1_r, temp2_r, temp3_r, temp4_r, temp5_r; \ 100 v4i32 temp0_l, temp1_l, temp2_l, temp3_l, temp4_l, temp5_l; \ 101 v4i32 sum0_r, sum1_r, sum2_r, sum3_r; \ 102 v4i32 sum0_l, sum1_l, sum2_l, sum3_l; \ 196 v4i32 temp0_r, temp1_r, temp0_l, temp1_l; \ 197 v4i32 sum0_ [all...] |
H A D | hevc_macros_msa.h | 38 v4i32 out_m; \ 57 v4i32 out_m; \
|
H A D | hevc_lpf_sao_msa.c | 81 cmp3 = (v2i64) __msa_ilvev_w((v4i32) cmp1, (v4i32) cmp0); in hevc_loopfilter_luma_hor_msa() 82 cmp3 = (v2i64) __msa_ceqi_w((v4i32) cmp3, 0); in hevc_loopfilter_luma_hor_msa() 130 cmp2 = (v2i64) __msa_ilvev_w((v4i32) cmp1, (v4i32) cmp0); in hevc_loopfilter_luma_hor_msa() 131 cmp2 = (v2i64) __msa_ceqi_w((v4i32) cmp2, 0); in hevc_loopfilter_luma_hor_msa() 691 tmp2 = __msa_copy_u_w((v4i32) dst0, 0); in hevc_loopfilter_luma_ver_msa() 692 tmp3 = __msa_copy_u_w((v4i32) dst0, 1); in hevc_loopfilter_luma_ver_msa() 698 tmp2 = __msa_copy_u_w((v4i32) dst0, 2); in hevc_loopfilter_luma_ver_msa() 699 tmp3 = __msa_copy_u_w((v4i32) dst in hevc_loopfilter_luma_ver_msa() [all...] |
H A D | h264dsp_msa.c | 577 src0 = (v16u8) __msa_insert_w((v4i32) src0, 0, load0); \ 578 src1 = (v16u8) __msa_insert_w((v4i32) src1, 0, load1); \ 579 src2 = (v16u8) __msa_insert_w((v4i32) src2, 0, load2); \ 580 src3 = (v16u8) __msa_insert_w((v4i32) src3, 0, load3); \ 644 src0 = (v16u8) __msa_insert_w((v4i32) src0, 0, load0); \ 645 src1 = (v16u8) __msa_insert_w((v4i32) src1, 0, load1); \ 1124 dst2_x = (v16u8) __msa_ilvl_w((v4i32) tmp2, (v4i32) tmp0); in avc_h_loop_filter_luma_mbaff_intra_msa() 1125 dst2_y = (v16u8) __msa_ilvl_w((v4i32) tmp3, (v4i32) tmp in avc_h_loop_filter_luma_mbaff_intra_msa() [all...] |
H A D | hpeldsp_msa.c | 71 out0 = __msa_copy_u_w((v4i32) res0, 0); in common_hz_bil_4w_msa() 72 out1 = __msa_copy_u_w((v4i32) res1, 0); in common_hz_bil_4w_msa() 233 tmp0 = (v16u8) __msa_insert_w((v4i32) tmp0, 0, dst0); in common_hz_bil_and_aver_dst_4w_msa() 234 tmp1 = (v16u8) __msa_insert_w((v4i32) tmp1, 0, dst1); in common_hz_bil_and_aver_dst_4w_msa() 239 out0 = __msa_copy_u_w((v4i32) res0, 0); in common_hz_bil_and_aver_dst_4w_msa() 240 out1 = __msa_copy_u_w((v4i32) res1, 0); in common_hz_bil_and_aver_dst_4w_msa() 311 out0 = __msa_copy_u_w((v4i32) res0, 0); in common_vt_bil_4w_msa() 312 out1 = __msa_copy_u_w((v4i32) res1, 0); in common_vt_bil_4w_msa() 460 tmp0 = (v16u8) __msa_insert_w((v4i32) tmp0, 0, dst0); in common_vt_bil_and_aver_dst_4w_msa() 461 tmp1 = (v16u8) __msa_insert_w((v4i32) tmp in common_vt_bil_and_aver_dst_4w_msa() [all...] |
/third_party/skia/third_party/externals/libwebp/src/dsp/ |
H A D | msa_macro.h | 27 #define ADDVI_W(a, b) __msa_addvi_w((v4i32)a, b) 30 #define SRAI_W(a, b) __msa_srai_w((v4i32)a, b) 32 #define SLLI_B(a, b) __msa_slli_b((v4i32)a, b) 58 #define LD_SW(...) LD_W(v4i32, __VA_ARGS__) 70 #define ST_SW(...) ST_W(v4i32, __VA_ARGS__) 275 #define LD_SW2(...) LD_W2(v4i32, __VA_ARGS__) 282 #define LD_SW3(...) LD_W3(v4i32, __VA_ARGS__) 289 #define LD_SW4(...) LD_W4(v4i32, __VA_ARGS__) 329 #define ST_SW2(...) ST_W2(v4i32, __VA_ARGS__) 336 #define ST_SW3(...) ST_W3(v4i32, __VA_ARGS_ [all...] |
H A D | lossless_enc_msa.c | 23 v4i32 t4, t5; \ 29 t4 = __msa_srli_w((v4i32)src0, 16); \ 30 t5 = __msa_srli_w((v4i32)src1, 16); \ 41 v4i32 t2; \ 44 t2 = __msa_srli_w((v4i32)src, 16); \ 83 const uint32_t pix_w = __msa_copy_s_w((v4i32)dst0, 2); in TransformColor_MSA() 90 const uint32_t pix_w = __msa_copy_s_w((v4i32)dst0, 0); in TransformColor_MSA()
|
H A D | lossless_msa.c | 40 pix_w = __msa_copy_s_w((v4i32)dst2, 0); \ 58 uint32_t pix_w = __msa_copy_s_w((v4i32)dst0, 2); \ 84 v4i32 t4, t5; \ 90 t4 = __msa_srli_w((v4i32)t0, 16); \ 91 t5 = __msa_srli_w((v4i32)t1, 16); \ 102 v4i32 t2; \ 105 t2 = __msa_srli_w((v4i32)t0, 16); \ 324 const uint32_t pix_w = __msa_copy_s_w((v4i32)dst0, 2); in TransformColorInverse_MSA() 331 const uint32_t pix_w = __msa_copy_s_w((v4i32)dst0, 0); in TransformColorInverse_MSA()
|
H A D | dec_msa.c | 25 v4i32 a1_m, b1_m, c1_m, d1_m; \ 26 v4i32 c_tmp1_m, c_tmp2_m, d_tmp1_m, d_tmp2_m; \ 27 const v4i32 cospi8sqrt2minus1 = __msa_fill_w(20091); \ 28 const v4i32 sinpi8sqrt2 = __msa_fill_w(35468); \ 45 v4i32 in0, in1, in2, in3, hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3; in TransformOne() 46 v4i32 res0, res1, res2, res3; in TransformOne() 66 res0 = (v4i32)__msa_pckev_b((v16i8)vt0, (v16i8)vt1); in TransformOne() 131 v4i32 tmp0 = { 0 }; in TransformAC3() 132 v4i32 out0 = __msa_fill_w(a + d4); in TransformAC3() 133 v4i32 out in TransformAC3() [all...] |
H A D | enc_msa.c | 26 v4i32 a1_m, b1_m, c1_m, d1_m; \ 27 const v4i32 cospi8sqrt2minus1 = __msa_fill_w(20091); \ 28 const v4i32 sinpi8sqrt2 = __msa_fill_w(35468); \ 29 v4i32 c_tmp1_m = in1 * sinpi8sqrt2; \ 30 v4i32 c_tmp2_m = in3 * cospi8sqrt2minus1; \ 31 v4i32 d_tmp1_m = in1 * cospi8sqrt2minus1; \ 32 v4i32 d_tmp2_m = in3 * sinpi8sqrt2; \ 47 v4i32 in0, in1, in2, in3, hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3; in ITransformOne() 48 v4i32 res0, res1, res2, res3; in ITransformOne() 68 res0 = (v4i32)__msa_pckev_ in ITransformOne() [all...] |
H A D | rescaler_msa.c | 57 dst = __msa_copy_s_w((v4i32)t0, 0); \ 84 dst = (v4u32)__msa_pckev_w((v4i32)out1, (v4i32)out0); \ 114 dst = __msa_copy_s_w((v4i32)t0, 0); \ 122 const v4i32 zero = { 0 }; in ExportRowExpand_0() 179 const v4i32 B1 = __msa_fill_w(B); in ExportRowExpand_1() 180 const v4i32 A1 = __msa_fill_w(A); in ExportRowExpand_1() 181 const v4i32 AB = __msa_ilvr_w(A1, B1); in ExportRowExpand_1() 272 const v4i32 zero = { 0 }; 356 const v4i32 zer [all...] |
/third_party/ffmpeg/libavutil/mips/ |
H A D | generic_macros_msa.h | 37 #define LD_SW(...) LD_V(v4i32, __VA_ARGS__) 45 #define ST_SW(...) ST_V(v4i32, __VA_ARGS__) 281 #define LD_SW2(...) LD_V2(v4i32, __VA_ARGS__) 300 #define LD_SW4(...) LD_V4(v4i32, __VA_ARGS__) 339 #define LD_SW8(...) LD_V8(v4i32, __VA_ARGS__) 367 #define ST_SW2(...) ST_V2(v4i32, __VA_ARGS__) 377 #define ST_SW4(...) ST_V4(v4i32, __VA_ARGS__) 393 #define ST_SW8(...) ST_V8(v4i32, __VA_ARGS__) 447 out0_m = __msa_copy_u_w((v4i32) in, idx); \ 453 out0_m = __msa_copy_u_w((v4i32) i [all...] |
/third_party/skia/third_party/externals/swiftshader/third_party/llvm-10.0/llvm/lib/Target/X86/ |
H A D | X86TargetTransformInfo.cpp | 198 { ISD::MUL, MVT::v4i32, 11 }, // pmulld in getArithmeticInstrCost() 222 if (Args.size() == 2 && ISD == ISD::MUL && LT.second == MVT::v4i32) { in getArithmeticInstrCost() 424 { ISD::SDIV, MVT::v4i32, 19 }, // pmuludq sequence in getArithmeticInstrCost() 425 { ISD::SREM, MVT::v4i32, 24 }, // pmuludq+mul+sub sequence in getArithmeticInstrCost() 428 { ISD::UDIV, MVT::v4i32, 15 }, // pmuludq sequence in getArithmeticInstrCost() 429 { ISD::UREM, MVT::v4i32, 20 }, // pmuludq+mul+sub sequence in getArithmeticInstrCost() 440 if (ISD == ISD::SDIV && LT.second == MVT::v4i32 && ST->hasSSE41()) in getArithmeticInstrCost() 442 if (ISD == ISD::SREM && LT.second == MVT::v4i32 && ST->hasSSE41()) in getArithmeticInstrCost() 467 { ISD::SHL, MVT::v4i32, 1 }, // pslld in getArithmeticInstrCost() 471 { ISD::SRL, MVT::v4i32, in getArithmeticInstrCost() [all...] |
/third_party/skia/third_party/externals/libpng/mips/ |
H A D | filter_msa_intrinsics.c | 331 out0 = (RTYPE) __msa_ilvev_w((v4i32) in1, (v4i32) in0); \ 332 out1 = (RTYPE) __msa_ilvev_w((v4i32) in3, (v4i32) in2); \ 474 src0 = (v16u8) __msa_insert_w((v4i32) zero, 0, inp0); in png_read_filter_row_sub4_msa() 515 src0 = (v16u8) __msa_insert_w((v4i32) zero, 0, inp0); in png_read_filter_row_sub3_msa() 533 out1 = __msa_copy_s_w((v4i32) dst0, 2); in png_read_filter_row_sub3_msa() 558 src0 = (v16u8) __msa_insert_w((v4i32) zero, 0, inp0); in png_read_filter_row_avg4_msa() 559 src1 = (v16u8) __msa_insert_w((v4i32) zero, 0, inp1); in png_read_filter_row_avg4_msa() 562 out0 = __msa_copy_s_w((v4i32) src in png_read_filter_row_avg4_msa() [all...] |
/third_party/skia/third_party/externals/swiftshader/third_party/llvm-10.0/llvm/lib/Target/ARM/ |
H A D | ARMTargetTransformInfo.cpp | 200 {ISD::SIGN_EXTEND, MVT::v4i32, MVT::v4i16, 0}, in getCastInstrCost() 201 {ISD::ZERO_EXTEND, MVT::v4i32, MVT::v4i16, 0}, in getCastInstrCost() 202 {ISD::SIGN_EXTEND, MVT::v4i32, MVT::v4i8, 0}, in getCastInstrCost() 203 {ISD::ZERO_EXTEND, MVT::v4i32, MVT::v4i8, 0}, in getCastInstrCost() 219 { ISD::SIGN_EXTEND, MVT::v4i32, MVT::v4i16, 0 }, in getCastInstrCost() 220 { ISD::ZERO_EXTEND, MVT::v4i32, MVT::v4i16, 0 }, in getCastInstrCost() 223 { ISD::TRUNCATE, MVT::v4i32, MVT::v4i64, 0 }, in getCastInstrCost() 224 { ISD::TRUNCATE, MVT::v4i16, MVT::v4i32, 1 }, in getCastInstrCost() 243 { ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i32, 1 }, in getCastInstrCost() 244 { ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i32, in getCastInstrCost() [all...] |
/third_party/skia/third_party/externals/swiftshader/third_party/llvm-10.0/llvm/lib/Target/AArch64/ |
H A D | AArch64TargetTransformInfo.cpp | 302 { ISD::TRUNCATE, MVT::v4i16, MVT::v4i32, 1 }, in getCastInstrCost() 303 { ISD::TRUNCATE, MVT::v4i32, MVT::v4i64, 0 }, in getCastInstrCost() 310 { ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i32, 2 }, in getCastInstrCost() 311 { ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i32, 2 }, in getCastInstrCost() 327 { ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i32, 1 }, in getCastInstrCost() 330 { ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i32, 1 }, in getCastInstrCost() 368 { ISD::FP_TO_SINT, MVT::v4i32, MVT::v4f32, 1 }, in getCastInstrCost() 371 { ISD::FP_TO_UINT, MVT::v4i32, MVT::v4f32, 1 }, in getCastInstrCost() 947 {ISD::ADD, MVT::v4i32, 1}, in getArithmeticReductionCost() member in MVT 967 { TTI::SK_Broadcast, MVT::v4i32, in getShuffleCost() [all...] |
H A D | AArch64ISelDAGToDAG.cpp | 613 case MVT::v4i32: in tryMLAV64LaneV128() 642 case MVT::v4i32: in tryMULLV64LaneV128() 653 case MVT::v4i32: in tryMULLV64LaneV128() 3197 } else if (VT == MVT::v4i32 || VT == MVT::v4f32) { in Select() 3224 } else if (VT == MVT::v4i32 || VT == MVT::v4f32) { in Select() 3251 } else if (VT == MVT::v4i32 || VT == MVT::v4f32) { in Select() 3278 } else if (VT == MVT::v4i32 || VT == MVT::v4f32) { in Select() 3305 } else if (VT == MVT::v4i32 || VT == MVT::v4f32) { in Select() 3332 } else if (VT == MVT::v4i32 || VT == MVT::v4f32) { in Select() 3359 } else if (VT == MVT::v4i32 || V in Select() [all...] |