diff options
| author | Jonathan Adamczewski <jadamcze@utas.edu.au> | 2009-05-21 08:18:03 -0600 |
|---|---|---|
| committer | Brian Paul <brianp@vmware.com> | 2009-05-21 08:18:03 -0600 |
| commit | b4824520ecf453cd8de90e57e839cb11a698d9c0 (patch) | |
| tree | 2ec3bd7a633fd704c9b04e6acf9f6760bc4418da /src/gallium/drivers/cell/spu/spu_tri.h | |
| parent | 5b27b4ad37bd992d2d3a6fd9d407277113544f30 (diff) | |
cell: unroll inner loop of spu_render.c:cmd_render()
It was taking approximately 50 cycles to extract the vertex indices,
calculate the vertex_header pointers and call tri_draw() for each
three vertices - .
Unrolled, it takes less than 100 cycles to extract, unpack,
calculate pointers and call tri_draw() eight times. It does have a
nasty jump-tabled switch. I'm sure that there's a better way...
Code size of spu_render.o gets larger due to the extra constants and
work in the inner loop, there are extra stack saves and loads
because there are more registers in use, and an assert. spu_tri.o
gets a little smaller.
Diffstat (limited to 'src/gallium/drivers/cell/spu/spu_tri.h')
| -rw-r--r-- | src/gallium/drivers/cell/spu/spu_tri.h | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/gallium/drivers/cell/spu/spu_tri.h b/src/gallium/drivers/cell/spu/spu_tri.h index aa694dd7c9..82e3b19ad7 100644 --- a/src/gallium/drivers/cell/spu/spu_tri.h +++ b/src/gallium/drivers/cell/spu/spu_tri.h @@ -31,7 +31,7 @@ extern boolean -tri_draw(const float *v0, const float *v1, const float *v2, uint tx, uint ty); +tri_draw(const qword vs, uint tx, uint ty); #endif /* SPU_TRI_H */ |