virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-12 19:18:44 +02:00
Code Issues Releases Activity
00ed04d614
FFmpeg/libavcodec/mips/simple_idct_mmi.c
Shiyou Yin df13b75aa1 avcodec/mips: [loongson] reoptimize simple idct with mmi. 
Performance of mpeg4 decoding improved about 23%(from 128fps to 158fps, tested on loongson 3A3000).
Reoptimized following functions with mmi.
1. ff_simple_idct_put_8_mmi
2. ff_simple_idct_add_8_mmi
3. ff_simple_idct_8_mmi

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2018-09-02 03:37:32 +02:00

424 lines
24 KiB
C
Raw Blame History

/*
* Loongson SIMD optimized simple idct
*
* Copyright (c) 2015 Loongson Technology Corporation Limited
* Copyright (c) 2015 Zhou Xiaoyong <zhouxiaoyong@loongson.cn>
* Zhang Shuangshuang <zhangshuangshuang@ict.ac.cn>
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "idctdsp_mips.h"
#include "constants.h"
#include "libavutil/mips/asmdefs.h"
#include "libavutil/mips/mmiutils.h"
#define W1 22725 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W2 21407 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W3 19266 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W4 16383 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W5 12873 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W6 8867 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W7 4520 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define ROW_SHIFT 11
#define COL_SHIFT 20
#define DC_SHIFT 3
DECLARE_ALIGNED(8, const int16_t, W_arr)[46] = {
W4, W2, W4, W6,
W1, W3, W5, W7,
W4, W6, -W4, -W2,
W3, -W7, -W1, -W5,
W4, -W6, -W4, W2,
W5, -W1, W7, W3,
W4, -W2, W4, -W6,
W7, -W5, W3, -W1,
1024, 0, 1024, 0, //ff_p32_1024 = 0x0000040000000400ULL
0, -1, -1, -1, //mask = 0xffffffffffff0000ULL
32, 32, 32, 32 //ff_p16_32 = 0x0020002000200020ULL
};
void ff_simple_idct_8_mmi(int16_t *block)
{
BACKUP_REG
__asm__ volatile (
#define IDCT_ROW_COND_DC(src1, src2) \
"dmfc1 $11, "#src1" \n\t" \
"dmfc1 $12, "#src2" \n\t" \
"and $11, $11, $9 \n\t" \
"or $10, $11, $12 \n\t" \
"beqz $10, 1f \n\t" \
\
"punpcklhw $f30, "#src1", "#src2" \n\t" \
"punpckhhw $f31, "#src1", "#src2" \n\t" \
/* s6, s4, s2, s0 */ \
"punpcklhw "#src1", $f30, $f31 \n\t" \
/* s7, s5, s3, s1 */ \
"punpckhhw "#src2", $f30, $f31 \n\t" \
\
"pmaddhw $f30, "#src1", $f18 \n\t" \
"pmaddhw $f31, "#src2", $f19 \n\t" \
"paddw $f28, $f30, $f31 \n\t" \
"psubw $f29, $f30, $f31 \n\t" \
"punpcklwd $f30, $f28, $f29 \n\t" \
"punpckhwd $f31, $f28, $f29 \n\t" \
"paddw $f26, $f30, $f31 \n\t" \
"paddw $f26, $f26, $f16 \n\t" \
/* $f26: src[7], src[0] */ \
"psraw $f26, $f26, $f17 \n\t" \
\
"pmaddhw $f30, "#src1", $f20 \n\t" \
"pmaddhw $f31, "#src2", $f21 \n\t" \
"paddw $f28, $f30, $f31 \n\t" \
"psubw $f29, $f30, $f31 \n\t" \
"punpcklwd $f30, $f28, $f29 \n\t" \
"punpckhwd $f31, $f28, $f29 \n\t" \
"paddw $f27, $f30, $f31 \n\t" \
"paddw $f27, $f27, $f16 \n\t" \
/* $f27: src[6], src[1] */ \
"psraw $f27, $f27, $f17 \n\t" \
\
"pmaddhw $f30, "#src1", $f22 \n\t" \
"pmaddhw $f31, "#src2", $f23 \n\t" \
"paddw $f28, $f30, $f31 \n\t" \
"psubw $f29, $f30, $f31 \n\t" \
"punpcklwd $f30, $f28, $f29 \n\t" \
"punpckhwd $f31, $f28, $f29 \n\t" \
"paddw $f28, $f30, $f31 \n\t" \
"paddw $f28, $f28, $f16 \n\t" \
/* $f28: src[5], src[2] */ \
"psraw $f28, $f28, $f17 \n\t" \
\
"pmaddhw $f30, "#src1", $f24 \n\t" \
"pmaddhw $f31, "#src2", $f25 \n\t" \
"paddw "#src1", $f30, $f31 \n\t" \
"psubw "#src2", $f30, $f31 \n\t" \
"punpcklwd $f30, "#src1", "#src2" \n\t" \
"punpckhwd $f31, "#src1", "#src2" \n\t" \
"paddw $f29, $f30, $f31 \n\t" \
"paddw $f29, $f29, $f16 \n\t" \
/* $f29: src[4], src[3] */ \
"psraw $f29, $f29, $f17 \n\t" \
\
"punpcklhw "#src1", $f26, $f27 \n\t" \
"punpckhhw $f30, $f27, $f26 \n\t" \
"punpcklhw $f31, $f28, $f29 \n\t" \
"punpckhhw "#src2", $f29, $f28 \n\t" \
/* src[3], src[2], src[1], src[0] */ \
"punpcklwd "#src1", "#src1", $f31 \n\t" \
/* src[7], src[6], src[5], src[4] */ \
"punpcklwd "#src2", "#src2", $f30 \n\t" \
"j 2f \n\t" \
\
"1: \n\t" \
"li $10, 3 \n\t" \
"dmtc1 $10, $f30 \n\t" \
"psllh $f28, "#src1", $f30 \n\t" \
"dmtc1 $9, $f31 \n\t" \
"punpcklhw $f29, $f28, $f28 \n\t" \
"and $f29, $f29, $f31 \n\t" \
"paddw $f28, $f28, $f29 \n\t" \
"punpcklwd "#src1", $f28, $f28 \n\t" \
"punpcklwd "#src2", $f28, $f28 \n\t" \
"2: \n\t" \
/* idctRowCondDC row0~8 */
/* load W */
"gslqc1 $f19, $f18, 0x00(%[w_arr]) \n\t"
"gslqc1 $f21, $f20, 0x10(%[w_arr]) \n\t"
"gslqc1 $f23, $f22, 0x20(%[w_arr]) \n\t"
"gslqc1 $f25, $f24, 0x30(%[w_arr]) \n\t"
"gslqc1 $f17, $f16, 0x40(%[w_arr]) \n\t"
/* load source in block */
"gslqc1 $f1, $f0, 0x00(%[block]) \n\t"
"gslqc1 $f3, $f2, 0x10(%[block]) \n\t"
"gslqc1 $f5, $f4, 0x20(%[block]) \n\t"
"gslqc1 $f7, $f6, 0x30(%[block]) \n\t"
"gslqc1 $f9, $f8, 0x40(%[block]) \n\t"
"gslqc1 $f11, $f10, 0x50(%[block]) \n\t"
"gslqc1 $f13, $f12, 0x60(%[block]) \n\t"
"gslqc1 $f15, $f14, 0x70(%[block]) \n\t"
/* $9: mask ; $f17: ROW_SHIFT */
"dmfc1 $9, $f17 \n\t"
"li $10, 11 \n\t"
"mtc1 $10, $f17 \n\t"
IDCT_ROW_COND_DC($f0,$f1)
IDCT_ROW_COND_DC($f2,$f3)
IDCT_ROW_COND_DC($f4,$f5)
IDCT_ROW_COND_DC($f6,$f7)
IDCT_ROW_COND_DC($f8,$f9)
IDCT_ROW_COND_DC($f10,$f11)
IDCT_ROW_COND_DC($f12,$f13)
IDCT_ROW_COND_DC($f14,$f15)
#define IDCT_COL_CASE1(src, out1, out2) \
"pmaddhw $f26, "#src", $f18 \n\t" \
"pmaddhw $f27, "#src", $f20 \n\t" \
"pmaddhw $f28, "#src", $f22 \n\t" \
"pmaddhw $f29, "#src", $f24 \n\t" \
\
"punpcklwd $f30, $f26, $f26 \n\t" \
"punpckhwd $f31, $f26, $f26 \n\t" \
/* $f26: src[0], src[56] */ \
"paddw $f26, $f30, $f31 \n\t" \
"punpcklwd $f30, $f27, $f27 \n\t" \
"punpckhwd $f31, $f27, $f27 \n\t" \
/* $f27: src[8], src[48] */ \
"paddw $f27, $f30, $f31 \n\t" \
"punpcklwd $f30, $f28, $f28 \n\t" \
"punpckhwd $f31, $f28, $f28 \n\t" \
/* $f28: src[16], src[40] */ \
"paddw $f28, $f30, $f31 \n\t" \
"punpcklwd $f30, $f29, $f29 \n\t" \
"punpckhwd $f31, $f29, $f29 \n\t" \
/* $f29: src[24], src[32] */ \
"paddw $f29, $f30, $f31 \n\t" \
\
/* out1: src[24], src[16], src[8], src[0] */ \
/* out2: src[56], src[48], src[40], src[32] */ \
"punpcklhw $f30, $f26, $f27 \n\t" \
"punpcklhw $f31, $f28, $f29 \n\t" \
"punpckhwd "#out1", $f30, $f31 \n\t" \
"psrah "#out1", "#out1", $f16 \n\t" \
"punpcklhw $f30, $f27, $f26 \n\t" \
"punpcklhw $f31, $f29, $f28 \n\t" \
"punpckhwd "#out2", $f31, $f30 \n\t" \
"psrah "#out2", "#out2", $f16 \n\t"
#define IDCT_COL_CASE2(src1, src2, out1, out2) \
"pmaddhw $f28, "#src1", $f18 \n\t" \
"pmaddhw $f29, "#src2", $f19 \n\t" \
"paddw $f30, $f28, $f29 \n\t" \
"psubw $f31, $f28, $f29 \n\t" \
"punpcklwd $f28, $f30, $f31 \n\t" \
"punpckhwd $f29, $f30, $f31 \n\t" \
"pmaddhw $f30, "#src1", $f20 \n\t" \
"pmaddhw $f31, "#src2", $f21 \n\t" \
/* $f26: src[0], src[56] */ \
"paddw $f26, $f28, $f29 \n\t" \
"paddw $f28, $f30, $f31 \n\t" \
"psubw $f29, $f30, $f31 \n\t" \
"punpcklwd $f30, $f28, $f29 \n\t" \
"punpckhwd $f31, $f28, $f29 \n\t" \
"pmaddhw $f28, "#src1", $f22 \n\t" \
"pmaddhw $f29, "#src2", $f23 \n\t" \
/* $f27: src[8], src[48] */ \
"paddw $f27, $f30, $f31 \n\t" \
"paddw $f30, $f28, $f29 \n\t" \
"psubw $f31, $f28, $f29 \n\t" \
"punpcklwd $f28, $f30, $f31 \n\t" \
"punpckhwd $f29, $f30, $f31 \n\t" \
"pmaddhw $f30, "#src1", $f24 \n\t" \
"pmaddhw $f31, "#src2", $f25 \n\t" \
/* $f28: src[16], src[40] */ \
"paddw $f28, $f28, $f29 \n\t" \
"paddw "#out1", $f30, $f31 \n\t" \
"psubw "#out2", $f30, $f31 \n\t" \
"punpcklwd $f30, "#out1", "#out2" \n\t" \
"punpckhwd $f31, "#out1", "#out2" \n\t" \
/* $f29: src[24], src[32] */ \
"paddw $f29, $f30, $f31 \n\t" \
\
/* out1: src[24], src[16], src[8], src[0] */ \
/* out2: src[56], src[48], src[40], src[32] */ \
"punpcklhw "#out1", $f26, $f27 \n\t" \
"punpckhhw "#out2", $f27, $f26 \n\t" \
"punpcklhw $f30, $f28, $f29 \n\t" \
"punpckhhw $f31, $f29, $f28 \n\t" \
"punpckhwd "#out1", "#out1", $f30 \n\t" \
"punpckhwd "#out2", $f31, "#out2" \n\t" \
"psrah "#out1", "#out1", $f16 \n\t" \
"psrah "#out2", "#out2", $f16 \n\t"
/* idctSparseCol col0~3 */
/* $f17: ff_p16_32; $f16: COL_SHIFT-16 */
"gsldlc1 $f17, 0x57(%[w_arr]) \n\t"
"gsldrc1 $f17, 0x50(%[w_arr]) \n\t"
"li $10, 4 \n\t"
"dmtc1 $10, $f16 \n\t"
"paddh $f0, $f0, $f17 \n\t"
/* Transpose row[0,2,4,6] */
"punpcklhw $f26, $f0, $f4 \n\t"
"punpckhhw $f27, $f0, $f4 \n\t"
"punpcklhw $f28, $f8, $f12 \n\t"
"punpckhhw $f29, $f8, $f12 \n\t"
"punpcklwd $f0, $f26, $f28 \n\t"
"punpckhwd $f4, $f26, $f28 \n\t"
"punpcklwd $f8, $f27, $f29 \n\t"
"punpckhwd $f12, $f27, $f29 \n\t"
"or $f26, $f2, $f6 \n\t"
"or $f26, $f26, $f10 \n\t"
"or $f26, $f26, $f14 \n\t"
"dmfc1 $10, $f26 \n\t"
"bnez $10, 1f \n\t"
/* case1: In this case, row[1,3,5,7] are all zero */
/* col0: $f0: col[24,16,8,0]; $f2: col[56,48,40,32] */
IDCT_COL_CASE1($f0, $f0, $f2)
/* col1: $f4: col[25,17,9,1]; $f6: col[57,49,41,33] */
IDCT_COL_CASE1($f4, $f4, $f6)
/* col2: $f8: col[26,18,10,2]; $f10: col[58,50,42,34] */
IDCT_COL_CASE1($f8, $f8, $f10)
/* col3: $f12: col[27,19,11,3]; $f14: col[59,51,43,35] */
IDCT_COL_CASE1($f12, $f12, $f14)
"j 2f \n\t"
"1: \n\t"
/* case2: row[1,3,5,7] are not all zero */
/* Transpose */
"punpcklhw $f26, $f2, $f6 \n\t"
"punpckhhw $f27, $f2, $f6 \n\t"
"punpcklhw $f28, $f10, $f14 \n\t"
"punpckhhw $f29, $f10, $f14 \n\t"
"punpcklwd $f2, $f26, $f28 \n\t"
"punpckhwd $f6, $f26, $f28 \n\t"
"punpcklwd $f10, $f27, $f29 \n\t"
"punpckhwd $f14, $f27, $f29 \n\t"
/* col0: $f0: col[24,16,8,0]; $f2: col[56,48,40,32] */
IDCT_COL_CASE2($f0, $f2, $f0, $f2)
/* col1: $f4: col[25,17,9,1]; $f6: col[57,49,41,33] */
IDCT_COL_CASE2($f4, $f6, $f4, $f6)
/* col2: $f8: col[26,18,10,2]; $f10: col[58,50,42,34] */
IDCT_COL_CASE2($f8, $f10, $f8, $f10)
/* col3: $f12: col[27,19,11,3]; $f14: col[59,51,43,35] */
IDCT_COL_CASE2($f12, $f14, $f12, $f14)
"2: \n\t"
/* Transpose */
"punpcklhw $f26, $f0, $f4 \n\t"
"punpckhhw $f27, $f0, $f4 \n\t"
"punpcklhw $f28, $f8, $f12 \n\t"
"punpckhhw $f29, $f8, $f12 \n\t"
"punpcklwd $f0, $f26, $f28 \n\t"
"punpckhwd $f4, $f26, $f28 \n\t"
"punpcklwd $f8, $f27, $f29 \n\t"
"punpckhwd $f12, $f27, $f29 \n\t"
/* Transpose */
"punpcklhw $f26, $f2, $f6 \n\t"
"punpckhhw $f27, $f2, $f6 \n\t"
"punpcklhw $f28, $f10, $f14 \n\t"
"punpckhhw $f29, $f10, $f14 \n\t"
"punpcklwd $f2, $f26, $f28 \n\t"
"punpckhwd $f6, $f26, $f28 \n\t"
"punpcklwd $f10, $f27, $f29 \n\t"
"punpckhwd $f14, $f27, $f29 \n\t"
/* idctSparseCol col4~7 */
"paddh $f1, $f1, $f17 \n\t"
/* Transpose */
"punpcklhw $f26, $f1, $f5 \n\t"
"punpckhhw $f27, $f1, $f5 \n\t"
"punpcklhw $f28, $f9, $f13 \n\t"
"punpckhhw $f29, $f9, $f13 \n\t"
"punpcklwd $f1, $f26, $f28 \n\t"
"punpckhwd $f5, $f26, $f28 \n\t"
"punpcklwd $f9, $f27, $f29 \n\t"
"punpckhwd $f13, $f27, $f29 \n\t"
"or $f26, $f3, $f7 \n\t"
"or $f26, $f26, $f11 \n\t"
"or $f26, $f26, $f15 \n\t"
"dmfc1 $10, $f26 \n\t"
"bnez $10, 1f \n\t"
/* case1: In this case, row[1,3,5,7] are all zero */
/* col4: $f1: col[24,16,8,0]; $f3: col[56,48,40,32] */
IDCT_COL_CASE1($f1, $f1, $f3)
/* col5: $f5: col[25,17,9,1]; $f7: col[57,49,41,33] */
IDCT_COL_CASE1($f5, $f5, $f7)
/* col6: $f9: col[26,18,10,2]; $f11: col[58,50,42,34] */
IDCT_COL_CASE1($f9, $f9, $f11)
/* col7: $f13: col[27,19,11,3]; $f15: col[59,51,43,35] */
IDCT_COL_CASE1($f13, $f13, $f15)
"j 2f \n\t"
"1: \n\t"
/* case2: row[1,3,5,7] are not all zero */
/* Transpose */
"punpcklhw $f26, $f3, $f7 \n\t"
"punpckhhw $f27, $f3, $f7 \n\t"
"punpcklhw $f28, $f11, $f15 \n\t"
"punpckhhw $f29, $f11, $f15 \n\t"
"punpcklwd $f3, $f26, $f28 \n\t"
"punpckhwd $f7, $f26, $f28 \n\t"
"punpcklwd $f11, $f27, $f29 \n\t"
"punpckhwd $f15, $f27, $f29 \n\t"
/* col4: $f1: col[24,16,8,0]; $f3: col[56,48,40,32] */
IDCT_COL_CASE2($f1, $f3, $f1, $f3)
/* col5: $f5: col[25,17,9,1]; $f7: col[57,49,41,33] */
IDCT_COL_CASE2($f5, $f7, $f5, $f7)
/* col6: $f9: col[26,18,10,2]; $f11: col[58,50,42,34] */
IDCT_COL_CASE2($f9, $f11, $f9, $f11)
/* col7: $f13: col[27,19,11,3]; $f15: col[59,51,43,35] */
IDCT_COL_CASE2($f13, $f15, $f13, $f15)
"2: \n\t"
/* Transpose */
"punpcklhw $f26, $f1, $f5 \n\t"
"punpckhhw $f27, $f1, $f5 \n\t"
"punpcklhw $f28, $f9, $f13 \n\t"
"punpckhhw $f29, $f9, $f13 \n\t"
"punpcklwd $f1, $f26, $f28 \n\t"
"punpckhwd $f5, $f26, $f28 \n\t"
"punpcklwd $f9, $f27, $f29 \n\t"
"punpckhwd $f13, $f27, $f29 \n\t"
/* Transpose */
"punpcklhw $f26, $f3, $f7 \n\t"
"punpckhhw $f27, $f3, $f7 \n\t"
"punpcklhw $f28, $f11, $f15 \n\t"
"punpckhhw $f29, $f11, $f15 \n\t"
"punpcklwd $f3, $f26, $f28 \n\t"
"punpckhwd $f7, $f26, $f28 \n\t"
"punpcklwd $f11, $f27, $f29 \n\t"
"punpckhwd $f15, $f27, $f29 \n\t"
/* Store */
"gssqc1 $f1, $f0, 0x00(%[block]) \n\t"
"gssqc1 $f5, $f4, 0x10(%[block]) \n\t"
"gssqc1 $f9, $f8, 0x20(%[block]) \n\t"
"gssqc1 $f13, $f12, 0x30(%[block]) \n\t"
"gssqc1 $f3, $f2, 0x40(%[block]) \n\t"
"gssqc1 $f7, $f6, 0x50(%[block]) \n\t"
"gssqc1 $f11, $f10, 0x60(%[block]) \n\t"
"gssqc1 $f15, $f14, 0x70(%[block]) \n\t"
: [block]"+&r"(block)
: [w_arr]"r"(W_arr)
: "memory"
);
RECOVER_REG
}
void ff_simple_idct_put_8_mmi(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
{
ff_simple_idct_8_mmi(block);
ff_put_pixels_clamped_mmi(block, dest, line_size);
}
void ff_simple_idct_add_8_mmi(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
{
ff_simple_idct_8_mmi(block);
ff_add_pixels_clamped_mmi(block, dest, line_size);
}
Reference in New Issue View Git Blame Copy Permalink