summaryrefslogtreecommitdiff
path: root/src/gallium/drivers
AgeCommit message (Collapse)Author
2009-08-29llvmpipe: Factor out shared test code into a separate module.José Fonseca
2009-08-29llvmpipe: Cleanup constant helpers.José Fonseca
2009-08-29llvmpipe: Move intrinsic helpers to a separate module.José Fonseca
2009-08-29llvmpipe: Collect richer blend data.José Fonseca
2009-08-29llvmpipe: Some notes about PSHUF.José Fonseca
2009-08-29llvmpipe: Optimize blend swizzles by using bitmasks instead of shuffles for ↵José Fonseca
ubytes.
2009-08-29llvmpipe: Write data to a tsv file for posterior analysis.José Fonseca
2009-08-29llvmpipe: Move p_build_context to lp_bld_type.hJosé Fonseca
As it will be shared with more modules.
2009-08-29llvmpipe: Don't use llvm.readcyclecounter.José Fonseca
LLVM during optimization reorders the rdtsc instructions next to each other, which makes it pretty useless.
2009-08-29llvmpipe: Separate constant building.José Fonseca
2009-08-29llvmpipe: Move type support functions into a separate file.José Fonseca
2009-08-29llvmpipe: Measure the number of cycles taken for blending.José Fonseca
2009-08-29llvmpipe: Minor comments / debug mnemonics.José Fonseca
2009-08-29llvmpipe: Optimize away min/max with equal operands.José Fonseca
2009-08-29llvmpipe: Improve generated 8bit blending code.José Fonseca
2009-08-29llvmpipe: Get blending of normalized 8bit unsigned integers working.José Fonseca
2009-08-29llvmpipe: Specialize arithmetic operations.José Fonseca
2009-08-29llvmpipe: Introduce a custom typing system.José Fonseca
Straightforward representation of floating-point/fixed-point/integer, normalized/scaled, signed/unsigned SIMD vector types.
2009-08-29llvmpipe: Avoid C99-isms.José Fonseca
2009-08-29llvmpipe: Factor out intrisic calling code into a reusable function.José Fonseca
2009-08-29llvmpipe: Cleanup test programs.José Fonseca
2009-08-29llvmpipe: Blending.José Fonseca
The code
2009-08-29llvmpipe: Arithmetic helpers.José Fonseca
2009-08-29llvmpipe: fastpath for interpolated z16 less depthtestingKeith Whitwell
Because this is interpolated (ie. early) depth, we can build in an assumption about the quads emitted by triangle setup, ie that they are actually linear spans. Interpolate z over those spans in z16 format to save on math & conversion.
2009-08-29llvmpipe: cope with nr_cbufs == 0Keith Whitwell
Disable blend code when no color buffer
2009-08-29llvmpipe: Translate logicops.José Fonseca
2009-08-29llvmpipe: Separate pixel packing/unpacking from loading/storing.José Fonseca
2009-08-29llvmpipe: Adjust for pixel layout changes.José Fonseca
2009-08-29llvmpipe: Fix typo in copyright.José Fonseca
2009-08-29llvmpipe: Add test cases for pixels.José Fonseca
Actually some of this test cases need more testing...
2009-08-29llvmpipe: Update SConscript.José Fonseca
2009-08-29llvmpipe: Simple shell script to aid port softpipe changes to llvmpipe.José Fonseca
2009-08-29llvmpipe: move all depth/stencil/alpha pixel processing into one stageKeith Whitwell
2009-08-29llvmpipe: fix error in scissor state dependenciesKeith Whitwell
2009-08-29llvmpipe: cleanup framebuffer state routine slightlyKeith Whitwell
2009-08-29llvmpipe: move all color-combine code into lp_quad_blend.cKeith Whitwell
Consolidate the read-modify-write color combining code from the blend, colormask and output stages.
2009-08-29llvmpipe: fix typoKeith Whitwell
2009-08-29llvmpipe: example fastpaths in blendingKeith Whitwell
2009-08-29llvmpipe: actually pass >1 quad from triangle routineKeith Whitwell
First attempt
2009-08-29llvmpipe: expand quad pipeline to process >1 quad at a timeKeith Whitwell
This is part one -- we still only pass a single quad down, but the code can now cope with more. The quads must all be from the same tile.
2009-08-29llvmpipe: rip out old mulithread supportKeith Whitwell
2009-08-29llvmpipe: avoid flushing depth buffer cache on swapbuffersKeith Whitwell
There's no need to push out depth buffer contents on swapbuffers. Note that this change doesn't throw away depth buffer changes, it simply holds them in the cache over calls to swapbuffers. The hope is that swapbuffers will be followed by a clear() which means in that case we won't have to write the changes out.
2009-08-29llvmpipe: also shortcircuit non-texture tile lookupsKeith Whitwell
2009-08-29llvmpipe: shortcircuit repeated lookups of the same tileKeith Whitwell
The lp_tile_cache is often called repeatedly to look up the same tile. Add a cache (to the cache) of the single tile most recently retreived and make a quick inline check to see if this matches the subsequent request. Add a tile_address bitfield struct to make this check easier.
2009-08-29llvmpipe: remove unused vars in lp_setup.cKeith Whitwell
2009-08-29llvmpipe: use bitwise logic to setup quad masks in lp_setupKeith Whitwell
2009-08-29llvmpipe: simplify flush_spansKeith Whitwell
No loss of performance, but simpler code.
2009-08-29llvmpipe: make some small steps to flush texture cache less frequentlyKeith Whitwell
No performance gain yet, but the code is a bit cleaner.
2009-08-29llvmpipe: remove backwards dependency from tilecache to llvmpipeKeith Whitwell
The tile cache is a utility, it shouldn't know anything about the entity which is making use of it (ie llvmpipe). Remove llvmpipe parameter to all the tilecache function calls, and also remove the need to keep a llvmpipe pointer in the sampler structs.
2009-08-29llvmpipe: Rename preprocessor symbols too.José Fonseca