generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Changed above and left context data layout

Browse Source 
The main reason for the change was to reduce cycles in the token
decoder. (~1.5% gain for 32 bit)  This layout should be more
cache friendly.

As a result of this change, the encoder had to be updated.

Change-Id: Id5e804169d8889da0378b3a519ac04dabd28c837
Note: dixie uses a similar layout
This commit is contained in:
Scott LaVarnway
2010-08-31 10:49:57 -04:00
parent aaad6d1b54
commit e85e631504
15 changed files with 216 additions and 359 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
vp8/decoder/threading.c
View File

@@ -157,7 +157,7 @@ THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)
int ithread = ((DECODETHREAD_DATA *)p_data)->ithread;
VP8D_COMP *pbi = (VP8D_COMP *)(((DECODETHREAD_DATA *)p_data)->ptr1);
MB_ROW_DEC *mbrd = (MB_ROW_DEC *)(((DECODETHREAD_DATA *)p_data)->ptr2);
ENTROPY_CONTEXT mb_row_left_context[4][4];
ENTROPY_CONTEXT_PLANES mb_row_left_context;
while (1)
{
@@ -197,12 +197,9 @@ THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)
recon_uvoffset = mb_row * recon_uv_stride * 8;
// reset above block coeffs
xd->above_context[Y1CONTEXT] = pc->above_context[Y1CONTEXT];
xd->above_context[UCONTEXT ] = pc->above_context[UCONTEXT];
xd->above_context[VCONTEXT ] = pc->above_context[VCONTEXT];
xd->above_context[Y2CONTEXT] = pc->above_context[Y2CONTEXT];
xd->left_context = mb_row_left_context;
vpx_memset(mb_row_left_context, 0, sizeof(mb_row_left_context));
xd->above_context = pc->above_context;
xd->left_context = &mb_row_left_context;
vpx_memset(&mb_row_left_context, 0, sizeof(mb_row_left_context));
xd->up_available = (mb_row != 0);
xd->mb_to_top_edge = -((mb_row * 16)) << 3;
@@ -260,10 +257,7 @@ THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)
++xd->mode_info_context; /* next mb */
xd->above_context[Y1CONTEXT] += 4;
xd->above_context[UCONTEXT ] += 2;
xd->above_context[VCONTEXT ] += 2;
xd->above_context[Y2CONTEXT] ++;
xd->above_context++;
//pbi->mb_row_di[ithread].current_mb_col = mb_col;
pbi->current_mb_col[mb_row] = mb_col;
@@ -604,15 +598,12 @@ void vp8_mtdecode_mb_rows(VP8D_COMP *pbi,
if (mb_row > 0)
last_row_current_mb_col = &pbi->current_mb_col[mb_row -1];
vpx_memset(pc->left_context, 0, sizeof(pc->left_context));
vpx_memset(&pc->left_context, 0, sizeof(pc->left_context));
recon_yoffset = mb_row * recon_y_stride * 16;
recon_uvoffset = mb_row * recon_uv_stride * 8;
// reset above block coeffs
xd->above_context[Y1CONTEXT] = pc->above_context[Y1CONTEXT];
xd->above_context[UCONTEXT ] = pc->above_context[UCONTEXT];
xd->above_context[VCONTEXT ] = pc->above_context[VCONTEXT];
xd->above_context[Y2CONTEXT] = pc->above_context[Y2CONTEXT];
xd->above_context = pc->above_context;
xd->up_available = (mb_row != 0);
xd->mb_to_top_edge = -((mb_row * 16)) << 3;
@@ -672,10 +663,7 @@ void vp8_mtdecode_mb_rows(VP8D_COMP *pbi,
++xd->mode_info_context; /* next mb */
xd->above_context[Y1CONTEXT] += 4;
xd->above_context[UCONTEXT ] += 2;
xd->above_context[VCONTEXT ] += 2;
xd->above_context[Y2CONTEXT] ++;
xd->above_context++;
//pbi->current_mb_col_main = mb_col;
pbi->current_mb_col[mb_row] = mb_col;