Age | Commit message (Collapse) | Author |
|
|
|
vec2 v; v.x = v.y = 1.0; // chained assignment
vec4 v; v.zx = vec2(a,b); // swizzled writemask
|
|
instruction saturate-write option. Not finished yet.
|
|
|
|
|
|
|
|
of registers, rather than crash.
|
|
|
|
|
|
share a single register. Clean-up needed.
|
|
|
|
|
|
|
|
|
|
|
|
Improved shader error handling.
|
|
we can either jump on zero, or non-zero predicate.
|
|
|
|
float[4] register slot.
|
|
p and q works correctly.
|
|
|
|
|
|
|
|
|
|
emit time, not codegen time.
|
|
|
|
and expression temporarires. Much better register utilization now.
Lots of other fixes.
The OpenGL GLSL "orange book" brick shader demo works now.
|
|
temporaries.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
properly.
|
|
|
|
|
|
textures nowadays
|
|
|
|
|
|
|
|
|
|
|
|
Implement _mesa_uniform_matrix()
Support for program parameters/uniforms with more than 4 elements.
Store 4x4 matrices in column-major order in registers.
Update mat mul built-in functions accordingly.
|
|
|
|
function in linker.
|
|
|
|
|