Age | Commit message (Collapse) | Author |
|
|
|
spe_splat()
|
|
|
|
|
|
|
|
|
|
Basically, do as the comment says.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Do code generation for alpha test, z test, stencil, blend, colormask
and framebuffer/tile read/write as a single code block.
Ian's previous blend/z/stencil test code is still there but mostly disabled
and will be removed soon.
|
|
Fix incorrect opcode for fsmbi.
Added "macro" functions for loading floats/ints, register complement, zero, move.
Added #defines for return address and stack pointer registers.
Added assertions to check that the instruction buffer doesn't overflow.
|
|
|
|
|
|
|
|
|
|
|
|
Conflicts:
configs/default
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
softpipe.
We want to make it env variable, or even better, autodetect as the feature makes
softpipe run slower on a single CPU.
|
|
Store only input and inout of a quad_header in job que.
|
|
parts.
|
|
Use condition vars to communicate between threads instead of stalling.
|
|
Configured for 2 cores.
|
|
|
|
|
|
This avoids splitting the bitmaps in many files.
|
|
This allows quick inspection of surfaces in mass scale.
|
|
|
|
|
|
|
|
|
|
For blocks whose contents are derived completely from ref surfaces, don't
bother zero-ing the corresponding block in the luma/chroma textures, except
for the first such luma, chroma Cb, and chroma Cr each frame. All later
zero blocks are textured from that first zero block.
|
|
Need to be able to address each texture separately to do zero-block.
|