Dean c3625c4789 Merge pull request #4120 from tibzejoker/fix/vector-window-direct-ssa ...

fix(shader): keep direct VGPR accesses in SSA when the window has only relative reads

2026-10-13 14:47:39 +13:00

..

ControlFlow

fix(shader): move the primitive class to request version 19 and compose the ancillary word with render target index 0

2026-10-13 13:18:51 +13:00

IntermediateRepresentation

fix(shader): keep direct VGPR accesses in SSA when the window has only relative reads

2026-10-13 13:59:35 +13:00

Optimization

fix(shader): read the ancillary render target index as 0

2026-10-13 13:23:56 +13:00

RdnaDecoder

Merge pull request #3827 from CouchRiv/fix-shader-combined-cvt-i16-f16-omod

2026-10-12 00:39:20 +02:00

SpirvBackend

feat(shader): give reads the masked-select pass keeps the whole pixel ancillary word

2026-10-13 13:06:21 +13:00

tests

fix(shader): handle cache file allocation failures

2026-10-11 19:57:45 +04:00

Translation

fix(shader): keep direct VGPR accesses in SSA when the window has only relative reads

2026-10-13 13:59:35 +13:00

BdaAbi.hpp

fix(agc): note BDA-written pages at full address width and read back only the claimed ones

2026-10-11 12:54:02 +05:30

CacheKey.hpp

fix(recompiler): run vertex programs as one-lane waves without vertex subgroups

2026-10-12 08:50:05 -04:00

CMakeLists.txt

perf(agc): build the shader pipeline libraries when graphics stages are resolved at registration

2026-10-12 09:50:41 +02:00

CompiledVariant.hpp

refactor(shader): remove resource specialization from artifact key

2026-10-08 17:26:38 +03:00

Deserialize.cpp

perf(recompiler): keep descriptor heaps uncompacted and read image modes at run time when modules are built at registration

2026-10-12 09:50:32 +02:00

FragmentOutputSpecialization.hpp

feat(agc): implement dual-source blending from the MRT1 export

2026-10-10 10:29:00 +02:00

PipelineSpecialization.hpp

feat(shader): give reads the masked-select pass keeps the whole pixel ancillary word

2026-10-13 13:06:21 +13:00

Recompiler.cpp

feat(shader): give reads the masked-select pass keeps the whole pixel ancillary word

2026-10-13 13:06:21 +13:00

Recompiler.hpp

feat(shader): give reads the masked-select pass keeps the whole pixel ancillary word

2026-10-13 13:06:21 +13:00

RuntimeAbi.hpp

perf(recompiler): read vertex attribute component selectors at run time when modules are built at registration

2026-10-12 09:50:41 +02:00

ShaderCacheDirectory.hpp

Merge pull request #3785 from MrMha/fix/shader-cache-read-allocation

2026-10-12 00:01:43 +02:00

ShaderCacheVersion.cmake

perf(shader): hash only the recompiler's inputs into the shader cache version

2026-10-04 03:02:24 +13:00

ShaderDiskCache.cpp

feat(shader): give reads the masked-select pass keeps the whole pixel ancillary word

2026-10-13 13:06:21 +13:00

ShaderDiskCache.hpp

feat(shader): give reads the masked-select pass keeps the whole pixel ancillary word

2026-10-13 13:06:21 +13:00

ThreadOwned.hpp

fix(agc): free per-thread driver state at thread exit without emutls

2026-10-04 19:08:54 -05:00

VertexInputSpecialization.hpp

perf(shader): restore prepared vertex inputs + harden buffer specialization

2026-10-08 17:28:55 +03:00