Dean e3d03fbae6 Merge pull request #7930 from AnthonyClemens/perf/agc-skip-empty-indirect ...

perf(agc): skip CPU-path indirect draws that draw nothing before buil…

2026-10-23 23:45:49 +13:00

..

execution

Merge pull request #7930 from AnthonyClemens/perf/agc-skip-empty-indirect

2026-10-23 23:45:49 +13:00

shaders

test(agc): retile distinct D32 sample readback

2026-10-21 16:10:59 +03:00

AllocationFailure.cpp

fix(agc): release unpublished buffer slabs on allocation failure

2026-10-16 17:40:21 +02:00

AllocationFailure.hpp

fix(agc): release unpublished buffer slabs on allocation failure

2026-10-16 17:40:21 +02:00

AsyncLog.cpp

fix(tests): keep the async log test's aborting children away from the crash handler

2026-10-19 00:42:45 +00:00

BarrierBatch.cpp

perf(agc): skip the staging and write-back leading barriers a trailing barrier already covers

2026-10-18 10:50:02 +02:00

BdaContracts.cpp

perf(shader): emit runtime V# byte writes through the looping helper

2026-10-18 13:55:47 +00:00

BdaDevice.cpp

Merge pull request #8167 from MarcZX10/codex/8157-bda-device-gpu

2026-10-21 15:44:18 +13:00

BdaExecution.cpp

Merge pull request #7829 from JJenkx/fix/shader-bda-anchor-read-sync114

2026-10-20 16:05:32 +02:00

BdaResources.cpp

Merge pull request #7647 from fma-rc/fix/agc-heap-mirror-resident

2026-10-23 12:45:26 +02:00

BdaShader.cpp

Merge main into fix/bda-unaligned

2026-10-18 18:53:03 +02:00

BdaShader.hpp

Merge main into fix/bda-unaligned

2026-10-18 18:53:03 +02:00

BdaTests.hpp

fix(agc): copy a sparse heap whose mirror cannot be made, as main does

2026-10-20 22:59:49 +02:00

BufferPool.cpp

perf(agc): keep released image memory in the buffer pool for reuse by type and size

2026-10-19 21:30:25 +03:00

BufferResourceVcc.cpp

…

CmaskFastClear.cpp

…

ColorTargetLayout.cpp

…

ColorTransferTests.cpp

…

ColorTransferTests.hpp

…

CommandBatch.cpp

fix(agc): retain submitted batches when profiling fails

2026-10-17 13:41:42 +02:00

CommandBatchFailure.cpp

fix(agc): retain submitted batches when profiling fails

2026-10-17 13:41:42 +02:00

CommandBatchFailure.hpp

fix(agc): retain submitted batches when profiling fails

2026-10-17 13:41:42 +02:00

ComputePipelineCapture.cpp

fix(agc): ignore Finder .DS_Store files in capture directories

2026-10-18 21:26:49 -03:00

ComputePipelineCaptureRun.cmake

…

ComputePipelineReplay.cpp

…

ConstantBitTest.cpp

fix(shader): keep the 24-bit factor reads of 16-bit multiplies

2026-10-20 15:18:13 +07:00

ConstantPhi.cpp

fix(shader): collect packed ancillary readers without recursion

2026-10-17 17:32:02 -04:00

ConstantSelect.cpp

perf(shader): fold a select nested in a select on the same condition

2026-10-17 17:03:09 +00:00

ControlFlow.cpp

fix(shader): refuse a call link escape only at the wave size it escapes under

2026-10-20 18:13:18 +03:00

DccKeySharing.cpp

…

DenormalFlushElimination.cpp

perf(shader): drop the result flush of a product behind its tiny-product check

2026-10-17 16:34:08 +00:00

DepthRangeRemap.cpp

test(agc): read the prepared module's specialization constants from the artifact's SPIR-V

2026-10-18 22:14:53 +13:00

DepthSurfaceReuse.cpp

Merge pull request #7318 from MarcZX10/codex/msaa-depth-sampling-shadow

2026-10-19 20:50:38 +02:00

DescriptorCacheFailure.cpp

fix(agc): release unpublished descriptor cache objects

2026-10-17 15:41:44 +02:00

DeviceFault.cpp

feat(agc): record GPU checkpoints through VK_AMD_buffer_marker

2026-10-19 08:15:43 -03:00

DeviceMemoryFallback.cpp

fix(agc): release the texture caches before placing any image in host memory

2026-10-18 19:29:47 +02:00

DisplayBufferDcc.cpp

…

DppWrite.cpp

…

DrawInputsInPlace.cpp

perf(agc): copy draw inputs on NVIDIA's Windows driver instead of binding host imports in place

2026-10-19 11:48:36 +02:00

DrawProgramTrace.cpp

feat(agc): trace shader addresses for each draw

2026-10-14 23:39:27 +02:00

EmbeddedVertexFetch.cpp

test(recompiler): reject encoded modified offset forms

2026-10-16 18:56:16 +02:00

EmbeddedVertexFetchAliasing.cpp

…

ExecutionTestNotBuilt.cpp

…

F16SingleConvert.cpp

perf(shader): convert one f16 value instead of an (x, 0.0) pair

2026-10-19 03:07:05 +02:00

F32DenormPreserve.cpp

feat(shader): request DenormPreserve 32 where the device preserves f32 denormals

2026-10-21 10:45:24 +13:00

Flip.cpp

feat(agcDriver): execute a flip submitted on a compute queue

2026-10-17 08:17:45 +02:00

Float16Conversion.cpp

…

Float32FlushControl.cpp

perf(agc): probe the f32 denormal flush on MoltenVK

2026-10-20 19:16:39 +02:00

FmaMixOnce.cpp

…

FormattedBufferByteStores.cpp

…

FormattedBufferOnce.cpp

perf(shader): share folded SPIR-V constants and drop identity bitcasts

2026-10-18 21:07:19 +02:00

FragmentBarycentric.cpp

…

FrameDumpIo.cpp

test(agc): write more pixels than a pipe holds in the closed-pipe case

2026-10-15 14:04:10 -07:00

GcnDecode.cpp

fix(shader): keep the literal word of a GCN s_setreg_imm32_b32

2026-10-22 12:17:44 +03:00

Graphics.cpp

Merge pull request #7647 from fma-rc/fix/agc-heap-mirror-resident

2026-10-23 12:45:26 +02:00

GraphicsTests.hpp

fix(agc): keep detiler descriptor pools unrecycled without a recorder

2026-10-15 10:14:29 +02:00

GuardedMemoryLoads.cpp

perf(recompiler): keep LDS loads branched and drop the LDS sentinel

2026-10-17 12:19:09 +13:00

GuestAllocations.cpp

perf(agc): cache the main image's page states on Linux

2026-10-19 19:26:32 +02:00

GuestMemoryLifetimeFixture.cpp

fix(agc): harden host import construction and mutation waits

2026-10-18 14:19:49 +13:00

GuestMemoryLifetimeSupport.hpp

fix(agc): harden host import construction and mutation waits

2026-10-18 14:19:49 +13:00

GuestSamplerResource.cpp

fix(agc): accept a min or max reduction sampler with anisotropic filtering

2026-10-15 15:16:41 +02:00

GuestTextureResource.cpp

Merge pull request #8962 from Celegans12/fix/agc-r128-texture-max-mip

2026-10-23 12:10:17 +13:00

HandoffSpin.cpp

perf(agc): count the processors the process may run on for the hand-off spin guard

2026-10-16 19:52:11 +02:00

HostAllocator.cpp

perf(agc): cache freed small blocks per thread on Windows

2026-10-17 01:47:24 +07:00

HostHdr.cpp

feat(videoout): allow opting out of HDR presentation

2026-10-18 05:43:09 -03:00

HostImportFailure.cpp

fix(agc): publish host import handles only after creation succeeds

2026-10-17 12:14:15 +02:00

HostImportFailure.hpp

fix(agc): publish host import handles only after creation succeeds

2026-10-17 12:14:15 +02:00

HostImportLifetime.cpp

fix(agc): retire host imports before range cuts

2026-10-21 16:42:05 +00:00

HostImportLifetime.hpp

fix(agc): harden host import construction and mutation waits

2026-10-18 14:19:49 +13:00

HostImportRetireHold.cpp

fix(agc): publish host import handles only after creation succeeds

2026-10-17 12:14:15 +02:00

HostImportSparseDmaBuf.cpp

Merge pull request #6442 from JJenkx/fix/retain-import-window-failures

2026-10-19 04:33:22 +02:00

IntegerFolding.cpp

perf(shader): fold integer compares by value sets and known bits

2026-10-18 21:06:52 +02:00

IntegerNarrowing.cpp

…

IrReclaim.cpp

perf(shader): keep only evaluable branch conditions in resource plans

2026-10-15 08:22:09 +13:00

KeyMiss.cpp

…

LayerInput.cpp

test(agc): cover a Layer input without geometry shaders

2026-10-14 17:45:15 -06:00

LdsLockBudget.cpp

fix(shader): fault on misaligned 64-bit device-memory LDS atomics

2026-10-18 12:31:11 +03:00

LiveStack.cpp

perf(agc): answer small guest read checks without /proc/self/maps

2026-10-18 16:45:07 +00:00

Locks.cpp

perf(agc): replace winpthreads locks and TLS on the draw path

2026-10-23 11:32:03 +03:00

MaskedSelect.cpp

fix(recompiler): replay every eight-bit stencil reference a pixel shader may export

2026-10-19 21:57:24 +03:00

Mesh.cpp

feat(agc): apply register and exported sizes to mesh points

2026-10-21 09:13:32 -04:00

MeshGeometryFixture.hpp

fix(agc): validate mesh index descriptors on cached draws (reapply #5791)

2026-10-15 22:43:15 +02:00

MeshOutputBudget.cpp

feat(shader): emulate per-vertex parameters on the mesh path

2026-10-19 17:26:37 +02:00

MubufLdsDecode.cpp

…

MultisampleTextureLayout.cpp

feat(agc): add fragment-preserving R_X and Z_X layouts

2026-10-21 16:06:32 +03:00

NarrowConstantStoreFixture.hpp

…

NopDataWritePacket.cpp

feat(libSceAgcDriver): build NOP data-write packets

2026-10-22 11:42:46 -07:00

PassHazards.cpp

fix(agc): seal a pass's write-only stores at guest flushes

2026-10-19 00:48:42 +03:00

PerformanceOverlay.cpp

Merge pull request #7628 from milescoviello/test/hex-failure-values

2026-10-20 01:37:32 +02:00

PerformanceOverlayStatistics.cpp

test(agc): name the overlay statistics test's tolerance check within

2026-10-15 08:03:49 +00:00

PerformanceTrace.cpp

test(trace): check again that a failed benchmark rewrite keeps the previous file

2026-10-20 02:47:24 -05:00

PipelineLinkOptimization.cpp

…

PipelineQueueFailure.cpp

fix(agc): count pipeline jobs after queue admission

2026-10-17 17:02:32 +02:00

PixelPositionExport.cpp

…

PlaceholderImages.cpp

fix(agc): omit unavailable storage MSAA placeholders

2026-10-18 16:02:26 +00:00

Pm4.cpp

Merge pull request #8643 from realshinmalphur/fix/gpu-command-validation

2026-10-23 21:15:35 +13:00

Pm4Replay.cpp

…

Pm4ReplayTool.cmake

fix(agc): take agc_pm4_replay's capture directory as a wide argument on Windows

2026-10-20 10:03:21 -03:00

PositionInvariance.cpp

fix(shader): leave out the all-zero position distance when guest distances fill all eight

2026-10-18 10:40:52 +02:00

PresentationPass.cpp

…

ProfileOutput.cpp

fix(agc): reject late profiling output after shutdown

2026-10-19 02:05:16 -04:00

RdnaTruncatedModifier.cpp

…

Recorder.cpp

fix(agc): publish the units a completion store covers partly before it lands

2026-10-23 09:33:29 +02:00

RecorderKeptBytes.cpp

fix(agc): discard recorder batches after submission failure

2026-10-15 06:06:00 +00:00

RectListPositionInvariant.cpp

test(shader): pin the rect-list vertex selected at every tessellation corner

2026-10-17 02:13:25 +02:00

RedundantValueElimination.cpp

perf(shader): keep redundant expression operand keys inline

2026-10-19 13:24:47 +02:00

RegisterKey.cpp

fix(agc): mix shader identity into trace key

2026-10-15 13:06:28 +02:00

ResidentPresent.cpp

perf(agc): present 8-bit integer resident images without a readback

2026-10-14 22:04:55 +03:00

ResidentPresent.hpp

…

SavedExecMaskLoop.hpp

fix(shader): reuse completed pixel lane proofs

2026-10-22 21:11:46 +02:00

ScalarSelectMask.cpp

…

Screenshot.cpp

fix(agc): read ANYPS5_SCREENSHOTS through the wide environment on Windows

2026-10-20 08:49:47 -03:00

ShaderDeviceProfile.cpp

fix(recompiler): hold 32 samplers in the runtime sampler heap

2026-10-17 16:36:17 +13:00

ShaderDiskCache.cpp

Merge pull request #7945 from veosett/codex/windows-cache-housekeeping-safety

2026-10-21 20:57:27 +13:00

ShaderMemory.cpp

Merge pull request #6801 from JJenkx/feat/geometry-points-sized

2026-10-23 14:38:10 +13:00

ShaderModeRegisters.cpp

…

ShaderReplayRequests.cpp

feat(agc): replay prepared shader capture and materialization offline

2026-10-16 20:34:23 +02:00

ShaderReplayTool.cmake

feat(agc): replay prepared shader capture and materialization offline

2026-10-16 20:34:23 +02:00

SharedWrites.cpp

…

SlabReturn.hpp

fix(agc): return slab slots without allocating host memory

2026-10-17 14:05:58 +02:00

SnapshotCopyFailure.cpp

fix(agc): roll back failed snapshot cache insertions

2026-10-19 17:47:44 +04:00

Spirv13Image.cpp

…

SpirvBranchTargets.cpp

…

SpirvFailureText.cpp

fix(shader): print the shader hash of a SPIR-V emission failure in hex

2026-10-18 16:56:23 +00:00

SpirvLdsAtomicLock.cpp

fix(shader): release the LDS lock with an exchange and drop fences inside it

2026-10-19 14:43:13 +02:00

SpirvModule.cpp

fix(shader): ballot each predicate once per block

2026-10-19 09:30:26 +02:00

SpirvNarrowConstantStore.cpp

…

SpirvOptimizer.cpp

fix(shader): ballot each predicate once per block

2026-10-19 09:30:26 +02:00

SpirvSubvectorLoopBegin.cpp

test(shader): check that s_subvector_loop_begin tests an empty exec mask with one compare

2026-10-20 14:59:55 +07:00

StackWalk.cpp

…

Structurizer.cpp

fix(shader): give a selection a merge when one arm returns through nested branches

2026-10-15 10:11:54 -04:00

SubgroupBarrier.cpp

fix(shader): order wave LDS accesses with a memory barrier on AMD's proprietary driver

2026-10-20 20:24:01 +05:00

SubgroupClock.cpp

…

Submit.cpp

Merge main into pr-5480

2026-10-19 22:25:30 +13:00

SubvectorLoopFixture.hpp

test(shader): check that s_subvector_loop_begin tests an empty exec mask with one compare

2026-10-20 14:59:55 +07:00

TessellationFactorRing.cpp

…

TessellationLowering.cpp

fix(shader): clamp a domain read at an immediate address to the last control point

2026-10-22 10:39:26 +07:00

test_guest_memory_lifetime.py

fix(agc): harden host import construction and mutation waits

2026-10-18 14:19:49 +13:00

test_shader_cache_housekeeping.py

test(shader): cover device paths within host file IO limits

2026-10-19 21:25:57 +06:00

test_shader_cache_prebuilt_discovery.py

test(shader): isolate prebuilt cache directory discovery

2026-10-20 23:11:57 +06:00

TextureDetiler.cpp

fix(agc): keep detiler descriptor pools unrecycled without a recorder

2026-10-15 10:14:29 +02:00

TextureFormat.cpp

feat(shader): decode PS4 (GCN) shaders by re-encoding them to RDNA

2026-10-22 12:17:44 +03:00

TextureTiling.cpp

test(agc): cover range-limited depth plane stores

2026-10-21 10:03:00 +00:00

ThreadPriority.cpp

…

UnsupportedDiagnostics.cpp

feat(libSceAgcDriver): implement diagnostic fallbacks

2026-10-22 11:42:42 -07:00

ValidateProgram.cpp

…

VectorWindow.cpp

perf(shader): drop window stores no relative read can see

2026-10-18 21:55:41 +02:00

VertexBarrier.cpp

fix(shader): emit nothing for a vertex s_barrier that precedes every memory access

2026-10-19 01:07:58 +00:00

VertexInput.cpp

fix(agc): follow RDNA 2's range checks for zero-stride vertex fetches

2026-10-21 20:16:30 +02:00

VertexInputValidation.cpp

feat(agc): draw point-list output of NGG geometry programs as mesh points

2026-10-20 22:26:30 -05:00

VertexLayerExport.cpp

…

Visual.cpp

…

Vop1ControlDecode.cpp

…

Vop1SdwaMoveDecode.cpp

…

Vop2EncodingDecode.cpp

…

Vop3InterpDecode.cpp

…

WaitShutdown.cpp

feat(libSceAgcDriver): implement command buffer submission

2026-10-19 11:28:13 +05:30

WaveControlDecode.cpp

feat(shader): model ds_ordered_count swap

2026-10-19 16:41:03 +02:00

WriteTracking.cpp

Merge pull request #7233 from MarcZX10/codex/bda-store-exact-coverage-7220

2026-10-23 01:53:33 +13:00

WriteWatchCoverage.cpp

fix(agc): restore write-watch coverage for remapped pages

2026-10-20 12:41:38 +04:00