[FFmpeg-devel] [PATCH 2/2] lavc/vorbisdec: use intermediate variables

[FFmpeg-devel] [PATCH 2/2] lavc/vorbisdec: use intermediate variables

[remi]

From: Rémi Denis-Courmont <remi@remlab.net>

The compiler cannot infer that the two float vectors do not alias,
causing unnecessary extra loads and serialisation. This patch caches
the two input values in local variables so that compiler can optimise
individual loop iterations.
---
 libavcodec/vorbisdec.c | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/libavcodec/vorbisdec.c b/libavcodec/vorbisdec.c
index 9728997370..74e71936ab 100644
--- a/libavcodec/vorbisdec.c
+++ b/libavcodec/vorbisdec.c
@@ -1583,21 +1583,21 @@ void ff_vorbis_inverse_coupling(float *mag, float *ang, intptr_t blocksize)
 {
     ptrdiff_t i;
     for (i = 0;  i < blocksize;  i++) {
-        if (mag[i] > 0.0) {
-            if (ang[i] > 0.0) {
-                ang[i] = mag[i] - ang[i];
+        float angi = ang[i], magi = mag[i];
+
+        if (magi > 0.f) {
+            if (angi > 0.f) {
+                ang[i] = magi - angi;
             } else {
-                float temp = ang[i];
-                ang[i]     = mag[i];
-                mag[i]    += temp;
+                ang[i] = magi;
+                mag[i] = magi + angi;
             }
         } else {
-            if (ang[i] > 0.0) {
-                ang[i] += mag[i];
+            if (angi > 0.f) {
+                ang[i] = magi + angi;
             } else {
-                float temp = ang[i];
-                ang[i]     = mag[i];
-                mag[i]    -= temp;
+                ang[i] = magi;
+                mag[i] = magi - angi;
             }
         }
     }
-- 
2.37.2

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-devel] [PATCH 2/2] lavc/vorbisdec: use intermediate variables

[Andreas Rheinhardt]

remi@remlab.net:
> From: Rémi Denis-Courmont <remi@remlab.net>
> 
> The compiler cannot infer that the two float vectors do not alias,
> causing unnecessary extra loads and serialisation. This patch caches
> the two input values in local variables so that compiler can optimise
> individual loop iterations.
> ---

We have av_restrict. (No opinion on this patch, it's just a remark.)

>  libavcodec/vorbisdec.c | 22 +++++++++++-----------
>  1 file changed, 11 insertions(+), 11 deletions(-)
> 
> diff --git a/libavcodec/vorbisdec.c b/libavcodec/vorbisdec.c
> index 9728997370..74e71936ab 100644
> --- a/libavcodec/vorbisdec.c
> +++ b/libavcodec/vorbisdec.c
> @@ -1583,21 +1583,21 @@ void ff_vorbis_inverse_coupling(float *mag, float *ang, intptr_t blocksize)
>  {
>      ptrdiff_t i;
>      for (i = 0;  i < blocksize;  i++) {
> -        if (mag[i] > 0.0) {
> -            if (ang[i] > 0.0) {
> -                ang[i] = mag[i] - ang[i];
> +        float angi = ang[i], magi = mag[i];
> +
> +        if (magi > 0.f) {
> +            if (angi > 0.f) {
> +                ang[i] = magi - angi;
>              } else {
> -                float temp = ang[i];
> -                ang[i]     = mag[i];
> -                mag[i]    += temp;
> +                ang[i] = magi;
> +                mag[i] = magi + angi;
>              }
>          } else {
> -            if (ang[i] > 0.0) {
> -                ang[i] += mag[i];
> +            if (angi > 0.f) {
> +                ang[i] = magi + angi;
>              } else {
> -                float temp = ang[i];
> -                ang[i]     = mag[i];
> -                mag[i]    -= temp;
> +                ang[i] = magi;
> +                mag[i] = magi - angi;
>              }
>          }
>      }

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".