openharmony/third_party_ffmpeg
1
0
Fork 0
mirror of https://gitee.com/openharmony/third_party_ffmpeg synced 2024-11-23 19:30:05 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

lavc/cook: get rid of wasteful pow in init_pow2table

Browse Source 
The table is highly structured, so pow (or exp2 for that matter) can entirely
be avoided, yielding a ~ 40x speedup with no loss of accuracy.

sample benchmark (Haswell, GNU/Linux):
new:
4449 decicycles in init_pow2table(loop 1000),     254 runs,      2 skips
4411 decicycles in init_pow2table(loop 1000),     510 runs,      2 skips
4391 decicycles in init_pow2table(loop 1000),    1022 runs,      2 skips

old:
183673 decicycles in init_pow2table(loop 1000),     256 runs,      0 skips
182142 decicycles in init_pow2table(loop 1000),     512 runs,      0 skips
182104 decicycles in init_pow2table(loop 1000),    1024 runs,      0 skips

Reviewed-by: Clément Bœsch <u@pkh.me>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
This commit is contained in:
Ganesh Ajjanagadde 2015-12-16 17:39:28 -08:00
parent b272c3a5aa
commit 05434b0eea
1 changed files with 9 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
libavcodec/cook.c
View File

@ -166,10 +166,17 @@ static float rootpow2tab[127];
/* table generator */ /* table generator */
static av_cold void init_pow2table(void) static av_cold void init_pow2table(void)
{ {
/* fast way of computing 2^i and 2^(0.5*i) for -63 <= i < 64 */
int i; int i;
static const float exp2_tab[2] = {1, M_SQRT2};
float exp2_val = powf(2, -63);
float root_val = powf(2, -32);
for (i = -63; i < 64; i++) { for (i = -63; i < 64; i++) {
pow2tab[63 + i] = pow(2, i); if (!(i & 1))
rootpow2tab[63 + i] = sqrt(pow(2, i)); root_val *= 2;
pow2tab[63 + i] = exp2_val;
rootpow2tab[63 + i] = root_val * exp2_tab[i & 1];
exp2_val *= 2;
} }
} }