Adrià f6eeca4f5f Merge pull request #6786 from fdfawef324/claude/agc-indirect-args-stale ...

fix(agc): wait for address-based writers before reading indirect dispatch arguments

2026-10-19 13:16:24 +02:00

..

AddressPasses.cpp

test(agc): check that a read through the BDA table sees a target as with its pass broken

2026-10-17 11:42:18 +00:00

AliasCopyLabel.cpp

fix(agc): store an image-copied buffer copy's destination before later labels

2026-10-15 13:41:36 +02:00

AlphaToCoverage.cpp

test(agc): distinguish stores beside resident DCC metadata

2026-10-17 15:04:14 +03:00

ArrayLayerTarget.cpp

fix(agc): preserve registered array slices on CPU write-back

2026-10-15 17:08:14 +04:00

BcBlockCopy.cpp

…

BdaAnchorReads.cpp

fix(shader): keep BDA anchors exact on tables whose ranges overlap

2026-10-14 10:09:08 -04:00

BdaSpanReads.cpp

…

BindlessImages.cpp

feat(agc): implement image and sampler tables indexed at runtime offsets

2026-10-14 16:43:57 +02:00

BitfieldMask.cpp

…

BranchPastEndpgm.cpp

…

BufferAtomicCsub.cpp

…

BufferAtomics.cpp

…

BufferAtomicsFloat.cpp

…

BufferAtomicsIncDec64.cpp

…

BufferAtomicsOrSwap64.cpp

…

BufferAtomicUnaligned.cpp

fix(shader): act on the containing dword for unaligned 32-bit buffer atomics

2026-10-18 17:54:44 +02:00

BufferDescriptorRuntime.cpp

fix(libSceAgcDriver): normalize reserved bit 30 in guest buffer descriptors

2026-10-18 16:38:59 +05:30

BufferFormat.cpp

…

BufferIntegerAtomics.cpp

…

BufferLoadStore.cpp

…

BufferResourceVcc.cpp

style(shader): trim descriptor fixture trailing blank lines

2026-10-14 12:26:01 +03:00

BufferUnalignedBase.cpp

test(agc): expect 32-bit buffer atomics off a DWORD boundary to act on their containing dword

2026-10-18 17:54:44 +02:00

Bvh64IntersectRay.cpp

…

CacheControl.cpp

…

ClampedGlobalFill.cpp

fix(agc): take RADV's fill-memory kernel as a buffer fill

2026-10-19 08:21:36 +02:00

Color565Target.cpp

feat(agc): render 5_6_5 UNORM color targets

2026-10-17 23:59:26 +07:00

Color1555Target.cpp

…

ComputeFsTableCalls.cpp

feat(agc): link calls through the AGC FS table

2026-10-17 16:47:09 +02:00

ComputePipelineOutsideGpuMutex.cpp

test(agc): give the new program of compute_pipeline_outside_gpu_mutex its own module

2026-10-16 12:52:28 +02:00

ComputeTableCalls.cpp

fix(shader): fault only where a table call misses its linked callees

2026-10-15 06:44:25 +13:00

ComputeUserDataCalls.cpp

…

ConservativeRasterization.cpp

…

ConstantDwordFill.cpp

perf(agc): take a dword fill of any inline constant as a buffer fill

2026-10-14 04:12:35 +02:00

ConstantStore.cpp

…

CopiedWriterCopy.cpp

…

CrossWorkgroupFlag.cpp

…

D16Memory.cpp

…

DccKeyPresync.cpp

perf(agc): pre-sync the DCC key writers of a dispatch's images before the GPU mutex

2026-10-16 14:18:46 +02:00

DccKeysFirstWrite.cpp

perf(agc): settle Windows imports under 1 MiB without waiting

2026-10-15 19:42:17 -07:00

DepthBiasPerFace.cpp

perf(agc): expand draw recipes to support depth targets

2026-10-14 02:00:40 +02:00

DepthBitsGuestMemory.cpp

test(agc): read the depth bits of a SW_64KB_Z_X surface from guest memory

2026-10-17 20:19:49 -07:00

DepthClearHle.cpp

test(agc): isolate the depth clear cases per recorder and cover plane fills against a UAV clear

2026-10-18 05:56:26 +02:00

DepthCompareCopy.cpp

revert(libSceAgcDriver): restore deferred writeback and main execution tests

2026-10-14 20:57:58 +01:00

DepthFastClear.cpp

fix(agc): read and store back the whole pipe-aligned HTILE outside the surfaces lock

2026-10-15 13:53:50 +02:00

DepthFastClearHarness.hpp

fix(agc): preserve fresh HTILE depth and stencil clears

2026-10-17 18:54:23 -04:00

DepthFastClearSubmit.cpp

test(agc): cover the depth and stencil fast clear a recorded HTILE fill leaves

2026-10-19 05:16:01 +02:00

DepthHtileWrite.cpp

test(agc): allocate full tiled depth fixture footprints

2026-10-17 18:54:24 -04:00

DepthPlaneFill.cpp

fix(agc): inherit pending plane fills on fresh attachments

2026-10-17 18:54:24 -04:00

DepthPlaneGuestLayers.cpp

test(agc): drop the refusal cases main's relaxed geometry check no longer produces

2026-10-14 20:01:37 +02:00

DepthPlaneSampling.cpp

test(agc): drop the comment above the guest depth-bits case

2026-10-18 09:55:38 +07:00

DepthRangeRemapDraw.cpp

test(agc): draw a fetchless mode 0 vertex shader on a remapping target

2026-10-19 00:59:33 +13:00

DepthStorage.cpp

test(agc): create fixture command pools with the reset flag

2026-10-18 01:00:16 +00:00

DepthStorageDispatch.cpp

feat(agc): storage image access to depth planes through a 2D array view from a later slice

2026-10-14 12:13:52 +02:00

DepthStorageReuse.cpp

fix(agc): retire a depth surface on a storage write only when the write covers a whole plane

2026-10-18 10:18:12 +02:00

DepthSurfaceFromStorage.cpp

test(agc): depth_surface_from_storage pins new depth surfaces taken from storage writes

2026-10-14 20:23:39 +02:00

DepthSurfaceLastWriter.cpp

fix(libSceAgcDriver): retire a depth surface once another writer covers its memory

2026-10-15 11:00:36 +02:00

DeviceSelection.cpp

…

DispatchMove.cpp

fix(agc): keep the template of another address when a dispatch move is refused

2026-10-17 13:52:16 +02:00

DispatchMoveDepth.cpp

perf(agc): serve compute dispatches whose read-only buffers moved from the cached template

2026-10-17 09:08:47 +02:00

DivFixupF16.cpp

…

DivFixupF16DenormalModes.cpp

fix(shader): flush v_div_fixup_f16 src1 and src2 by the f16 denormal field

2026-10-17 13:21:18 +07:00

DivFixupF64DenormalModes.cpp

fix(shader): flush v_div_fixup_f64 src1 and src2 by the f64 denormal field

2026-10-17 14:08:08 +07:00

Division.cpp

…

DivisionOutputModifiers.cpp

…

DivisionResultModifiers.cpp

…

Dpp8.cpp

…

Dpp8PkFmac.cpp

…

Dpp16InactiveSource.cpp

…

Dpp16Operands.cpp

…

Dpp16RowShareXmask.cpp

…

DppNoForm.cpp

…

DrawCacheRelocation.cpp

Merge main into perf/shader-image-table-reuse-v3-pr

2026-10-17 16:05:44 -07:00

DrawCacheReuse.cpp

fix(agc): stop the pipeline library worker before the device is reset

2026-10-15 18:02:44 -07:00

DrawOffsetIndices.cpp

perf(agc): draw an indirect rect list from its records on the GPU

2026-10-15 08:17:50 +02:00

DrawPipelineTrace.cpp

fix(agc): include draw committer work in CPU traces

2026-10-19 02:27:57 +02:00

DrawUnsupportedBufferType.cpp

test(agc): check that a draw refused for a buffer type names its shader and V#

2026-10-15 01:22:23 +03:00

DrawWriteRanges.cpp

test(agc): run the failure and draw cache reuse tests with pipelined draws, pin CMASK write ranges

2026-10-15 11:34:29 +02:00

DsAppendConsume.cpp

…

DsPermute.cpp

…

DsSwizzle.cpp

…

DsWrxchg2.cpp

…

DsWrxchg2B64.cpp

…

DualSourceBlend.cpp

…

DwordPatternFill.cpp

perf(agc): match the dword pattern-fill kernel with its two compares in either order

2026-10-15 14:40:16 -03:00

Dx10Clamp.cpp

…

EmulatedColorCompare.cpp

test(agc): skip the emulated color compare clamp cases without MinLod

2026-10-17 12:26:38 +03:00

ExecHostLanes.cpp

…

F16ConversionFloatMode.cpp

fix(shader): round and flush f16 conversion results by FLOAT_MODE in every denormal mode

2026-10-14 16:25:23 +00:00

F16DenormalFlush.cpp

…

F16FmaDenormalModes.cpp

test(agc): accept any NaN where the f16 fma and fma_mix tests expect one

2026-10-17 18:13:41 +07:00

F16MinMaxNan.cpp

…

F16PackDenormalModes.cpp

fix(shader): flush v_cvt_pkrtz_f16_f32 and v_pack_b32_f16 f16 denormals

2026-10-17 13:24:16 +07:00

F16PackTowardZero.cpp

…

F16ToF32Unpack.cpp

perf(shader): skip the f16 NaN rebuild for operands that are only compared

2026-10-13 15:11:54 +02:00

F32ArithmeticNan.cpp

…

F32CompareConstant.cpp

…

F32ConstantFactorTininess.cpp

…

F32DenormalFlush.cpp

perf(shader): flush f32 denormals in hardware when a startup probe matches the emulation

2026-10-19 07:00:13 +02:00

F32Dot2Rounding.cpp

…

F32MinMaxNan.cpp

…

F32OutputModifier.cpp

…

F32PackRounding.cpp

…

F32ResultClamp.cpp

…

F64Arithmetic.cpp

test(agc): respect f64 denormal preservation support

2026-10-18 02:35:26 +02:00

F64Conversions.cpp

…

F64Division.cpp

…

F64DivisionModifiers.cpp

…

F64FmaSubnormals.cpp

perf(shader): map cached SPIR-V from the shader cache files

2026-10-14 10:36:24 +04:00

F64NanConversions.cpp

…

F64Rounding.cpp

…

F64Transcendental.cpp

…

FillCoverKeys.cpp

…

FlatApertures.cpp

…

FlatApertureSources.cpp

feat(recompiler): read the scalar aperture sources as values

2026-10-13 22:39:49 -04:00

FlatCacheBits.cpp

…

FlatD16Loads.cpp

…

FlatGlobalLoadsRegion.hpp

…

FlatLoads.cpp

…

FlatLoadsX2X3.cpp

…

FlatLoadsX4.cpp

…

FlatStore.cpp

fix(agc): run the AGC GPU tests in an ANYPS5_ENABLE_ASAN build

2026-10-17 14:49:28 +00:00

FlipStoresSmallResults.cpp

test(agc): give the flip store test the deadline its siblings use

2026-10-19 09:54:08 +03:00

Float16Dot.cpp

…

Float16FmaRounding.cpp

…

Float16MinMaxBits.cpp

feat(shader): translate packed f16 in the f16 denormal modes that flush

2026-10-17 21:23:10 +13:00

Float16Misc.cpp

…

Float16Modifiers.cpp

…

Float16Native.cpp

test(agc): word the native f16 skip as a device limitation

2026-10-13 17:56:02 -04:00

Float16ResultModifiers.cpp

…

Float16Rounding.cpp

…

Float16TernaryClamp.cpp

…

FmaLegacy.cpp

…

FmaMixF16DenormalModes.cpp

test(agc): accept any NaN where the f16 fma and fma_mix tests expect one

2026-10-17 18:13:41 +07:00

FmaRounding.cpp

…

FmaskImageReads.cpp

fix(agc): read uncompressed FMASK images from guest memory

2026-10-16 09:40:41 +06:00

FormatD16.cpp

…

FormatD16Store.cpp

…

FormattedStore.cpp

…

FormattedStoreUnaligned.cpp

perf(shader): store 8- and 16-bit buffer components as bytes in compute

2026-10-13 16:36:56 +02:00

Fp16Overflow.cpp

…

FragmentUserDataTail.cpp

feat(shader): link raw fragment user-data tails with a constant high half

2026-10-13 21:34:59 +02:00

FrontFaceAllBits.cpp

fix(shader): honour SPI_BARYC_CNTL.FRONT_FACE_ALL_BITS for the front-face input

2026-10-14 21:37:48 +02:00

GatedDispatchCapture.cpp

…

GdsAppendConsume.cpp

fix(shader): add the ordered count of a first active lane in the upper wave64 half

2026-10-15 07:37:40 +00:00

GlobalAtomicCsub.cpp

…

GlobalAtomics64.cpp

…

GlobalAtomics.cpp

…

GlobalAtomicsFloat.cpp

…

GlobalAtomicsLanes.cpp

…

GlobalLoadAddtid.cpp

…

GlobalLoads.cpp

perf(agc): create address-table compute pipelines before the GPU lock from the described set layout

2026-10-15 07:27:37 +13:00

GlobalLoadsX2X3.cpp

…

GlobalLoadsX4.cpp

…

GlobalStoreAddtid.cpp

…

GlobalVccBase.cpp

…

GpuBufferFormat.cpp

test(shader): pin GPU-selected formatted loads for every buffer format

2026-10-15 10:32:47 +13:00

GpuIndirectArgumentsMisalignedWriterAhead.cpp

test(agc): align indirect argument storage at runtime

2026-10-15 18:54:28 +03:00

GpuIndirectArgumentsWrittenAhead.cpp

fix(agc): resolve a GPU-read indirect dispatch at the gate when the run writes its arguments

2026-10-15 06:50:57 -06:00

GpuMapOnFiber.cpp

fix(agc): run the AGC GPU tests in an ANYPS5_ENABLE_ASAN build

2026-10-17 14:49:28 +00:00

GuardedNullPointer.cpp

…

GuestMemoryLifetime.cpp

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

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

HalfExpansion.cpp

…

HostImportAtMap.cpp

…

HostImportTeardown.cpp

…

HostImportWindows.cpp

test(agc): host import windows, whole-range builds and budget refusals

2026-10-14 14:51:48 +13:00

IdleStoreOverUnmap.cpp

fix(agc): hold a guest allocation lease while a queue stores idle results

2026-10-15 09:01:16 +02:00

ImageAddressDimension.cpp

fix(agc): accept 2D-array image loads and stores on 3D T#s

2026-10-13 18:44:22 -05:00

ImageAtomics64.cpp

…

ImageAtomics.cpp

…

ImageAtomicsBitwise.cpp

…

ImageAtomicsFloat.cpp

…

ImageBy.cpp

…

ImageCacheBits.cpp

…

ImageConvertedUnorm.cpp

feat(shader): sample converted unorm images in the shader

2026-10-18 17:43:35 +02:00

ImageGatherOneDimensional.cpp

…

ImageGatherOneDimensionalArray.cpp

…

ImageLoadPacked.cpp

…

ImageOneDimensionalArray.cpp

…

ImagePacked16Load.cpp

…

ImagePackedHorizontalGather.cpp

…

ImagePckN.cpp

…

ImagePipeBankXor.cpp

…

ImagePlaneAddressOnCube.cpp

fix(shader): read cube T#s through DIM 2D image_sample and image_load as face 0

2026-10-14 23:37:33 +02:00

ImagePlaneAddressOnLine.cpp

fix(shader): read explicit-LOD and level-zero 2D-array samples through 1D T#s as measured

2026-10-15 15:36:02 +02:00

ImagePlaneAddressOnLineImplicitLod.cpp

…

ImagePlaneBiasOnSingleLevelVolume.cpp

fix(shader): decode forced degamma on biased RGBA8_UINT samples

2026-10-14 23:25:53 +02:00

ImagePlaneCompareOnLine.cpp

fix(shader): compare through 2D image_sample_c on 1D T#s

2026-10-14 21:54:00 +02:00

ImageRepackedFloat.cpp

feat(agc): sample texture format 43 through a lossless B10G11R11 copy

2026-10-17 16:53:37 -04:00

ImageSampleAnisoOverride.cpp

…

ImageSampleBiasA16.cpp

…

ImageSampleBorderColor.cpp

…

ImageSampleCubeDerivatives.cpp

fix(shader): keep cube image gradients out of the face coordinate bias

2026-10-13 15:26:04 +02:00

ImageSampleDerivatives.cpp

…

ImageSampleForceDegamma.cpp

…

ImageSampleLodClamp.cpp

…

ImageSampleReduction.cpp

fix(test): sample the anisotropic min/max rows without a mip filter

2026-10-18 07:38:34 +05:30

ImageSampleSharedSampler.cpp

…

ImageSampleUnnormalized.cpp

feat(libSceAgcDriver): support secondary LOD bias and relax unnormalized coordinate filters

2026-10-14 17:07:46 +05:30

ImageSampleVccSampler.cpp

test(agc): exercise VCC-backed sampler words on Vulkan

2026-10-16 15:59:37 +03:00

ImageSint8.cpp

…

ImageSrgbLoad.cpp

perf(shader): initialize the sRGB decode table once per module

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

ImageSrgbSampleCopy.cpp

test(agc): force the sRGB sample copy with APS5_SRGB_SHADER_DECODE=copy

2026-10-18 07:27:15 -04:00

ImageStoreArrayThroughVolume.cpp

fix(agc): accept 2D-array image loads and stores on 3D T#s

2026-10-13 18:44:22 -05:00

ImageStoreLineThroughPlane.cpp

…

ImageStorePacked.cpp

…

ImageStoreSint.cpp

feat(shader): store D16 data to SINT storage images

2026-10-16 13:18:38 +02:00

ImageTable.cpp

Merge pull request #6417 from UnhealthyKraken/fix/invariant-image-table-phis

2026-10-19 02:40:40 +02:00

ImageTableCache.cpp

feat(agc): implement image and sampler tables indexed at runtime offsets

2026-10-14 16:43:57 +02:00

ImageTableCacheBounds.cpp

test(agc): create the driver before the queued-draw vouch takes its stamp

2026-10-18 17:03:09 -07:00

ImageTableReuse.cpp

test(agc): expect one resolution for a CPU write beside an image table on Windows too

2026-10-18 22:59:54 -04:00

ImageTexelStatus.cpp

…

ImageVolumeAddressOnLine.cpp

fix(shader): read 1D T#s through 3D image_sample and image_load accesses

2026-10-14 18:32:22 +01:00

ImageVolumeAddressOnPlane.cpp

…

ImageVolumeAddressPlanarTexture.cpp

…

ImageXorSwizzleVolume.cpp

feat(agc): support thick 4 KiB standard XOR volumes

2026-10-18 01:22:36 +03:00

InactiveVertexPointSize.cpp

…

IndirectArgumentsPendingProducer.cpp

test(agc): keep direct argument case off the stopped driver

2026-10-19 10:18:11 +02:00

IndirectArgumentsWrittenAhead.cpp

…

IndirectDispatchLimit.cpp

test(agc): exceed the device's own indirect dispatch limit

2026-10-19 18:57:39 +13:00

IndirectPatchedStage.cpp

perf(agc): draw stages use their shared compiled result instead of a copy

2026-10-14 12:11:54 +05:30

IntegerDot.cpp

…

IntegerToF32Rounding.cpp

…

InterpolationF16.cpp

test(agc): draw 16-bit interpolation through the emulated barycentrics

2026-10-17 15:18:41 +03:00

InterpolationModes.cpp

test(agc): compare 16-bit interpolation with a CPU reference on random rows

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

JobManagerCompute.cpp

fix(libSceJobManager): refuse shutdown with compute work in flight and check the handler wait

2026-10-18 02:14:02 +02:00

LayeredVolumeDraw.cpp

feat(agc): support thick 4 KiB standard XOR volumes

2026-10-18 01:22:36 +03:00

LayeredVolumeSubmit.cpp

…

LdexpF16InlineConstants.cpp

…

LdexpF32Denormal.cpp

…

LdexpF32DenormalModes.cpp

fix(shader): translate v_ldexp_f32 in f32 denormal modes 1 and 2

2026-10-14 21:39:55 +03:00

LdsAddF32.cpp

fix(shader): flush ds_add_f32 denormals and NaNs as the hardware does

2026-10-14 22:48:34 +02:00

LdsAddtid.cpp

…

LdsAtomics64.cpp

…

LdsAtomics.cpp

…

LdsBeyondDeviceLimit.cpp

fix(shader): use 64-bit buffer atomics for device-memory LDS

2026-10-18 05:47:30 +03:00

LdsCondxchg.cpp

…

LdsFloatAtomicModes.cpp

…

LdsIntegerAtomics.cpp

…

LdsMaskWaterfall.cpp

test(agc): run the LDS mask waterfall in both wave modes

2026-10-16 14:12:58 +02:00

LdsMixedWidthAtomics.cpp

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

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

LdsReadWrite.cpp

…

LdsSrc2.cpp

…

LinearTargetResident.cpp

perf(agc): linear color targets over imported memory get a resident image

2026-10-16 20:42:45 +02:00

LogNearOne.cpp

…

LookBackScan.cpp

…

LoopEndingExits.cpp

…

MacLegacy.cpp

…

ManyBuffers.cpp

…

MaskedGlobalStore.cpp

…

MemoryValueFill.cpp

perf(agc): take the memory-value dword fill kernel as a buffer fill

2026-10-14 00:44:41 +02:00

MeshDrawRelocation.cpp

fix(agc): keep the mesh index descriptor on relocated draw stages

2026-10-17 13:36:47 +02:00

MeshIndexCache.cpp

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

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

MeshProvokingLayer.cpp

fix(shader): write a mesh shader's layer through a per-primitive block

2026-10-19 03:07:32 +07:00

MeshProvokingVertex.cpp

fix(agc): draw geometry programs with the last provoking vertex

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

MixPrecision.cpp

test(agc): run the mix precision test without Float64

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

MsaaEnableSingleSample.cpp

…

MulLoU16HighHalf.cpp

fix(shader): read low-half 16-bit multiply factors as 24-bit fields

2026-10-17 12:52:58 +07:00

MultisampledPassOrder.cpp

fix(agc): name in-place draw inputs and multisampled targets in pass hazards

2026-10-18 23:14:36 +02:00

MultisampleFillHostOnly.cpp

fix(agc): keep multisample ownership on its recording device

2026-10-19 01:08:38 +02:00

MultisampleReadback.cpp

test(agc): check the sample reader excludes cache snapshots

2026-10-18 04:25:27 +03:00

MultisampleRetirement.cpp

fix(agc): reclaim retired multisample color targets

2026-10-19 01:08:37 +02:00

NanQuieting.cpp

…

NativeDepthBdaClear.cpp

test(agc): classify native clear import skips

2026-10-17 21:25:34 +03:00

NativeMultisampleDepth.cpp

fix(agc): keep the first native multisample depth sampling refusal

2026-10-19 02:45:32 +02:00

NggVertexWaveInfo.cpp

…

NggVertexWaveReduction.cpp

…

OversizedDescriptorSets.cpp

test(agc): create fixture command pools with the reset flag

2026-10-18 01:00:16 +00:00

OversizedDispatch.cpp

fix(agc): split oversized dispatches with workgroup offset specialization

2026-10-18 22:54:33 -05:00

PackedAlu.cpp

feat(shader): translate packed f16 in the f16 denormal modes that flush

2026-10-17 21:23:10 +13:00

PackedF16DenormalModes.cpp

feat(shader): translate packed f16 in the f16 denormal modes that flush

2026-10-17 21:23:10 +13:00

PackedFloat16Native.cpp

…

PackedFloatColorTarget.cpp

feat(agc): implement 11_11_10 float color targets through 32-bit uint attachments

2026-10-17 17:29:21 -04:00

PackedIntegerClamp.cpp

…

PackedIntegerNegate.cpp

…

PackedMipTailTarget.cpp

…

PackedUnormColorTarget.cpp

…

PartiallyResidentSnapshot.cpp

test(agc): create fixture command pools with the reset flag

2026-10-18 01:00:16 +00:00

PassInputHazard.cpp

fix(agc): name in-place draw inputs and multisampled targets in pass hazards

2026-10-18 23:14:36 +02:00

PatternBufferFill.cpp

…

PendingExactSurface.cpp

revert(libSceAgcDriver): restore deferred writeback and main execution tests

2026-10-14 20:57:58 +01:00

Permlane16.cpp

…

PipelinedDispatch.cpp

test(agc): wait for the pipeline to empty between descriptor rounds

2026-10-16 12:53:38 +02:00

PipelinedDma.cpp

fix(agc): keep pipelined DMA opt-in and order copies against queued writes

2026-10-18 00:14:02 +02:00

PipelinedFill.cpp

fix(agc): a pipelined fill reads its inputs after queued writes and never waits for room under the GPU lock

2026-10-15 11:34:28 +02:00

PipelinedPredicatedLabelOrder.cpp

fix(agc): a label packet behind a deferred label is not enqueued ahead of it

2026-10-15 11:34:28 +02:00

PipelinedPredicationRead.cpp

fix(agc): a predicated packet drains queued writes to its predicate before reading it

2026-10-15 11:34:28 +02:00

PipelinedPrefetch.cpp

perf(agc): a DMA_DATA prefetch keeps the draw pipeline

2026-10-19 07:43:15 +02:00

PipelineJournal.cpp

perf(agc): journal the pipelines of a run and build them ahead on the next one

2026-10-16 12:52:28 +02:00

PipelineOptimizedReadback.cpp

fix(agc): stop and join the compare helpers and the pipeline library worker at shutdown

2026-10-15 15:09:33 -07:00

Pixel32ArExport.cpp

fix(shader): take the 32_AR export alpha from channel 1 as GFX10 does

2026-10-14 17:56:48 +13:00

PixelAncillaryWord.cpp

…

PixelBarycentricEmulation.cpp

test(agc): cover the tessellation route of the barycentric emulation

2026-10-18 02:16:51 +04:00

PixelFixedPointPosition.cpp

…

PixelHelperLanes.cpp

…

PixelImplicitDerivatives.cpp

fix(shader): preserve pixel helper lanes in EXEC copies

2026-10-15 14:42:26 +05:30

PixelInterlock.cpp

…

PixelLds.cpp

…

PixelLdsLaneSlots.cpp

…

PixelLdsReadModifyWrite.cpp

test(agc): keep the pixel LDS read-modify-write slots within 1 KiB

2026-10-14 22:37:42 +13:00

PixelRowReduction.cpp

fix(shader): reduce a DPP row scan read at a row's last lane over the active lanes

2026-10-17 22:05:35 -03:00

PixelSampleCoverage.cpp

feat(agc): draw radeonsi line and polygon smoothing through merged mixed samples

2026-10-19 01:25:33 +00:00

PixelSampleIteration.cpp

test(agc): cover pixel shaders run once per pixel and once per sample

2026-10-15 16:06:38 +00:00

PixelUint16Export.cpp

…

PixelValidMask.cpp

…

PixelWaveReduction.cpp

…

PknormF16Denormals.cpp

fix(shader): flush v_cvt_pknorm_*_f16 denormal inputs by FLOAT_MODE

2026-10-14 16:18:32 +00:00

PooledBufferRange.cpp

test(agc): create fixture command pools with the reset flag

2026-10-18 01:00:16 +00:00

PreparationScopeSettle.cpp

test(agc): wait for preparation workers before linking

2026-10-17 02:00:23 +00:00

PreparedExports.cpp

…

PreparedGraphics.cpp

fix(agc): read unregistered graphics programs beside registered shaders

2026-10-15 06:24:14 +05:30

PreparedPublication.cpp

perf(agc): share unchanged shader registry chunks between published versions

2026-10-17 21:22:30 +00:00

PreparedShaders.cpp

Merge pull request #6366 from EagleBearerYS/fix/agc-defer-unpreparable-abi-stages-v3

2026-10-19 03:26:59 +02:00

PreparedUnregisteredGraphics.cpp

fix(agc): read unregistered graphics programs beside registered shaders

2026-10-15 06:24:14 +05:30

ProvokingVertex.cpp

fix(agc): draw rect lists in last provoking vertex mode

2026-10-18 02:38:26 +07:00

QuadPixelMask.cpp

fix(agc): demote quad-mask pixels to helpers instead of killing them

2026-10-15 12:59:39 +13:00

QueueZeroIdleWait.cpp

fix(agc): queue 0 waits instead of spinning when every queue is idle

2026-10-14 14:43:25 +05:30

QueueZeroLateInterrupt.cpp

test(agc): order the late interrupt's commit after queue 0 waits

2026-10-17 16:19:04 +03:00

ReadFirstLane.cpp

…

ReadLane.cpp

…

RecordedKeptBytes.cpp

…

RectListCorner.cpp

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

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

RectListIndirect.cpp

test(agc): require the GPU route for the indirect rect-list draws

2026-10-16 23:20:51 +02:00

RegistrationPrepareShutdown.cpp

…

RegistrationTrace.cpp

fix(agc): trace the title threads that register shaders

2026-10-17 21:33:53 +00:00

RelativeAncillary.cpp

…

RelativeIndexing.cpp

…

RelativeIndexingKnownBits.cpp

test(shader): skip RelativeIndexingKnownBits when the subgroup holds less than a wave32

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

RelativeStageInputs.cpp

…

ResourceCacheGrowth.cpp

test(agc): keep the resource cache growth test short

2026-10-18 03:08:58 +02:00

ReversedOrderBlend.cpp

test(agc): cover moving the export into a reversed-order blend factor

2026-10-15 12:51:27 +02:00

RuntimeImages.cpp

…

SameQueueLabelWait.cpp

…

SampledCacheBudget.cpp

fix(agc): omit unavailable storage MSAA placeholders

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

SampleMaskDepthStencil.cpp

feat(agc): emulate per-pixel multisample coverage masks

2026-10-15 13:00:34 +13:00

SampleMaskDraw.cpp

feat(agc): emulate per-pixel multisample coverage masks

2026-10-15 13:00:34 +13:00

SampleMaskMultisample.cpp

feat(agc): carry pixel iteration and sample locations in recompile requests

2026-10-15 16:06:38 +00:00

SampleMultisampleTarget.cpp

fix(agc): sample undrawn multisampled Z and D 64 KiB swizzle textures, and Z ones whose HTILE is cleared

2026-10-18 22:44:13 +02:00

ScalarBitCount.cpp

…

ScalarGetreg.cpp

…

ScalarProgramEnd.cpp

…

ScalarProgramEndSaved.cpp

…

ScalarProgramFlow.cpp

…

ScalarRelativeIndexing.cpp

…

ScalarSethalt.cpp

…

ScalarSetpc.cpp

…

ScalarSop1.cpp

…

ScalarSop2.cpp

…

ScalarSopkCompare.cpp

…

ScalarSopkMisc.cpp

…

ScalarSopkWaitcnt.cpp

…

ScalarStaticCall.cpp

…

ScalarSwappc.cpp

…

ScalarSwappcFetch.cpp

…

ScalarWideLoads.cpp

…

SdwaCvtF16U16.cpp

…

SdwaCvtF32Ubyte.cpp

fix(recompiler): accept SDWA source selects on v_cvt_f32_ubyte1, ubyte2 and ubyte3

2026-10-14 09:36:10 +13:00

SdwaDestination.cpp

…

SdwaFloatSelectors.cpp

…

SdwaIntegerClamp.cpp

…

SdwaIntegerSelectors.cpp

…

SdwaLdexpF16.cpp

…

SdwaMoveSelectors.cpp

…

ShaderClock.cpp

…

ShutdownThreads.cpp

fix(agc): initialize the async log's condition variables before its worker starts

2026-10-16 08:40:30 -07:00

SinCosNearZero.cpp

…

SinCosReduction.cpp

…

SkipShader.cpp

…

SparseBufferHoles.cpp

test(agc): create fixture command pools with the reset flag

2026-10-18 01:00:16 +00:00

SparseDmaBufImport.cpp

test(agc): check the unaligned import's lead only for a dma-buf import

2026-10-17 15:03:33 -03:00

SparseIndirectIndices.cpp

test(agc): reproduce sparse indirect index ranges

2026-10-16 04:27:23 +06:00

Srgb8ColorTarget.cpp

…

StagedClearSampledRead.cpp

test(libSceAgcDriver): page-align staged-clear write-watch registration on POSIX

2026-10-14 21:48:25 +01:00

StencilNullFeedback.cpp

test(agc): make the null-image fixture module and layout agree

2026-10-16 15:07:46 -04:00

StencilRefExport.cpp

…

StencilRefExportUnused.cpp

test(agc): check draws with an unused stencil reference

2026-10-15 12:51:25 +02:00

StencilReplay.cpp

test(agc): keep stencil replay subgroup contract controls

2026-10-14 22:08:48 +02:00

StorageCacheBudget.cpp

fix(agc): account for batch-retained evicted storage images

2026-10-18 14:34:09 +00:00

StorageFileWrite.cpp

test(agc): keep the queue-idle store out of the file write cases

2026-10-19 01:56:57 -03:00

StorageFlushOrdering.cpp

fix(agc): store pending image results after the range's writers finish

2026-10-15 15:11:12 +02:00

StorageOneSliceAlias.cpp

revert(libSceAgcDriver): restore deferred writeback and main execution tests

2026-10-14 20:57:58 +01:00

StorageWriteBackOverhang.cpp

fix(agc): store a storage image whose last block row runs past its allocation

2026-10-14 01:19:55 +01:00

StoreIdleResults.cpp

…

StoreOverUntracked.cpp

Merge pull request #6744 from Ondra97/codex/fix-store-untracked-lifetime

2026-10-17 17:56:11 +13:00

StreamCopy.cpp

test(agc): stream copy kernel match, refusals and HLE path

2026-10-14 13:53:16 +02:00

SubvectorLoop.cpp

…

SwapAltBlend.cpp

…

TbufferFormatD16.cpp

…

Tessellation.cpp

fix(agc): accept fractional-even tessellation partitioning

2026-10-14 21:16:21 +02:00

TgSizeWaveId.cpp

fix(shader): put the compute wave index in TG_SIZE[11:6]

2026-10-13 17:42:50 +02:00

TypedHeaps.cpp

…

Uint8x4ColorTarget.cpp

feat(agc): render into SW_64KB_Z_X color targets

2026-10-13 22:36:27 +01:00

UnimportedLabelOrder.cpp

…

VCmpEqF16Clamp.cpp

…

VCmpF16DenormalModes.cpp

fix(shader): flush f16 compare inputs by the FLOAT_MODE f16 denormal field

2026-10-17 13:11:24 +07:00

VCmpF32DenormalModes.cpp

fix(shader): translate f32 compares in f32 denormal modes 1 and 2

2026-10-14 21:39:54 +03:00

VCmpF32Denormals.cpp

test(agc): compile each v_cmp denormals program once for its eight input groups

2026-10-13 15:29:43 +02:00

VCvtI16F16OutputModifiers.cpp

…

VCvtPknormF16SourceModifiers.cpp

…

VectorWindowDirectAccess.cpp

…

VectorWindowNarrowing.cpp

perf(recompiler): keep the vector window's unindexed slots out of the array

2026-10-14 22:35:34 +13:00

VertexClipCullDistances.cpp

test(agc): keep the pixel user data alive through its recompile

2026-10-17 14:33:32 +00:00

VertexClipCullDraw.cpp

fix(agc): preserve position export layout on undecodable tail

2026-10-15 15:25:40 +02:00

VertexFetch.cpp

…

VertexUserDataTail.cpp

test(shader): preserve valid positions in vertex tail state checks

2026-10-13 21:29:52 +02:00

ViewPastLastMip.cpp

…

Vop1Control.cpp

…

Vop1FloatUnary.cpp

…

Vop1Vop2Integer.cpp

…

Vop2ModifierForms.cpp

…

Vop3AddSubClamp.cpp

…

Vop3Aliases.cpp

…

Vop3AlignOpSel.cpp

…

Vop3CarryOut.cpp

…

Vop3CompareModifiers.cpp

…

Vop3CubeClamp.cpp

…

Vop3Integer16TernaryOpSel.cpp

…

Vop3Integer.cpp

…

Vop3IntegerAlu.cpp

…

Vop3IntegerClamp.cpp

…

Vop3IntegerClampWide.cpp

…

Vop3LegacySourceModifiers.cpp

…

Vop3Lerp.cpp

…

Vop3MadClamp.cpp

…

Vop3MadI64.cpp

…

Vop3PackLdexp.cpp

…

Vop3QuadSad.cpp

…

Vop3Sad.cpp

…

VopcCompare.cpp

…

VopcCompareF64.cpp

…

VopcCompareTable.cpp

…

VulkanTestDevice.hpp

…

Wave32ExecLaneBits.cpp

…

Wave32Subgroup.cpp

…

WaveIdInWorkgroup.cpp

…

WaveSleep.cpp

…

WaveUniform.cpp

…

WindowLaneSpill.cpp

fix(recompiler): read lane spills back through relative-addressing window registers

2026-10-15 21:52:10 +13:00

WriteDataWrittenAhead.cpp

…

WriteLane.cpp

…

ZeroPositionVertex.cpp

…

ZeroStrideVertexFetch.cpp

perf(shader): map cached SPIR-V from the shader cache files

2026-10-14 10:36:24 +04:00