Manish Kausik H 69192e0193 [LegalizeDAG] Optimize CodeGen for ISD::CTLZ_ZERO_UNDEF (#83039) ...

Previously we had the same instructions being generated for `ISD::CTLZ` and `ISD::CTLZ_ZERO_UNDEF` which did not take advantage of the fact that zero is an invalid input for `ISD::CTLZ_ZERO_UNDEF`. This commit separates codegen for the two cases to allow for the optimization for the latter case.

The details of the optimization are outlined in #82075

Fixes #82075

Co-authored-by: Manish Kausik H <hmamishkausik@gmail.com>

2024-07-08 14:01:32 +01:00

..

CMakeLists.txt

…

DAGCombiner.cpp

[DAGCombiner] Remove unnecessary assert from getShiftAmountTy wrapper. NFC

2024-07-04 19:05:54 -07:00

FastISel.cpp

LLVM_FALLTHROUGH => [[fallthrough]]. NFC

2024-04-25 17:50:59 -07:00

FunctionLoweringInfo.cpp

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

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

InstrEmitter.cpp

[ConstraintFP] Model rounding control registers for inline asm (#92846)

2024-05-23 22:50:39 +08:00

InstrEmitter.h

…

LegalizeDAG.cpp

[LegalizeDAG] Optimize CodeGen for ISD::CTLZ_ZERO_UNDEF (#83039)

2024-07-08 14:01:32 +01:00

LegalizeFloatTypes.cpp

Revert "Intrinsic: introduce minimumnum and maximumnum (#93841)"

2024-06-21 08:34:04 +02:00

LegalizeIntegerTypes.cpp

[LegalizeDAG] Optimize CodeGen for ISD::CTLZ_ZERO_UNDEF (#83039)

2024-07-08 14:01:32 +01:00

LegalizeTypes.cpp

[CodeGen] Use range-based for loops (NFC) (#96777)

2024-06-26 16:49:00 -07:00

LegalizeTypes.h

Revert "Intrinsic: introduce minimumnum and maximumnum (#93841)"

2024-06-21 08:34:04 +02:00

LegalizeTypesGeneric.cpp

…

LegalizeVectorOps.cpp

Revert "Intrinsic: introduce minimumnum and maximumnum (#93841)"

2024-06-21 08:34:04 +02:00

LegalizeVectorTypes.cpp

DAG: Fix assert when legalizing v3f16 ldexp (#97098)

2024-06-30 08:29:20 +02:00

MatchContext.h

[RISCV][VP] Introduce vp saturating addition/subtraction and RISC-V support. (#82370)

2024-02-23 14:17:15 +08:00

ResourcePriorityQueue.cpp

…

ScheduleDAGFast.cpp

[CodeGen] Use range-based for loops (NFC) (#96777)

2024-06-26 16:49:00 -07:00

ScheduleDAGRRList.cpp

[llvm] Move CodeGenTypes library to its own directory (#79444)

2024-01-25 12:01:31 -05:00

ScheduleDAGSDNodes.cpp

[IR] Memory Model Relaxation Annotations (#78569)

2024-04-24 08:52:25 +02:00

ScheduleDAGSDNodes.h

[llvm] Move CodeGenTypes library to its own directory (#79444)

2024-01-25 12:01:31 -05:00

ScheduleDAGVLIW.cpp

…

SDNodeDbgValue.h

…

SelectionDAG.cpp

[SelectionDAG] Let ComputeKnownSignBits handle (shl (ext X), C) (#97695)

2024-07-05 22:37:26 +02:00

SelectionDAGAddressAnalysis.cpp

[CodeGen] Update for scalable MemoryType in MMO (#70452)

2024-03-23 12:56:25 +00:00

SelectionDAGBuilder.cpp

[X86][CodeGen] Convert masked.load/store to CLOAD/CSTORE node only when vector size = 1

2024-07-05 15:50:21 +08:00

SelectionDAGBuilder.h

[IR] Remove support for icmp and fcmp constant expressions (#93038)

2024-06-04 08:31:03 +02:00

SelectionDAGDumper.cpp

[PAC][AArch64] Lower ptrauth constants in code (#96879)

2024-06-28 07:29:38 +03:00

SelectionDAGISel.cpp

[llvm][CodeGen] Avoid 'raw_string_ostream::str' (NFC) (#97318)

2024-07-01 21:52:37 -04:00

SelectionDAGPrinter.cpp

[llvm][CodeGen] Avoid 'raw_string_ostream::str' (NFC) (#97318)

2024-07-01 21:52:37 -04:00

SelectionDAGTargetInfo.cpp

…

StatepointLowering.cpp

[llvm][SelectionDAG] Fix up chains in lowerInvokeable. rdar://113994760 (#94004)

2024-06-13 18:05:52 -07:00

StatepointLowering.h

…

TargetLowering.cpp

[SelectionDAG] Remove LegalTypes argument from getShiftAmountConstant. (#97653)

2024-07-04 18:33:25 -07:00