Johannes Reifferscheid a6a9215b93

Lower shuffle to single-result form if possible. (#84321) ...

We currently always lower shuffle to the struct-returning variant. I saw
some cases where this survived all the way through ptx, resulting in
increased register usage. The easiest fix is to simply lower to the
single-result version when the predicate is unused.

2024-03-21 10:33:49 +01:00

..

AffineToStandard

[mlir] Fix shared builds. NFC

2024-02-28 09:31:05 -05:00

AMDGPUToROCDL

[mlir][AMDGPU] Improve amdgpu.lds_barrier, add warnings (#77942)

2024-03-11 10:06:49 -05:00

ArithCommon

…

ArithToAMDGPU

[mlir] Use create instead of createOrFold for ConstantOp as folding has no effect (NFC) (#80129)

2024-01-31 23:40:37 -08:00

ArithToArmSME

[mlir][ArmSME] Add arith-to-arm-sme conversion pass (#78197)

2024-01-22 09:23:11 +00:00

ArithToEmitC

[mlir][EmitC] Add an emitc.conditional operator (#84883)

2024-03-12 11:27:26 +01:00

ArithToLLVM

[mlir] [arith] add shl overflow flag in Arith and lower to SPIR-V and LLVMIR (#79828)

2024-01-30 11:11:29 +00:00

ArithToSPIRV

[mlir] [arith] add shl overflow flag in Arith and lower to SPIR-V and LLVMIR (#79828)

2024-01-30 11:11:29 +00:00

ArmNeon2dToIntr

…

ArmSMEToLLVM

[mlir][ArmSME] Support 4-way widening outer products (#79288)

2024-02-07 08:17:47 +00:00

ArmSMEToSCF

[mlir][ArmSME] Fix loop bounds of masked loads/stores (#78983)

2024-01-26 09:39:43 +00:00

AsyncToLLVM

[mlir] Use OpBuilder::createBlock in op builders and patterns (#82770)

2024-02-24 09:10:07 +01:00

BufferizationToMemRef

…

ComplexToLibm

…

ComplexToLLVM

…

ComplexToSPIRV

…

ComplexToStandard

[mlir][complex] Support Fastmath flag in conversion of complex.sqrt to standard (#85019)

2024-03-14 15:53:28 +09:00

ControlFlowToLLVM

…

ControlFlowToSCF

[mlir] Use OpBuilder::createBlock in op builders and patterns (#82770)

2024-02-24 09:10:07 +01:00

ControlFlowToSPIRV

…

ConvertToLLVM

…

FuncToEmitC

[mlir][EmitC] Correct comment (NFC)

2024-03-06 12:46:29 +00:00

FuncToLLVM

[mlir] Use OpBuilder::createBlock in op builders and patterns (#82770)

2024-02-24 09:10:07 +01:00

FuncToSPIRV

…

GPUCommon

[mlir][LLVM] Use int32_t to indirectly construct GEPArg (#79562)

2024-01-26 14:27:51 +01:00

GPUToNVVM

Lower shuffle to single-result form if possible. (#84321)

2024-03-21 10:33:49 +01:00

GPUToROCDL

[mlir][ROCDL] Set the LLVM data layout when lowering to ROCDL LLVM (#74501)

2024-02-27 09:59:50 -06:00

GPUToSPIRV

[mlir][spirv] Use AttrTypeReplacer in map-memref-storage-class. NFC. (#80055)

2024-01-30 20:32:26 -05:00

GPUToVulkan

…

IndexToLLVM

…

IndexToSPIRV

…

LinalgToStandard

[MLIR] Add missing MLIRLinalgTransforms to LinalgToStandard conv (#84545)

2024-03-20 09:26:16 +00:00

LLVMCommon

…

MathToFuncs

…

MathToLibm

…

MathToLLVM

[mlir][math] Propagate scalability in convert-math-to-llvm (#82635)

2024-02-23 09:48:58 +00:00

MathToSPIRV

…

MemRefToLLVM

[MLIR] Add missing MLIRFuncDialect dep to MLIRMemRefToLLVM (#84546)

2024-03-19 17:25:37 -07:00

MemRefToSPIRV

[mlir][spirv] Retain nontemporal attribute when converting memref load/store (#82119)

2024-03-02 15:49:18 -08:00

NVGPUToNVVM

[mlir][nvgpu] Make phaseParity of mbarrier.try_wait i1 (#81460)

2024-02-13 09:50:34 +01:00

NVVMToLLVM

[MLIR] Add missing MLIRFuncDialect dep to MLIRNVVMToLLVM (#84548)

2024-03-11 23:07:49 +00:00

OpenACCToSCF

…

OpenMPToLLVM

[MLIR][OpenMP] NFC: Uniformize OpenMP ops names (#85393)

2024-03-20 11:19:38 +00:00

PDLToPDLInterp

Add llvm::min/max_element and use it in llvm/ and mlir/ directories. (#84678)

2024-03-10 20:00:13 -07:00

ReconcileUnrealizedCasts

…

SCFToControlFlow

…

SCFToEmitC

[mlir][EmitC] Drop unused code (NFC) (#80325)

2024-02-04 18:51:47 +01:00

SCFToGPU

…

SCFToOpenMP

[MLIR][OpenMP] NFC: Uniformize OpenMP ops names (#85393)

2024-03-20 11:19:38 +00:00

SCFToSPIRV

[mlir] Use OpBuilder::createBlock in op builders and patterns (#82770)

2024-02-24 09:10:07 +01:00

ShapeToStandard

[mlir] Use arith max or min ops instead of cmp + select (#82178)

2024-02-21 12:28:05 -08:00

SPIRVCommon

…

SPIRVToLLVM

[mlir][spirv] Fix a crash of typeConverter with non supported type (#79955)

2024-01-30 22:36:56 -05:00

TensorToLinalg

…

TensorToSPIRV

…

TosaToArith

…

TosaToLinalg

[TOSA] TosaToLinalg: fix int64_t min/max lowering of clamp (#82641)

2024-02-22 21:16:33 +01:00

TosaToMLProgram

…

TosaToSCF

…

TosaToTensor

[mlir] Use create instead of createOrFold for ConstantOp as folding has no effect (NFC) (#80129)

2024-01-31 23:40:37 -08:00

UBToLLVM

…

UBToSPIRV

…

VectorToArmSME

[mlir][ArmSME] Add arith-to-arm-sme conversion pass (#78197)

2024-01-22 09:23:11 +00:00

VectorToGPU

[MLIR] Fix crash in AffineMap::replace for zero result maps (#80930)

2024-02-08 19:16:29 +05:30

VectorToLLVM

[mlir][ArmNeon] Implements LowerVectorToArmNeon Pattern for SMMLA (#81895)

2024-03-08 14:50:13 -08:00

VectorToSCF

…

VectorToSPIRV

…

CMakeLists.txt

[mlir][EmitC] Add Arith to EmitC conversions (#84151)

2024-03-07 11:34:11 +01:00