shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions 6 Packages Projects Releases Wiki Activity
llvm-project/llvm/test/Transforms
History
Alexey Bataev 94ec7ffa46
[SLP] Do not skip tiny trees with gathered loads to vectorize 
The isTreeTinyAndNotFullyVectorizable check for 2-node trees
(insertelement root + gather child) was too aggressive: it rejected
trees even when LoadEntriesToVectorize was non-empty, preventing
gathered loads from being vectorized into masked loads/strided loads, etc.

Reviewers: RKSimon, hiraditya

Pull Request: https://github.com/llvm/llvm-project/pull/190040
2026-04-02 06:47:53 -04:00
..
ADCE
…
AddDiscriminators
[CSSPGO] Preserve pseudo-probe discriminators in cloneWithBaseDiscriminator (#178267)
2026-01-28 10:24:56 -08:00
AggressiveInstCombine
[AggressiveInstCombine] Recognize table based log2 and replace with ctlz+sub. (#185160)
2026-03-14 17:38:56 -07:00
AlignmentFromAssumptions
…
ArgumentPromotion
[ArgPromotion] Add DW_CC_nocall to DISubprogram (#178973)
2026-02-03 13:55:29 -08:00
AtomicExpand
[atomicrmw] fminimumnum/fmaximumnum support (#187030)
2026-03-18 09:35:49 +01:00
Attributor
ValueTracking: llvm.amdgcn.fract cannot introduce overflow (#189002)
2026-04-01 21:07:57 +00:00
BDCE
…
BlockExtractor
…
BranchFolding
…
CalledValuePropagation
…
CallSiteSplitting
…
CanonicalizeAliases
…
CanonicalizeFreezeInLoops
…
CodeExtractor
[DebugInfo] Add Verifier check for incorrectly-scoped retainedNodes (#166855)
2025-11-10 13:13:49 +01:00
CodeGenPrepare
[CodeGenPrepare][NFC] Reland: Update the dominator tree instead of rebuilding it (#179040)
2026-03-31 09:01:11 +08:00
ConstantHoisting
…
ConstantMerge
…
ConstraintElimination
[ConstraintElim] Infer linear constraints from udiv and urem (#180689)
2026-02-10 22:25:13 +08:00
Coroutines
Reland "[CoroSplit] Erase trivially dead allocas after spilling (#189295)" (#190124)
2026-04-02 07:45:13 +00:00
CorrelatedValuePropagation
[CVP] Stop CVP constant propagation from destroying llvm.assume (#183688)
2026-03-02 14:55:43 +00:00
CrossDSOCFI
…
DCE
…
DeadArgElim
…
DeadStoreElimination
[DSE] Handle provenance when eliminating tautological assignments
2026-03-04 10:35:52 +01:00
DFAJumpThreading
[IR] Reland Optimize PHINode::removeIncomingValue() and PHINode::removeIncomingValueIf() to use the swapping strategy. (#174274)
2026-01-04 09:24:53 +08:00
DivRemPairs
…
DropUnnecessaryAssumes
[DropAssumes] Drop dereferenceable assumptions after vectorization. (#166947)
2025-11-10 10:50:15 +00:00
DwarfEHPrepare
DwarfEHPrepare: Use LibcallLoweringInfo analysis (#176380)
2026-01-18 07:12:58 +01:00
DXILUpgrade
…
EarlyCSE
[LLVM] Refine MemoryEffect handling for target-specific intrinsics (#155590)
2026-02-25 13:53:44 +00:00
EliminateAvailableExternally
…
EmbedBitcode
[llvm][LLD][COFF] Add fat-lto-object support for COFF targets (#165529)
2025-12-18 22:53:25 +02:00
EntryExitInstrumenter
[CHERI] Allow @llvm.returnaddress to return a pointer in any address space. (#188464)
2026-03-25 13:59:38 +00:00
ExpandIRInsts
[CodeGen] Expand power-of-2 div/rem at IR level in ExpandIRInsts. (#180654)
2026-02-25 22:54:34 +01:00
ExpandMemCmp
Move ExpandMemCmp and MergeIcmp to the middle end (#77370)
2026-04-02 09:57:00 +02:00
ExpandVariadics
[profcheck] Propagate profile metadata to Wrapper function in optimize mode of ExpandVariadic. (#168161)
2025-11-21 11:17:28 -08:00
FatLTOCleanup
…
FixIrreducible
[AMDGPU][FixIrreducible][UnifyLoopExits] Support callbr with inline-asm (#149308)
2025-10-30 14:22:42 +01:00
Float2Int
[Float2Int] Make sure the CFP can be represented in the integer type (#167699)
2025-11-13 20:54:32 +08:00
ForcedFunctionAttrs
[ForceFunctionAttrs] Fix handling of conflicts for more attributes (#186304)
2026-03-17 02:10:16 +08:00
FunctionAttrs
[LLVM] Improve IR parsing and printing for target memory locations (#176968)
2026-03-19 17:29:54 +00:00
FunctionImport
Reland "[DebugMetadata][DwarfDebug] Support function-local types in lexical block scopes (4/7)" (#165032)
2026-02-04 00:34:52 +01:00
FunctionSpecialization
[Passes][FuncSpec] Move optsize/minsize handling into pass (#189952)
2026-04-02 09:32:39 +02:00
GCOVProfiling
…
GlobalDCE
…
GlobalMerge
GlobalMerge: Do not merge globals with non-dbg metadata.
2026-01-16 11:21:50 -08:00
GlobalOpt
[LLVM] Improve IR parsing and printing for target memory locations (#176968)
2026-03-19 17:29:54 +00:00
GlobalSplit
…
GuardWidening
…
GVN
[HLSL] Implement Texture2D::operator[] (#186110)
2026-03-23 15:51:22 +00:00
GVNHoist
IR: Add !nofpclass metadata (#177140)
2026-01-22 20:49:34 +01:00
GVNSink
…
HardwareLoops
…
HelloNew
…
HipStdPar
…
HotColdSplit
[DebugInfo] Add Verifier check for incorrectly-scoped retainedNodes (#166855)
2025-11-10 13:13:49 +01:00
IndirectBrExpand
…
IndVarSimplify
[NFC] [IndVars] test for multiple bounds for predicate-loop-traps
2026-02-13 10:13:31 -08:00
InferAddressSpaces
[InferAS] Infer the address space of inttoptr (#173244)
2026-02-20 09:49:00 +08:00
InferAlignment
[InferAlign] Eliminate trunc ptr to log2(align) pattern (#176562)
2026-01-17 17:00:07 +01:00
InferFunctionAttrs
[LLVM] Improve IR parsing and printing for target memory locations (#176968)
2026-03-19 17:29:54 +00:00
Inline
[LLVM] Add flatten function attribute to LLVM IR and implement recursive inlining in AlwaysInliner (#174899)
2026-03-19 11:25:46 -07:00
InstCombine
[SimplifyLibCalls] Prevent orphaned global string literals (#189502)
2026-03-31 12:12:15 -07:00
InstNamer
…
InstSimplify
[LLVM][ConstantFold] Ensure scalar type is queried when determining denormal mode. (#189600)
2026-03-31 14:11:23 +01:00
InterleavedAccess
…
Internalize
…
IRCE
[IRCE] Prove predicates with and without loop-guarded SCEVs (#174843)
2026-01-08 15:16:55 +01:00
IRNormalizer
[IR] Split Br into UncondBr and CondBr (#184027)
2026-03-11 12:31:10 +00:00
IROutliner
[LLVM] Remove "no-nans-fp-math" attribute support (#186285)
2026-03-13 09:29:28 +00:00
JumpTableToSwitch
[JTS] Add statistics (#183431)
2026-02-26 06:12:43 -08:00
JumpThreading
[JT] tryToUnfoldSelectInCurrBB should update BFI & BPI if present (#188097)
2026-03-24 14:12:31 -07:00
KCFI
…
LCSSA
[DebugInfo] Add Verifier check for incorrectly-scoped retainedNodes (#166855)
2025-11-10 13:13:49 +01:00
LICM
[LICM] Reassociate add/sub expressions to hoist invariant computations (#183082)
2026-04-01 18:29:40 +01:00
LoadStoreVectorizer
[LoadStoreVectorizer] Add tests for mixed-type vectorization (#189716)
2026-03-31 17:51:42 +00:00
LoopBoundSplit
…
LoopDataPrefetch
…
LoopDeletion
[LoopInfo] Preserve profile information in makeLoopInvariant (#174171)
2026-02-27 10:19:03 -08:00
LoopDistribute
[NFC][LoopDistribute] Add test for doubly nested loop (#179263)
2026-02-05 13:32:12 +05:30
LoopFlatten
…
LoopFusion
[LoopFusion] Remove the InvalidDependencies duplicates (#187744)
2026-03-23 16:13:46 -04:00
LoopIdiom
[LoopIdiomVectorize] Preserve address space in FindFirstByte (#185226)
2026-03-13 22:57:07 +00:00
LoopInstSimplify
…
LoopInterchange
[DA] Test AddRecs are nsw before strong SIV test (#183421)
2026-03-09 12:39:37 -04:00
LoopLoadElim
…
LoopPredication
…
LoopRotate
[Passes][LoopRotate] Move minsize handling fully into pass (#189956)
2026-04-02 09:32:56 +02:00
LoopSimplify
[InstCombine] Canonicalize GEP source element types (#180745)
2026-03-06 14:48:01 +00:00
LoopSimplifyCFG
[IR] Don't allow successors() over block without terminators (#186646)
2026-03-15 20:34:38 +00:00
LoopStrengthReduce
[AMDGPU] Implement LSR cost model for GFX9+ (#184138)
2026-03-23 12:18:11 +01:00
LoopTransformWarning
…
LoopUnroll
[LoopUnrollPass] Don't pre-set UP.Count before legality checks in computeUnrollCount() (#185979)
2026-03-31 19:50:52 +00:00
LoopUnrollAndJam
[DA] Test AddRecs are nsw before strong SIV test (#183421)
2026-03-09 12:39:37 -04:00
LoopVectorize
[VPlan] Avoid unnecessary BTC SymbolicValue creation (NFC) (#189929)
2026-04-01 16:25:48 +00:00
LoopVersioning
llvm: Convert some assorted lit tests to opaque pointers (#174564)
2026-01-06 11:41:27 +00:00
LoopVersioningLICM
[InstCombine] Canonicalize GEP source element types (#180745)
2026-03-06 14:48:01 +00:00
LowerAllowCheck
[LowerAllowCheck] Add llvm.allow.sanitize.* intrinsics (#172029)
2026-01-03 16:20:43 +01:00
LowerAtomic
…
LowerConstantIntrinsics
[NFCI][IRBuilder] Add CreateAllocationSize helper (#178346)
2026-01-28 22:46:25 -05:00
LowerExpectIntrinsic
…
LowerGlobalDestructors
…
LowerGuardIntrinsic
…
LowerIFunc
…
LowerInvoke
…
LowerMatrixIntrinsics
[Matrix] Handle load/store with different AS in getNonAliasingPointer. (#188721)
2026-03-28 20:52:54 +00:00
LowerSwitch
[IR] Reland Optimize PHINode::removeIncomingValue() and PHINode::removeIncomingValueIf() to use the swapping strategy. (#174274)
2026-01-04 09:24:53 +08:00
LowerTypeTests
[LLVM] Improve IR parsing and printing for target memory locations (#176968)
2026-03-19 17:29:54 +00:00
LowerWidenableCondition
…
MakeGuardsExplicit
…
Mem2Reg
…
MemCpyOpt
Reapply "[MemCpyOpt] support offset slices for performStackMoveOptzn and processMemCpy" (#180002)
2026-02-24 09:33:16 -05:00
MemProfContextDisambiguation
[MemProf] Enhance thin link optimization remarks (#184829)
2026-03-05 23:01:58 -08:00
MergedLoadStoreMotion
…
MergeFunc
[MergeFunc] Do not merge kernel functions (#174254)
2026-01-04 13:35:16 +01:00
MergeICmps
[IR] Reland Optimize PHINode::removeIncomingValue() and PHINode::removeIncomingValueIf() to use the swapping strategy. (#174274)
2026-01-04 09:24:53 +08:00
MetaRenamer
…
MoveAutoInit
[MoveAutoInit] Fix for miscompilation for #150120 (#173961)
2026-01-25 16:21:52 +01:00
NameAnonGlobals
…
NaryReassociate
…
NewGVN
…
ObjCARC
[ObjCARC] Don't sink objc_retain past atomic writes (#184113)
2026-03-30 14:47:31 +01:00
OpenMP
[OpenMP] Only generate call to __kmpc_global_thread_num when needed (#182669)
2026-02-25 06:56:00 -05:00
PartialInlining
…
PartiallyInlineLibCalls
[PILC][profcheck] Bias branch weights when optimizing sqrt (#167742)
2025-11-13 12:46:22 -08:00
PGOProfile
Revert "[MemProf] Dump inline call stacks as optimization remarks (#188678)"
2026-03-30 16:48:38 +02:00
PhaseOrdering
Move ExpandMemCmp and MergeIcmp to the middle end (#77370)
2026-04-02 09:57:00 +02:00
PlaceSafepoints
…
PreISelIntrinsicLowering
[LowerMemIntrinsics][AMDGPU] Optimize memset.pattern lowering (#185901)
2026-03-13 10:37:33 +01:00
Reassociate
[DebugInfo][Reassociate] Use debug records instead of intrinsics in test (#185031)
2026-03-07 00:11:34 +05:30
Reg2Mem
…
RelLookupTableConverter
…
RewriteStatepointsForGC
…
SafeStack
SafeStack: Add missing lit.local.cfg to SPARC subdirectory
2026-02-24 09:06:00 +01:00
SampleProfile
[Psuedoprobe][MachO] Enable pseudo probes emission for MachO (#185758)
2026-03-31 16:27:58 -07:00
SandboxVectorizer
[SandboxVec][LoadStoreVec] Initial pass implementation (#188308)
2026-03-31 15:15:43 -07:00
ScalarizeMaskedMemIntrin
…
Scalarizer
[Scalarizer] Fix out-of-bounds crash (#180359)
2026-02-21 17:55:30 +00:00
SCCP
[LLVM] Improve IR parsing and printing for target memory locations (#176968)
2026-03-19 17:29:54 +00:00
SeparateConstOffsetFromGEP
[LLVM] [SeparateConstOffsetFromGEP] Fix sep-const-offset-from-gep invalid assumption (#183402)
2026-03-11 12:14:32 -07:00
SimpleLoopUnswitch
[InstCombine] Canonicalize GEP source element types (#180745)
2026-03-06 14:48:01 +00:00
SimplifyCFG
[ProfInfo] Fix integer overflow in getDisjunctionWeights (#189079)
2026-03-27 15:19:53 -07:00
SimplifyTypeTests
…
Sink
…
SLPVectorizer
[SLP] Do not skip tiny trees with gathered loads to vectorize
2026-04-02 06:47:53 -04:00
SpeculativeExecution
…
SROA
SROA: Fix tree merge IRBuilder insert point (#189680)
2026-04-01 14:45:05 +01:00
StackProtector
StackProtector: Use LibcallLoweringInfo analysis (#170329)
2026-01-18 07:12:17 +01:00
StraightLineStrengthReduce
Reland "Redesign Straight-Line Strength Reduction (SLSR) (#162930)" (#169614)
2025-12-08 16:07:27 -06:00
StripDeadPrototypes
…
StripSymbols
…
StructurizeCFG
[StructurizeCFG] Fix incorrect zero-cost hoisting in nested control flow (#183792)
2026-03-14 09:26:40 -04:00
TailCallElim
…
ThinLTOBitcodeWriter
[NFC][ThinLTO] Check that refs between split modules have the same GUID (#182982)
2026-02-24 18:16:35 +00:00
TypePromotion
…
UnifyFunctionExitNodes
…
UnifyLoopExits
[llvm][UnifyLoopExits] Avoid optimization if no exit block is found (#165343)
2026-01-26 01:31:28 +01:00
Util
[LIBM][AMDLIBM] - Add new vector call support for fveclib=AMDLIBM (#180896)
2026-03-12 13:24:11 +00:00
VectorCombine
[VectorCombine] Fix transitive Uses in foldShuffleToIdentity (#188989)
2026-04-01 14:53:04 +01:00
WholeProgramDevirt
WholeProgramDevirt: Import/export the CVP byte directly in the summary (#188979)
2026-04-02 09:28:32 +02:00