RepoMirrors
/
ffmpeg
mirror of https://git.ffmpeg.org/ffmpeg.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

lavc/aacpsdsp: use restrict qualifier 

Except for add_squares, telling the compiler that the output vector(s)
cannot alias helps quite a bit (cycles on SiFive U74-MC):

ps_add_squares_c: 98277.7
ps_add_squares_r: 98320.2

ps_hybrid_analysis_c: 3731.2
ps_hybrid_analysis_r: 2495.7

ps_hybrid_analysis_ileave_c: 20478.0
ps_hybrid_analysis_ileave_r: 16092.2

ps_hybrid_synthesis_deint_c: 19051.5
ps_hybrid_synthesis_deint_r: 15420.0

ps_mul_pair_single_c: 122941.2
ps_mul_pair_single_r: 91035.0
This commit is contained in:
Rémi Denis-Courmont 2022-09-22 19:00:09 +03:00
parent 9c471c76cb
commit 51945d5ace
1 changed files with 15 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
libavcodec/aacpsdsp_template.c
View File

@ -26,24 +26,25 @@
#include "libavutil/attributes.h" #include "libavutil/attributes.h"
#include "aacpsdsp.h" #include "aacpsdsp.h"
static void ps_add_squares_c(INTFLOAT *dst, const INTFLOAT (*src)[2], int n) static void ps_add_squares_c(INTFLOAT *av_restrict dst,
const INTFLOAT (*src)[2], int n)
{ {
int i; for (int i = 0; i < n; i++)
for (i = 0; i < n; i++)
dst[i] += (UINTFLOAT)AAC_MADD28(src[i][0], src[i][0], src[i][1], src[i][1]); dst[i] += (UINTFLOAT)AAC_MADD28(src[i][0], src[i][0], src[i][1], src[i][1]);
} }
static void ps_mul_pair_single_c(INTFLOAT (*dst)[2], INTFLOAT (*src0)[2], INTFLOAT *src1, static void ps_mul_pair_single_c(INTFLOAT (*av_restrict dst)[2],
INTFLOAT (*src0)[2], INTFLOAT *src1,
int n) int n)
{ {
int i; for (int i = 0; i < n; i++) {
for (i = 0; i < n; i++) {
dst[i][0] = AAC_MUL16(src0[i][0], src1[i]); dst[i][0] = AAC_MUL16(src0[i][0], src1[i]);
dst[i][1] = AAC_MUL16(src0[i][1], src1[i]); dst[i][1] = AAC_MUL16(src0[i][1], src1[i]);
} }
} }
static void ps_hybrid_analysis_c(INTFLOAT (*out)[2], INTFLOAT (*in)[2], static void ps_hybrid_analysis_c(INTFLOAT (*av_restrict out)[2],
INTFLOAT (*in)[2],
const INTFLOAT (*filter)[8][2], const INTFLOAT (*filter)[8][2],
ptrdiff_t stride, int n) ptrdiff_t stride, int n)
{ {
@ -76,13 +77,12 @@ static void ps_hybrid_analysis_c(INTFLOAT (*out)[2], INTFLOAT (*in)[2],
} }
} }
static void ps_hybrid_analysis_ileave_c(INTFLOAT (*out)[32][2], INTFLOAT L[2][38][64], static void ps_hybrid_analysis_ileave_c(INTFLOAT (*av_restrict out)[32][2],
int i, int len) INTFLOAT L[2][38][64],
int i, int len)
{ {
int j;
for (; i < 64; i++) { for (; i < 64; i++) {
for (j = 0; j < len; j++) { for (int j = 0; j < len; j++) {
out[i][j][0] = L[0][j][i]; out[i][j][0] = L[0][j][i];
out[i][j][1] = L[1][j][i]; out[i][j][1] = L[1][j][i];
} }
@ -90,13 +90,11 @@ static void ps_hybrid_analysis_ileave_c(INTFLOAT (*out)[32][2], INTFLOAT L[2][38
} }
static void ps_hybrid_synthesis_deint_c(INTFLOAT out[2][38][64], static void ps_hybrid_synthesis_deint_c(INTFLOAT out[2][38][64],
INTFLOAT (*in)[32][2], INTFLOAT (*av_restrict in)[32][2],
int i, int len) int i, int len)
{ {
int n;
for (; i < 64; i++) { for (; i < 64; i++) {
for (n = 0; n < len; n++) { for (int n = 0; n < len; n++) {
out[0][n][i] = in[i][n][0]; out[0][n][i] = in[i][n][0];
out[1][n][i] = in[i][n][1]; out[1][n][i] = in[i][n][1];
} }