Ramkumar Ramachandra 03b70cb1f4

[InstCombine] Restrict foldICmpOfVectorReduce to one-use (#182833) ...

Follow up on 279b3dbe ([InstCombine] Fold icmp (vreduce_(or|and) %x),
(0|-1), #182684) to fix a regression by restricting the fold to one-use.

Regression: https://godbolt.org/z/f38b169MM

2026-02-23 19:31:22 +00:00

..

CMakeLists.txt

…

InstCombineAddSub.cpp

[InstCombine] Preserve fdiv metadata on fneg folds (#179157)

2026-02-03 08:27:18 +08:00

InstCombineAndOrXor.cpp

AMDGPU: Strip sign bit operations on llvm.amdgcn.trig.preop uses (#179712)

2026-02-19 18:44:33 +01:00

InstCombineAtomicRMW.cpp

[NFC][InstCombine] Fix namespace usage in InstCombine (#161902)

2025-10-04 05:28:24 -07:00

InstCombineCalls.cpp

[InstCombine] Transform splat before n x i1 for vec.reduce.add (#182213)

2026-02-21 15:03:25 +00:00

InstCombineCasts.cpp

InstCombine: Fold bitcast of vector with constant to scalar (#179042)

2026-02-19 21:35:31 +01:00

InstCombineCompares.cpp

[InstCombine] Restrict foldICmpOfVectorReduce to one-use (#182833)

2026-02-23 19:31:22 +00:00

InstCombineInternal.h

[InstCombine] Optimise the expression (C && A) || (!C && B) with FoldOrOfLogicalAnds (#178438)

2026-02-16 01:08:30 +05:30

InstCombineLoadStoreAlloca.cpp

[InstCombine] Replace getAllocatedType() with getAllocationSize() (#177435)

2026-01-26 22:31:29 -05:00

InstCombineMulDivRem.cpp

InstCombine: Use SimplifyDemandedFPClass on fmul (#177490)

2026-02-10 09:49:31 +00:00

InstCombineNegator.cpp

[InstCombine] Optimize sub(sext(add(x,y)),sext(add(x,z))). (#144174)

2025-06-19 10:13:58 -07:00

InstCombinePHI.cpp

[InstCombine] Avoid invalid bitcast across address spaces in foldIntegerTypedPHI (#181064)

2026-02-15 15:35:43 -08:00

InstCombineSelect.cpp

InstCombine: Fold absorbing fmul of compared 0 into select (#172381)

2026-02-21 18:42:09 +00:00

InstCombineShifts.cpp

[InstCombine] Fix crash on large integers in visitShl (#176723)

2026-01-19 12:07:03 +01:00

InstCombineSimplifyDemanded.cpp

InstCombine: Implement SimplifyDemandedFPClass for ldexp (#180703)

2026-02-19 10:24:46 +00:00

InstCombineVectorOps.cpp

[InstCombine] Limit canonicalization of extractelement(cast) to constant index or same basic block (#166227)

2026-01-08 22:43:13 +01:00

InstructionCombining.cpp

[InstCombine] Only ignore first zero index during GEP canonicalization (#180764)

2026-02-23 08:55:00 +00:00