Re: [FFmpeg-devel] [PATCH 1/9] lavfi/nlmeans: random code shuffling to help compiler

[Michael Niedermayer]

On Sun, May 06, 2018 at 01:40:52PM +0200, Clément Bœsch wrote:
> This makes nlmeans_slice() slightly faster at least on GCC 7.3.
> ---
>  libavfilter/vf_nlmeans.c | 8 
>  1 file changed, 4 insertions(+), 4 deletions(-)

LGTM

thx

[...]

-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Modern terrorism, a quick summary: Need oil, start war with country that
has oil, kill hundread thousand in war. Let country fall into chaos,
be surprised about raise of fundamantalists. Drop more bombs, kill more
people, be surprised about them taking revenge and drop even more bombs
and strip your own citizens of their rights and freedoms. to be continued


signature.asc
Description: PGP signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH 1/9] lavfi/nlmeans: random code shuffling to help compiler

[Clément Bœsch]

This makes nlmeans_slice() slightly faster at least on GCC 7.3.
---
 libavfilter/vf_nlmeans.c | 8 
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/libavfilter/vf_nlmeans.c b/libavfilter/vf_nlmeans.c
index e4952e187e..d222d3913e 100644
--- a/libavfilter/vf_nlmeans.c
+++ b/libavfilter/vf_nlmeans.c
@@ -368,7 +368,6 @@ static int nlmeans_slice(AVFilterContext *ctx, void *arg, 
int jobnr, int nb_jobs
 int x, y;
 NLMeansContext *s = ctx->priv;
 const struct thread_data *td = arg;
-const uint8_t *src = td->src;
 const int src_linesize = td->src_linesize;
 const int process_h = td->endy - td->starty;
 const int slice_start = (process_h *  jobnr   ) / nb_jobs;
@@ -377,14 +376,15 @@ static int nlmeans_slice(AVFilterContext *ctx, void *arg, 
int jobnr, int nb_jobs
 const int endy   = td->starty + slice_end;
 
 for (y = starty; y < endy; y++) {
+const uint8_t *src = td->src + y*src_linesize;
+struct weighted_avg *wa = s->wa + y*s->wa_linesize;
 for (x = td->startx; x < td->endx; x++) {
 const int patch_diff_sq = get_integral_patch_value(td->ii_start, 
s->ii_lz_32, x, y, td->p);
 if (patch_diff_sq < s->max_meaningful_diff) {
-struct weighted_avg *wa = >wa[y*s->wa_linesize + x];
 const int weight_lut_idx = patch_diff_sq * s->pdiff_lut_scale;
 const double weight = s->weight_lut[weight_lut_idx]; // 
exp(-patch_diff_sq * s->pdiff_scale)
-wa->total_weight += weight;
-wa->sum += weight * src[y*src_linesize + x];
+wa[x].total_weight += weight;
+wa[x].sum += weight * src[x];
 }
 }
 }
-- 
2.17.0

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel