Lines Matching refs:coeffs
330 void ff_hevc_idct_4x4_lsx(int16_t *coeffs, int col_limit)
337 in0 = __lsx_vld(coeffs, 0);
338 in1 = __lsx_vld(coeffs, 16);
356 __lsx_vst(in0, coeffs, 0);
357 __lsx_vst(in1, coeffs, 16);
360 void ff_hevc_idct_8x8_lsx(int16_t *coeffs, int col_limit)
365 DUP4_ARG2(__lsx_vld, coeffs, 0, coeffs, 16, coeffs, 32,
366 coeffs, 48, in0, in1, in2, in3);
367 DUP4_ARG2(__lsx_vld, coeffs, 64, coeffs, 80, coeffs, 96,
368 coeffs, 112, in4, in5, in6, in7);
376 __lsx_vst(in0, coeffs, 0);
377 __lsx_vst(in1, coeffs, 16);
378 __lsx_vst(in2, coeffs, 32);
379 __lsx_vst(in3, coeffs, 48);
380 __lsx_vst(in4, coeffs, 64);
381 __lsx_vst(in5, coeffs, 80);
382 __lsx_vst(in6, coeffs, 96);
383 __lsx_vst(in7, coeffs, 112);
386 void ff_hevc_idct_16x16_lsx(int16_t *coeffs, int col_limit)
391 int16_t *src = coeffs;
428 buf_ptr = coeffs;
457 buf_ptr = coeffs + 8;
461 DUP4_ARG2(__lsx_vld, coeffs, 0, coeffs, 32, coeffs, 64, coeffs, 96,
463 DUP4_ARG2(__lsx_vld, coeffs, 128, coeffs, 160, coeffs, 192, coeffs, 224,
467 __lsx_vst(vec0, coeffs, 0);
468 __lsx_vst(vec1, coeffs, 32);
469 __lsx_vst(vec2, coeffs, 64);
470 __lsx_vst(vec3, coeffs, 96);
471 __lsx_vst(vec4, coeffs, 128);
472 __lsx_vst(vec5, coeffs, 160);
473 __lsx_vst(vec6, coeffs, 192);
474 __lsx_vst(vec7, coeffs, 224);
476 src = coeffs + 8;
482 src = coeffs + 128;
498 src = coeffs + 8;
508 src = coeffs + 136;
525 static void hevc_idct_8x32_column_lsx(int16_t *coeffs, int32_t buf_pitch,
538 int16_t *src0 = (coeffs + buf_pitch);
539 int16_t *src1 = (coeffs + buf_pitch_2);
540 int16_t *src2 = (coeffs + buf_pitch_4);
541 int16_t *src3 = (coeffs);
749 __lsx_vst(in0, (coeffs + i * buf_pitch), 0);
755 __lsx_vst(in0, (coeffs + (31 - i) * buf_pitch), 0);
761 static void hevc_idct_transpose_32x8_to_8x32(int16_t *coeffs, int16_t *tmp_buf)
767 DUP4_ARG2(__lsx_vld, coeffs, 0, coeffs, 64, coeffs, 128,
768 coeffs, 192, in0, in1, in2, in3);
769 DUP4_ARG2(__lsx_vld, coeffs, 256, coeffs, 320, coeffs, 384,
770 coeffs, 448, in4, in5, in6, in7);
771 coeffs += 8;
786 static void hevc_idct_transpose_8x32_to_32x8(int16_t *tmp_buf, int16_t *coeffs)
799 __lsx_vst(in0, coeffs, 0);
800 __lsx_vst(in1, coeffs, 64);
801 __lsx_vst(in2, coeffs, 128);
802 __lsx_vst(in3, coeffs, 192);
803 __lsx_vst(in4, coeffs, 256);
804 __lsx_vst(in5, coeffs, 320);
805 __lsx_vst(in6, coeffs, 384);
806 __lsx_vst(in7, coeffs, 448);
807 coeffs += 8;
811 void ff_hevc_idct_32x32_lsx(int16_t *coeffs, int col_limit)
814 int16_t *src = coeffs;
828 hevc_idct_8x32_column_lsx((coeffs + col_cnt * 8), buf_pitch, round);
836 src = (coeffs + 32 * 8 * row_cnt);