diff options
Diffstat (limited to 'third_party/aom/av1/encoder/mips/msa/fdct_msa.h')
-rw-r--r-- | third_party/aom/av1/encoder/mips/msa/fdct_msa.h | 117 |
1 files changed, 117 insertions, 0 deletions
diff --git a/third_party/aom/av1/encoder/mips/msa/fdct_msa.h b/third_party/aom/av1/encoder/mips/msa/fdct_msa.h new file mode 100644 index 000000000..52bcf790c --- /dev/null +++ b/third_party/aom/av1/encoder/mips/msa/fdct_msa.h @@ -0,0 +1,117 @@ +/* + * Copyright (c) 2016, Alliance for Open Media. All rights reserved + * + * This source code is subject to the terms of the BSD 2 Clause License and + * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License + * was not distributed with this source code in the LICENSE file, you can + * obtain it at www.aomedia.org/license/software. If the Alliance for Open + * Media Patent License 1.0 was not distributed with this source code in the + * PATENTS file, you can obtain it at www.aomedia.org/license/patent. + */ + +#ifndef AV1_ENCODER_MIPS_MSA_AV1_FDCT_MSA_H_ +#define AV1_ENCODER_MIPS_MSA_AV1_FDCT_MSA_H_ + +#include "aom_dsp/mips/fwd_txfm_msa.h" +#include "aom_dsp/mips/txfm_macros_msa.h" +#include "aom_ports/mem.h" + +#define AOM_ADST8(in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, out2, \ + out3, out4, out5, out6, out7) \ + { \ + v8i16 cnst0_m, cnst1_m, cnst2_m, cnst3_m, cnst4_m; \ + v8i16 vec0_m, vec1_m, vec2_m, vec3_m, s0_m, s1_m; \ + v8i16 coeff0_m = { cospi_2_64, cospi_6_64, cospi_10_64, cospi_14_64, \ + cospi_18_64, cospi_22_64, cospi_26_64, cospi_30_64 }; \ + v8i16 coeff1_m = { cospi_8_64, -cospi_8_64, cospi_16_64, -cospi_16_64, \ + cospi_24_64, -cospi_24_64, 0, 0 }; \ + \ + SPLATI_H2_SH(coeff0_m, 0, 7, cnst0_m, cnst1_m); \ + cnst2_m = -cnst0_m; \ + ILVEV_H2_SH(cnst0_m, cnst1_m, cnst1_m, cnst2_m, cnst0_m, cnst1_m); \ + SPLATI_H2_SH(coeff0_m, 4, 3, cnst2_m, cnst3_m); \ + cnst4_m = -cnst2_m; \ + ILVEV_H2_SH(cnst2_m, cnst3_m, cnst3_m, cnst4_m, cnst2_m, cnst3_m); \ + \ + ILVRL_H2_SH(in0, in7, vec1_m, vec0_m); \ + ILVRL_H2_SH(in4, in3, vec3_m, vec2_m); \ + DOT_ADD_SUB_SRARI_PCK(vec0_m, vec1_m, vec2_m, vec3_m, cnst0_m, cnst1_m, \ + cnst2_m, cnst3_m, in7, in0, in4, in3); \ + \ + SPLATI_H2_SH(coeff0_m, 2, 5, cnst0_m, cnst1_m); \ + cnst2_m = -cnst0_m; \ + ILVEV_H2_SH(cnst0_m, cnst1_m, cnst1_m, cnst2_m, cnst0_m, cnst1_m); \ + SPLATI_H2_SH(coeff0_m, 6, 1, cnst2_m, cnst3_m); \ + cnst4_m = -cnst2_m; \ + ILVEV_H2_SH(cnst2_m, cnst3_m, cnst3_m, cnst4_m, cnst2_m, cnst3_m); \ + \ + ILVRL_H2_SH(in2, in5, vec1_m, vec0_m); \ + ILVRL_H2_SH(in6, in1, vec3_m, vec2_m); \ + \ + DOT_ADD_SUB_SRARI_PCK(vec0_m, vec1_m, vec2_m, vec3_m, cnst0_m, cnst1_m, \ + cnst2_m, cnst3_m, in5, in2, in6, in1); \ + BUTTERFLY_4(in7, in0, in2, in5, s1_m, s0_m, in2, in5); \ + out7 = -s0_m; \ + out0 = s1_m; \ + \ + SPLATI_H4_SH(coeff1_m, 0, 4, 1, 5, cnst0_m, cnst1_m, cnst2_m, cnst3_m); \ + \ + ILVEV_H2_SH(cnst3_m, cnst0_m, cnst1_m, cnst2_m, cnst3_m, cnst2_m); \ + cnst0_m = __msa_ilvev_h(cnst1_m, cnst0_m); \ + cnst1_m = cnst0_m; \ + \ + ILVRL_H2_SH(in4, in3, vec1_m, vec0_m); \ + ILVRL_H2_SH(in6, in1, vec3_m, vec2_m); \ + DOT_ADD_SUB_SRARI_PCK(vec0_m, vec1_m, vec2_m, vec3_m, cnst0_m, cnst2_m, \ + cnst3_m, cnst1_m, out1, out6, s0_m, s1_m); \ + \ + SPLATI_H2_SH(coeff1_m, 2, 3, cnst0_m, cnst1_m); \ + cnst1_m = __msa_ilvev_h(cnst1_m, cnst0_m); \ + \ + ILVRL_H2_SH(in2, in5, vec1_m, vec0_m); \ + ILVRL_H2_SH(s0_m, s1_m, vec3_m, vec2_m); \ + out3 = DOT_SHIFT_RIGHT_PCK_H(vec0_m, vec1_m, cnst0_m); \ + out4 = DOT_SHIFT_RIGHT_PCK_H(vec0_m, vec1_m, cnst1_m); \ + out2 = DOT_SHIFT_RIGHT_PCK_H(vec2_m, vec3_m, cnst0_m); \ + out5 = DOT_SHIFT_RIGHT_PCK_H(vec2_m, vec3_m, cnst1_m); \ + \ + out1 = -out1; \ + out3 = -out3; \ + out5 = -out5; \ + } + +#define AOM_FADST4(in0, in1, in2, in3, out0, out1, out2, out3) \ + { \ + v4i32 s0_m, s1_m, s2_m, s3_m, constant_m; \ + v4i32 in0_r_m, in1_r_m, in2_r_m, in3_r_m; \ + \ + UNPCK_R_SH_SW(in0, in0_r_m); \ + UNPCK_R_SH_SW(in1, in1_r_m); \ + UNPCK_R_SH_SW(in2, in2_r_m); \ + UNPCK_R_SH_SW(in3, in3_r_m); \ + \ + constant_m = __msa_fill_w(sinpi_4_9); \ + MUL2(in0_r_m, constant_m, in3_r_m, constant_m, s1_m, s0_m); \ + \ + constant_m = __msa_fill_w(sinpi_1_9); \ + s0_m += in0_r_m * constant_m; \ + s1_m -= in1_r_m * constant_m; \ + \ + constant_m = __msa_fill_w(sinpi_2_9); \ + s0_m += in1_r_m * constant_m; \ + s1_m += in3_r_m * constant_m; \ + \ + s2_m = in0_r_m + in1_r_m - in3_r_m; \ + \ + constant_m = __msa_fill_w(sinpi_3_9); \ + MUL2(in2_r_m, constant_m, s2_m, constant_m, s3_m, in1_r_m); \ + \ + in0_r_m = s0_m + s3_m; \ + s2_m = s1_m - s3_m; \ + s3_m = s1_m - s0_m + s3_m; \ + \ + SRARI_W4_SW(in0_r_m, in1_r_m, s2_m, s3_m, DCT_CONST_BITS); \ + PCKEV_H4_SH(in0_r_m, in0_r_m, in1_r_m, in1_r_m, s2_m, s2_m, s3_m, s3_m, \ + out0, out1, out2, out3); \ + } +#endif // AV1_ENCODER_MIPS_MSA_AV1_FDCT_MSA_H_ |