Jay Foad 2012b25420 [AMDGPU][GlobalISel] Disable fixed-point iteration in all Combiners (#105517) ...

Disable fixed-point iteration in all AMDGPU Combiners after #102163.

This saves around 2% compile time in ad hoc testing on some large
graphics shaders. I did not notice any regressions in the generated
code, just a bunch of harmless differences in instruction selection and
register allocation.

2024-08-22 17:14:53 +01:00

..

AsmParser

[AMDGPU][NFC] Rename isHi() to isHi16Reg() for clarity. (#103888)

2024-08-14 17:04:15 +01:00

Disassembler

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

MCA

[AMDGPU] Concatenate nested namespaces. NFC.

2024-07-16 15:36:49 +01:00

MCTargetDesc

[MC] Replace HasAggressiveSymbolFolding with SetDirectiveSuppressesReloc. NFC

2024-08-15 13:29:47 -07:00

TargetInfo

…

Utils

[AMDGPU] Move AMDGPUMemoryUtils out of Utils. NFC. (#104930)

2024-08-20 16:15:46 +01:00

AMDGPU.h

AMDGPU/NewPM: Port AMDGPULateCodeGenPrepare to new pass manager (#102806)

2024-08-12 15:09:12 +04:00

AMDGPU.td

[AMDGPU] GFX12 VMEM loads can write VGPR results out of order (#105549)

2024-08-22 11:46:51 +01:00

AMDGPUAliasAnalysis.cpp

…

AMDGPUAliasAnalysis.h

[IR] Don't include Module.h in Analysis.h (NFC) (#97023)

2024-06-28 14:30:47 +02:00

AMDGPUAlwaysInlinePass.cpp

[AMDGPU] Always Inline preserved analyses (#91198)

2024-05-09 06:17:01 -04:00

AMDGPUAnnotateKernelFeatures.cpp

…

AMDGPUAnnotateUniformValues.cpp

[AMDGPU] Move AMDGPUMemoryUtils out of Utils. NFC. (#104930)

2024-08-20 16:15:46 +01:00

AMDGPUArgumentUsageInfo.cpp

AMDGPU: Add plumbing for private segment size argument (#96445)

2024-06-25 16:20:51 +02:00

AMDGPUArgumentUsageInfo.h

AMDGPU: Add plumbing for private segment size argument (#96445)

2024-06-25 16:20:51 +02:00

AMDGPUAsanInstrumentation.cpp

[AMDGPU] Update instrumentAddress method to support aligned size and unusual size accesses. (#104804)

2024-08-21 10:31:42 +05:30

AMDGPUAsanInstrumentation.h

[AMDGPU] Update instrumentAddress method to support aligned size and unusual size accesses. (#104804)

2024-08-21 10:31:42 +05:30

AMDGPUAsmPrinter.cpp

[AMDGPU] MCExpr printing helper with KnownBits support (#95951)

2024-08-15 13:43:13 +01:00

AMDGPUAsmPrinter.h

[AsmPrinter] Don't EmitToStreamer instructions lowered by tblgenned code (#100803)

2024-07-29 19:18:18 +03:00

AMDGPUAtomicOptimizer.cpp

[AMDGPU] Enable atomic optimizer for divergent i64 and double values (#96934)

2024-07-15 17:49:09 +05:30

AMDGPUAttributes.def

…

AMDGPUAttributor.cpp

Reapply "[Attributor][AMDGPU] Enable AAIndirectCallInfo for AMDAttributor (#100952)"

2024-08-14 17:16:47 -04:00

AMDGPUCallingConv.td

Revert "Update amdgpu_gfx functions to use s0-s3 for inreg SGPR arguments on targets using scratch instructions for stack #78226" (#86273)

2024-03-26 11:01:08 +01:00

AMDGPUCallLowering.cpp

[AMDGPU] Use llvm::any_of, llvm::all_of, and llvm::none_of (NFC) (#103007)

2024-08-13 00:07:54 -07:00

AMDGPUCallLowering.h

…

AMDGPUCodeGenPrepare.cpp

AMDGPU: Fix assert from wrong address space size assumption (#97267)

2024-07-02 23:18:25 +02:00

AMDGPUCombine.td

[AMDGPU] Fix typo in function name

2024-05-06 08:55:09 +01:00

AMDGPUCombinerHelper.cpp

AMDGPU/GlobalISel: Use correct type for intrinsic ID

2024-05-30 14:31:19 +02:00

AMDGPUCombinerHelper.h

[NFC] Eliminate top-level "using namespace" from some headers. (#102751)

2024-08-11 13:10:48 -07:00

AMDGPUCtorDtorLowering.cpp

[llvm] Replace calls to Type::getPointerTo (NFC)

2023-11-30 13:18:51 -06:00

AMDGPUCtorDtorLowering.h

…

AMDGPUExportClustering.cpp

…

AMDGPUExportClustering.h

…

AMDGPUFeatures.td

…

AMDGPUFrameLowering.cpp

…

AMDGPUFrameLowering.h

…

AMDGPUGenRegisterBankInfo.def

Silence illegal address computation warning (#83244)

2024-03-01 21:41:12 +05:30

AMDGPUGISel.td

AMDGPU: Add tonearest and towardzero roundings for intrinsic llvm.fptrunc.round (#104486)

2024-08-17 11:22:47 -07:00

AMDGPUGlobalISelDivergenceLowering.cpp

[CodeGen][NewPM] Split MachinePostDominators into a concrete analysis result (#95113)

2024-06-12 14:29:22 +08:00

AMDGPUGlobalISelUtils.cpp

[AMDGPU] Remove hasAtomicFaddRtnForTy as it is not used anywhere (#82841)

2024-02-23 21:14:38 -05:00

AMDGPUGlobalISelUtils.h

[AMDGPU] Remove hasAtomicFaddRtnForTy as it is not used anywhere (#82841)

2024-02-23 21:14:38 -05:00

AMDGPUHSAMetadataStreamer.cpp

[AMDGPU] Concatenate nested namespaces. NFC.

2024-07-16 15:36:49 +01:00

AMDGPUHSAMetadataStreamer.h

[AMDGPU] MCExpr-ify AMDGPU HSAMetadata (#94788)

2024-06-26 16:39:08 +01:00

AMDGPUIGroupLP.cpp

[AMDGPU] Use llvm::any_of, llvm::all_of, and llvm::none_of (NFC) (#103007)

2024-08-13 00:07:54 -07:00

AMDGPUIGroupLP.h

[AMDGPU] Do not attempt to fallback to default mutations (#83208)

2024-02-27 18:04:59 -08:00

AMDGPUImageIntrinsicOptimizer.cpp

[AMDGPU] Use llvm::none_of (NFC)

2024-01-24 00:27:37 -08:00

AMDGPUInsertDelayAlu.cpp

…

AMDGPUInsertSingleUseVDST.cpp

[AMDGPU] Exclude certain opcodes from being marked as single use (#91802)

2024-06-12 10:43:23 +01:00

AMDGPUInstCombineIntrinsic.cpp

[AMDGPU] Constant folding of llvm.amdgcn.trig.preop (#98562)

2024-07-18 09:40:37 -07:00

AMDGPUInstrInfo.cpp

…

AMDGPUInstrInfo.h

AMDGPU/GlobalISel: Use correct type for intrinsic ID

2024-05-30 14:31:19 +02:00

AMDGPUInstrInfo.td

[AMDGPU] Add a trap lowering workaround for gfx11 (#85854)

2024-04-24 09:43:54 -04:00

AMDGPUInstructions.td

AMDGPU: Remove flat/global atomic fadd v2bf16 intrinsics (#97050)

2024-08-21 14:26:42 +04:00

AMDGPUInstructionSelector.cpp

[AMDGPU][True16][CodeGen] Support AND/OR/XOR and LDEXP True16 format (#102620)

2024-08-13 12:23:39 -04:00

AMDGPUInstructionSelector.h

AMDGPU/GlobalISel: Fix isExtractHiElt when selecting fma_mix (#102130)

2024-08-07 12:13:39 +02:00

AMDGPUISelDAGToDAG.cpp

[KnownBits] Add KnownBits::add and KnownBits::sub helper wrappers. (#99468)

2024-08-12 10:21:28 +01:00

AMDGPUISelDAGToDAG.h

[NFC] Eliminate top-level "using namespace" from some headers. (#102751)

2024-08-11 13:10:48 -07:00

AMDGPUISelLowering.cpp

AMDGPU: Add tonearest and towardzero roundings for intrinsic llvm.fptrunc.round (#104486)

2024-08-17 11:22:47 -07:00

AMDGPUISelLowering.h

AMDGPU: Add tonearest and towardzero roundings for intrinsic llvm.fptrunc.round (#104486)

2024-08-17 11:22:47 -07:00

AMDGPULateCodeGenPrepare.cpp

AMDGPU/NewPM: Port AMDGPULateCodeGenPrepare to new pass manager (#102806)

2024-08-12 15:09:12 +04:00

AMDGPULegalizerInfo.cpp

[AMDGPU][GlobalISel] Save a copy in one case of addrspacecast (#104789)

2024-08-19 18:22:29 +01:00

AMDGPULegalizerInfo.h

AMDGPU: Remove ds_fmin/ds_fmax intrinsics (#96739)

2024-06-27 15:35:24 +02:00

AMDGPULibCalls.cpp

[AMDGPU] Use range-based for loops. NFC. (#99047)

2024-07-17 10:18:03 +01:00

AMDGPULibFunc.cpp

[AMDGPU] clang-tidy: use std::make_unique. NFC.

2024-07-17 07:58:09 +01:00

AMDGPULibFunc.h

…

AMDGPULowerBufferFatPointers.cpp

[AMDGPU] Add intrinsic for raw atomic buffer loads (#97707)

2024-07-22 18:04:49 +02:00

AMDGPULowerKernelArguments.cpp

[IR] Add getDataLayout() helpers to Function and GlobalValue (#96919)

2024-06-28 08:36:49 +02:00

AMDGPULowerKernelAttributes.cpp

[PatternMatch] Use m_SpecificCmp matchers. NFC. (#100878)

2024-07-29 10:04:06 +08:00

AMDGPULowerModuleLDSPass.cpp

[AMDGPU] Move AMDGPUMemoryUtils out of Utils. NFC. (#104930)

2024-08-20 16:15:46 +01:00

AMDGPUMachineCFGStructurizer.cpp

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

AMDGPUMachineFunction.cpp

[AMDGPU] Use member initializers. NFC.

2024-07-16 15:29:10 +01:00

AMDGPUMachineFunction.h

[AMDGPU] Add dynamic LDS size implicit kernel argument to CO-v5 (#65273)

2024-01-04 19:05:12 +05:30

AMDGPUMachineModuleInfo.cpp

[IR] Don't include Module.h in Analysis.h (NFC) (#97023)

2024-06-28 14:30:47 +02:00

AMDGPUMachineModuleInfo.h

[IR] Remove Function.h include from Analysis.h (NFC) (#97056)

2024-06-28 16:41:00 +02:00

AMDGPUMacroFusion.cpp

[DAG] clang-format createBranchMacroFusionDAGMutation calls. NFC.

2023-11-20 12:13:09 +00:00

AMDGPUMacroFusion.h

…

AMDGPUMarkLastScratchLoad.cpp

[CodeGen][NewPM] Port LiveIntervals to new pass manager (#98118)

2024-07-10 19:34:48 +08:00

AMDGPUMCInstLower.cpp

[AsmPrinter] Don't EmitToStreamer instructions lowered by tblgenned code (#100803)

2024-07-29 19:18:18 +03:00

AMDGPUMCInstLower.h

…

AMDGPUMemoryUtils.cpp

[AMDGPU] Move AMDGPUMemoryUtils out of Utils. NFC. (#104930)

2024-08-20 16:15:46 +01:00

AMDGPUMemoryUtils.h

[AMDGPU] Move AMDGPUMemoryUtils out of Utils. NFC. (#104930)

2024-08-20 16:15:46 +01:00

AMDGPUMIRFormatter.cpp

[AMDGPU] Introduce a pseudo mnemonic for S_DELAY_ALU in MIR. (#96004)

2024-06-20 15:59:08 -04:00

AMDGPUMIRFormatter.h

[AMDGPU] Introduce a pseudo mnemonic for S_DELAY_ALU in MIR. (#96004)

2024-06-20 15:59:08 -04:00

AMDGPUOpenCLEnqueuedBlockLowering.cpp

[NFC] Remove Type::getInt8PtrTy (#71029)

2023-11-07 17:26:26 +01:00

AMDGPUPassRegistry.def

AMDGPU/NewPM: Port AMDGPULateCodeGenPrepare to new pass manager (#102806)

2024-08-12 15:09:12 +04:00

AMDGPUPerfHintAnalysis.cpp

NewPM/AMDGPU: Port AMDGPUPerfHintAnalysis to new pass manager (#102645)

2024-08-11 15:11:10 +04:00

AMDGPUPerfHintAnalysis.h

NewPM/AMDGPU: Port AMDGPUPerfHintAnalysis to new pass manager (#102645)

2024-08-11 15:11:10 +04:00

AMDGPUPostLegalizerCombiner.cpp

[AMDGPU][GlobalISel] Disable fixed-point iteration in all Combiners (#105517)

2024-08-22 17:14:53 +01:00

AMDGPUPredicateControl.td

[AMDGPU][TableGen][NFC] Combine predicates without using classes. (#82346)

2024-02-21 11:45:36 +00:00

AMDGPUPreLegalizerCombiner.cpp

[AMDGPU][GlobalISel] Disable fixed-point iteration in all Combiners (#105517)

2024-08-22 17:14:53 +01:00

AMDGPUPrintfRuntimeBinding.cpp

Reapply "[OpenMP][libc] Remove special handling for OpenMP printf (#98940)"

2024-07-26 17:21:56 -05:00

AMDGPUPromoteAlloca.cpp

AMDGPU: Stop promoting allocas with addrspacecast users (#104051)

2024-08-14 21:53:38 +04:00

AMDGPUPromoteKernelArguments.cpp

[AMDGPU] Move AMDGPUMemoryUtils out of Utils. NFC. (#104930)

2024-08-20 16:15:46 +01:00

AMDGPUPTNote.h

…

AMDGPURegBankCombiner.cpp

[AMDGPU][GlobalISel] Disable fixed-point iteration in all Combiners (#105517)

2024-08-22 17:14:53 +01:00

AMDGPURegBankSelect.cpp

Revert "[CodeGen] Remove applySplitCriticalEdges in MachineDominatorTree (#97055)"

2024-07-11 09:13:37 +02:00

AMDGPURegBankSelect.h

…

AMDGPURegisterBankInfo.cpp

AMDGPU: Remove flat/global atomic fadd v2bf16 intrinsics (#97050)

2024-08-21 14:26:42 +04:00

AMDGPURegisterBankInfo.h

[AMDGPU] Make isScalarLoadLegal a member of AMDGPURegisterBankInfo. NFC.

2024-01-09 11:53:12 +00:00

AMDGPURegisterBanks.td

[AMDGPU][True16] Don't use the VGPR_LO/HI16 register classes. (#76440)

2023-12-28 11:48:25 +00:00

AMDGPURemoveIncompatibleFunctions.cpp

[AMDGPU] Fix and add namespace closing comments. NFC.

2024-07-16 16:56:31 +01:00

AMDGPUResourceUsageAnalysis.cpp

AMDGPU: Remove a pointless SIFunctionResourceInfo::getTotalNumVgprs o… (#101158)

2024-07-30 20:04:39 +02:00

AMDGPUResourceUsageAnalysis.h

AMDGPU: Remove a pointless SIFunctionResourceInfo::getTotalNumVgprs o… (#101158)

2024-07-30 20:04:39 +02:00

AMDGPURewriteOutArguments.cpp

[llvm] Stop including llvm/ADT/SmallSet.h (NFC)

2023-11-11 12:32:15 -08:00

AMDGPURewriteUndefForPHI.cpp

[AMDGPU] Do not preserve UniformityInfo (#76696)

2024-01-02 13:42:56 +05:30

AMDGPUSearchableTables.td

AMDGPU: Remove flat/global atomic fadd v2bf16 intrinsics (#97050)

2024-08-21 14:26:42 +04:00

AMDGPUSetWavePriority.cpp

…

AMDGPUSplitModule.cpp

[AMDGPU][llvm-split] Fix another division by zero (#104421)

2024-08-15 12:54:26 +01:00

AMDGPUSplitModule.h

[NFC][AMDGPU] Make AMDGPUSplitModule a ModulePass (#95773)

2024-06-18 09:16:32 +02:00

AMDGPUSubtarget.cpp

[AMDGPU] Split GCNSubtarget into its own file. NFC. (#105525)

2024-08-21 19:11:02 +01:00

AMDGPUSubtarget.h

[AMDGPU] clang-format: pass Triple by value and std::move it. NFC.

2024-07-16 15:52:58 +01:00

AMDGPUTargetMachine.cpp

AMDGPU: Temporarily stop adding AtomicExpand to new PM passes

2024-08-21 00:19:37 +04:00

AMDGPUTargetMachine.h

AMDGPU/NewPM: Start filling out addIRPasses (#102884)

2024-08-20 23:38:05 +04:00

AMDGPUTargetObjectFile.cpp

[llvm] Use StringRef::{starts,ends}_with (NFC) (#74956)

2023-12-11 21:01:36 -08:00

AMDGPUTargetObjectFile.h

…

AMDGPUTargetTransformInfo.cpp

AMDGPU: Correct costs of saturating add/sub intrinsics (#100808)

2024-08-09 12:55:15 +04:00

AMDGPUTargetTransformInfo.h

[LowerMemIntrinsics][NFC] Use Align in TTI::getMemcpyLoopLoweringType (#100984)

2024-07-29 13:40:53 +02:00

AMDGPUUnifyDivergentExitNodes.cpp

[AMDGPU] clang-tidy: use emplace_back instead of push_back. NFC.

2024-07-17 08:27:35 +01:00

AMDGPUUnifyDivergentExitNodes.h

…

AMDGPUUnifyMetadata.cpp

…

AMDKernelCodeT.h

…

BUFInstructions.td

[LLVM][AMDGPU] Copy isConvergent from Pseudo to Real instructions (#99658)

2024-07-25 18:01:07 +02:00

CaymanInstructions.td

…

CMakeLists.txt

[AMDGPU] Split GCNSubtarget into its own file. NFC. (#105525)

2024-08-21 19:11:02 +01:00

DSDIRInstructions.td

AMDGPU: Use PseudoInstr to name SIMCInstr for DSDIR and SOPs, NFC (#87537)

2024-04-03 11:35:09 -07:00

DSInstructions.td

[LLVM][AMDGPU] Copy isConvergent from Pseudo to Real instructions (#99658)

2024-07-25 18:01:07 +02:00

EvergreenInstructions.td

DAG: Replace bitwidth with type in suffix in atomic tablegen ops (#94845)

2024-06-13 11:52:22 +02:00

EXPInstructions.td

[AMDGPU] Remove unnecessary predicates from aliases. NFC. (#91602)

2024-05-10 15:04:38 +01:00

FLATInstructions.td

AMDGPU: Remove flat/global atomic fadd v2bf16 intrinsics (#97050)

2024-08-21 14:26:42 +04:00

GCNCreateVOPD.cpp

[AMDGPU] clang-tidy: define trivial constructors with = default. NFC.

2024-07-16 15:41:54 +01:00

GCNDPPCombine.cpp

[AMDGPU] Support byte_sel modifier for v_cvt_f32_fp8 and v_cvt_f32_bf8 (#90887)

2024-05-02 12:03:51 -07:00

GCNHazardRecognizer.cpp

[AMDGPU] Implement workaround for GFX11.5 export priority (#99273)

2024-07-23 17:03:21 +09:00

GCNHazardRecognizer.h

[AMDGPU] Implement workaround for GFX11.5 export priority (#99273)

2024-07-23 17:03:21 +09:00

GCNILPSched.cpp

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

GCNIterativeScheduler.cpp

[AMDGPU] Fix and add namespace closing comments. NFC.

2024-07-16 16:56:31 +01:00

GCNIterativeScheduler.h

…

GCNMinRegStrategy.cpp

…

GCNNSAReassign.cpp

[AMDGPU] Use using instead of typedef. NFC.

2024-07-16 16:44:12 +01:00

GCNPreRALongBranchReg.cpp

…

GCNPreRAOptimizations.cpp

[CodeGen][NewPM] Port LiveIntervals to new pass manager (#98118)

2024-07-10 19:34:48 +08:00

GCNProcessors.td

[AMDGPU] Add a new target gfx1152 (#94534)

2024-06-06 12:16:11 -04:00

GCNRegPressure.cpp

[CodeGen][NewPM] Port LiveIntervals to new pass manager (#98118)

2024-07-10 19:34:48 +08:00

GCNRegPressure.h

[AMDGPU] NFC: Add helper to query archVGPR (#102690)

2024-08-12 14:01:12 -07:00

GCNRewritePartialRegUses.cpp

[AMDGPU] Use using instead of typedef. NFC.

2024-07-16 16:44:12 +01:00

GCNSchedStrategy.cpp

[AMDGPU] Use llvm::any_of, llvm::all_of, and llvm::none_of (NFC) (#103007)

2024-08-13 00:07:54 -07:00

GCNSchedStrategy.h

[NFCI][AMDGPU] Try to use PressureDiff to Calculate RegPressure. (#94221)

2024-06-14 08:34:53 +02:00

GCNSubtarget.cpp

[AMDGPU] Split GCNSubtarget into its own file. NFC. (#105525)

2024-08-21 19:11:02 +01:00

GCNSubtarget.h

[AMDGPU] GFX12 VMEM loads can write VGPR results out of order (#105549)

2024-08-22 11:46:51 +01:00

GCNVOPDUtils.cpp

[AMDGPU] Update VOP instructions for GFX12 (#74853)

2023-12-12 11:38:24 +01:00

GCNVOPDUtils.h

…

InstCombineTables.td

…

MIMGInstructions.td

[AMDGPU] Add no return image_sample intrinsics and instructions (#97542)

2024-07-20 17:26:58 +09:00

R600.h

…

R600.td

[AMDGPU][NFC] Extend PredicateControl to support True16 predicates. (#82245)

2024-02-20 11:37:44 +00:00

R600AsmPrinter.cpp

…

R600AsmPrinter.h

…

R600ClauseMergePass.cpp

…

R600ControlFlowFinalizer.cpp

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

R600Defines.h

…

R600EmitClauseMarkers.cpp

[AMDGPU] Use range-based for loops. NFC. (#99047)

2024-07-17 10:18:03 +01:00

R600ExpandSpecialInstrs.cpp

…

R600FrameLowering.cpp

…

R600FrameLowering.h

…

R600InstrFormats.td

…

R600InstrInfo.cpp

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

R600InstrInfo.h

…

R600InstrInfo.td

…

R600Instructions.td

[AMDGPU] Generic lowering for rint and nearbyint (#69596)

2023-11-14 18:49:21 +01:00

R600ISelDAGToDAG.cpp

[NFC] Eliminate top-level "using namespace" from some headers. (#102751)

2024-08-11 13:10:48 -07:00

R600ISelLowering.cpp

[AMDGPU][R600] Move R600CodeGenPassBuilder into R600TargetMachine(NFC). (#103721)

2024-08-19 20:40:12 +05:30

R600ISelLowering.h

…

R600MachineCFGStructurizer.cpp

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

R600MachineFunctionInfo.cpp

…

R600MachineFunctionInfo.h

…

R600MachineScheduler.cpp

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

R600MachineScheduler.h

[NFC] Eliminate top-level "using namespace" from some headers. (#102751)

2024-08-11 13:10:48 -07:00

R600MCInstLower.cpp

…

R600OpenCLImageTypeLoweringPass.cpp

[AMDGPU] Use range-based for loops. NFC. (#99047)

2024-07-17 10:18:03 +01:00

R600OptimizeVectorRegisters.cpp

[AMDGPU] clang-tidy: use emplace_back instead of push_back. NFC.

2024-07-17 08:27:35 +01:00

R600Packetizer.cpp

[Target] Use range-based for loops (NFC) (#98705)

2024-07-13 17:40:51 -07:00

R600Processors.td

…

R600RegisterInfo.cpp

…

R600RegisterInfo.h

…

R600RegisterInfo.td

…

R600Schedule.td

…

R600Subtarget.cpp

…

R600Subtarget.h

…

R600TargetMachine.cpp

[AMDGPU][R600] Move R600CodeGenPassBuilder into R600TargetMachine(NFC). (#103721)

2024-08-19 20:40:12 +05:30

R600TargetMachine.h

[AMDGPU][R600] Move R600CodeGenPassBuilder into R600TargetMachine(NFC). (#103721)

2024-08-19 20:40:12 +05:30

R600TargetTransformInfo.cpp

[IR] Add getDataLayout() helpers to Function and GlobalValue (#96919)

2024-06-28 08:36:49 +02:00

R600TargetTransformInfo.h

…

R700Instructions.td

…

SIAnnotateControlFlow.cpp

AMDGPU/NewPM: Port SIAnnotateControlFlow to new pass manager (#102653)

2024-08-10 07:02:21 +04:00

SIDefines.h

[AMDGPU][NFC] Rename isHi() to isHi16Reg() for clarity. (#103888)

2024-08-14 17:04:15 +01:00

SIFixSGPRCopies.cpp

AMDGPU/NewPM: Port SIFixSGPRCopies to new pass manager (#102614)

2024-08-09 17:52:41 +04:00

SIFixSGPRCopies.h

AMDGPU/NewPM: Port SIFixSGPRCopies to new pass manager (#102614)

2024-08-09 17:52:41 +04:00

SIFixVGPRCopies.cpp

…

SIFoldOperands.cpp

[AMDGPU][True16][CodeGen] support v_mov_b16 and v_swap_b16 in true16 format (#102198)

2024-08-08 16:52:59 -04:00

SIFormMemoryClauses.cpp

[AMDGPU] Use using instead of typedef. NFC.

2024-07-16 16:44:12 +01:00

SIFrameLowering.cpp

Reapply "AMDGPU: Move attributor into optimization pipeline (#83131)" and follow up commit "clang/AMDGPU: Defeat attribute optimization in attribute test" (#98851)

2024-07-15 11:51:44 +04:00

SIFrameLowering.h

Revert "[AMDGPU] Compiler should synthesize private buffer resource descriptor from flat_scratch_init" (#81234)

2024-02-09 09:57:38 +01:00

SIInsertHardClauses.cpp

[Codegen] Make Width in getMemOperandsWithOffsetWidth a LocationSize. (#83875)

2024-03-06 17:40:13 +00:00

SIInsertWaitcnts.cpp

[AMDGPU] GFX12 VMEM loads can write VGPR results out of order (#105549)

2024-08-22 11:46:51 +01:00

SIInstrFormats.td

[AMDGPU][True16][MC] Support v_swap_b16. (#100442)

2024-08-01 22:08:45 +04:00

SIInstrInfo.cpp

[AMDGPU] Disable inline constants for pseudo scalar transcendentals (#104395)

2024-08-17 16:52:38 +09:00

SIInstrInfo.h

[AMDGPU] Disable inline constants for pseudo scalar transcendentals (#104395)

2024-08-17 16:52:38 +09:00

SIInstrInfo.td

AMDGPU: Add tonearest and towardzero roundings for intrinsic llvm.fptrunc.round (#104486)

2024-08-17 11:22:47 -07:00

SIInstructions.td

AMDGPU: Add tonearest and towardzero roundings for intrinsic llvm.fptrunc.round (#104486)

2024-08-17 11:22:47 -07:00

SIISelLowering.cpp

AMDGPU: Remove flat/global atomic fadd v2bf16 intrinsics (#97050)

2024-08-21 14:26:42 +04:00

SIISelLowering.h

AMDGPU: Add tonearest and towardzero roundings for intrinsic llvm.fptrunc.round (#104486)

2024-08-17 11:22:47 -07:00

SILateBranchLowering.cpp

Revert "[CodeGen] Remove applySplitCriticalEdges in MachineDominatorTree (#97055)"

2024-07-11 09:13:37 +02:00

SILoadStoreOptimizer.cpp

[AMDGPU] add missing checks in processBaseWithConstOffset (#102310)

2024-08-12 11:54:02 +04:00

SILowerControlFlow.cpp

[CodeGen][NewPM] Port LiveIntervals to new pass manager (#98118)

2024-07-10 19:34:48 +08:00

SILowerI1Copies.cpp

AMDGPU/NewPM: Port SILowerI1Copies to new pass manager (#102663)

2024-08-10 07:08:22 +04:00

SILowerI1Copies.h

AMDGPU/GlobalISelDivergenceLowering: select divergent i1 phis (#80003)

2024-02-05 14:07:01 +01:00

SILowerSGPRSpills.cpp

[AMDGPU] clang-tidy: use emplace_back instead of push_back. NFC.

2024-07-17 08:27:35 +01:00

SILowerWWMCopies.cpp

[CodeGen][NewPM] Port LiveIntervals to new pass manager (#98118)

2024-07-10 19:34:48 +08:00

SIMachineFunctionInfo.cpp

[AMDGPU] Move AMDGPUCodeGenPassBuilder into AMDGPUTargetMachine(NFC) (#103720)

2024-08-19 20:32:55 +05:30

SIMachineFunctionInfo.h

AMDGPU: Add plumbing for private segment size argument (#96445)

2024-06-25 16:20:51 +02:00

SIMachineScheduler.cpp

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

SIMachineScheduler.h

…

SIMemoryLegalizer.cpp

AMDGPU: Query MachineModuleInfo from PM instead of MachineFunction (#99679)

2024-07-22 08:55:39 +04:00

SIModeRegister.cpp

AMDGPU: Add tonearest and towardzero roundings for intrinsic llvm.fptrunc.round (#104486)

2024-08-17 11:22:47 -07:00

SIModeRegisterDefaults.cpp

Reapply "AMDGPU: Implement llvm.set.rounding (#88587)" series (#91113)

2024-05-06 09:09:19 +02:00

SIModeRegisterDefaults.h

Reapply "AMDGPU: Implement llvm.set.rounding (#88587)" series (#91113)

2024-05-06 09:09:19 +02:00

SIOptimizeExecMasking.cpp

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

SIOptimizeExecMaskingPreRA.cpp

[CodeGen][NewPM] Port LiveIntervals to new pass manager (#98118)

2024-07-10 19:34:48 +08:00

SIOptimizeVGPRLiveRange.cpp

[CodeGen][NewPM] Port LiveVariables to new pass manager (#97880)

2024-07-09 10:50:43 +08:00

SIPeepholeSDWA.cpp

[AMDGPU] clang-tidy: no else after return etc. NFC. (#99298)

2024-07-17 21:11:00 +01:00

SIPostRABundler.cpp

…

SIPreAllocateWWMRegs.cpp

[CodeGen][NewPM] Port LiveIntervals to new pass manager (#98118)

2024-07-10 19:34:48 +08:00

SIPreEmitPeephole.cpp

[AMDGPU] Update EXECZ retention in SIPreEmitPeephole for GFX10/12 (#97676)

2024-07-09 14:44:31 +09:00

SIProgramInfo.cpp

[AMDGPU] MCExpr-ify AMDGPU HSAMetadata (#94788)

2024-06-26 16:39:08 +01:00

SIProgramInfo.h

[AMDGPU] MCExpr-ify AMDGPU HSAMetadata (#94788)

2024-06-26 16:39:08 +01:00

SIRegisterInfo.cpp

AMDGPU: Handle folding frame indexes into s_add_i32 (#101694)

2024-08-22 09:16:37 +04:00

SIRegisterInfo.h

[CodeGen] Make some includes explicit (NFC)

2023-11-24 14:43:18 +01:00

SIRegisterInfo.td

[AMDGPU][NFC] Rename isHi() to isHi16Reg() for clarity. (#103888)

2024-08-14 17:04:15 +01:00

SISchedule.td

AMDGPU: Correct cycle counts for f64 mfma on gfx940 (#83782)

2024-03-06 09:36:01 +05:30

SIShrinkInstructions.cpp

[AMDGPU][True16][CodeGen] Support AND/OR/XOR and LDEXP True16 format (#102620)

2024-08-13 12:23:39 -04:00

SIWholeQuadMode.cpp

[AMDGPU] SIWholeQuadMode: avoid execz effects in exact regions (#101157)

2024-08-01 18:46:36 +09:00

SMInstructions.td

[AMDGPU][AsmParser][NFCI] All NamedIntOperands to be of the i32 type. (#102616)

2024-08-09 15:20:48 +01:00

SOPInstructions.td

AMDGPU: Support VALU add instructions in localstackalloc (#101692)

2024-08-08 23:22:48 +04:00

VIInstrFormats.td

…

VINTERPInstructions.td

[AMDGPU][NFC] Rename the clamp modifier definition to follow the prevailing convention. (#94353)

2024-06-04 16:31:27 +01:00

VOP1Instructions.td

[AMDGPU][True16][CodeGen] Support AND/OR/XOR and LDEXP True16 format (#102620)

2024-08-13 12:23:39 -04:00

VOP2Instructions.td

[AMDGPU][True16][CodeGen] Support AND/OR/XOR and LDEXP True16 format (#102620)

2024-08-13 12:23:39 -04:00

VOP3Instructions.td

[AMDGPU][True16][CodeGen] Support AND/OR/XOR and LDEXP True16 format (#102620)

2024-08-13 12:23:39 -04:00

VOP3PInstructions.td

AMDGPU: Use defvar instead of defaulted multiclass argument

2024-07-12 21:15:12 +04:00

VOPCInstructions.td

[AMDGPU][True16][MC] added VOPC realtrue/faketrue flag and fake16 instructions (#104739)

2024-08-21 17:47:36 +03:00

VOPDInstructions.td

[AMDGPU] Update VOP instructions for GFX12 (#74853)

2023-12-12 11:38:24 +01:00

VOPInstructions.td

[LLVM][AMDGPU] Copy isConvergent from Pseudo to Real instructions (#99658)

2024-07-25 18:01:07 +02:00