zdoom
/
raze-gles
mirror of https://github.com/ZDoom/raze-gles.git
0
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

VPX: in 3 planes -> packed format conversion code, group together the 

three individual loops and compile the enclosing function at -O3 (-O1 for
debugging builds). Now, the time for this conversion ranges from 7 to 18
ms per frame across various tested machines, a clear improvement.

git-svn-id: https://svn.eduke32.com/eduke32@2042 1a8010ca-5511-0410-912e-c29ae57300e0
This commit is contained in:
helixhorned 2011-09-25 15:11:11 +00:00
parent fb2f01911c
commit 195ed466e3
1 changed files with 7 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
polymer/eduke32/source/animvpx.c
View File

@ -208,6 +208,11 @@ const char *animvpx_nextpic_errmsg[] = {
// retrieves one picture-frame from the stream
// pic format: lines of [Y U V 0] pixels
// *picptr==NULL means EOF has been reached
#ifdef DEBUGGINGAIDS
ATTRIBUTE((optimize("O1")))
#else
ATTRIBUTE((optimize("O3")))
#endif
int32_t animvpx_nextpic(animvpx_codec_ctx *codec, uint8_t **picptr)
{
int32_t ret, corrupted;
@ -277,14 +282,11 @@ read_ivf_frame:
/*** 3 planes --> packed conversion ***/
for (y=0; y<img->d_h; y++)
for (x=0; x<img->d_w; x++)
{
codec->pic[(img->d_w*y + x)<<2] = img->planes[VPX_PLANE_Y][img->stride[VPX_PLANE_Y]*y + x];
for (y=0; y<img->d_h; y++)
for (x=0; x<img->d_w; x++)
codec->pic[((img->d_w*y + x)<<2) + 1] = img->planes[VPX_PLANE_U][img->stride[VPX_PLANE_U]*(y>>1) + (x>>1)];
for (y=0; y<img->d_h; y++)
for (x=0; x<img->d_w; x++)
codec->pic[((img->d_w*y + x)<<2) + 2] = img->planes[VPX_PLANE_V][img->stride[VPX_PLANE_V]*(y>>1) + (x>>1)];
}
*picptr = codec->pic;
return 0;