Veera 6f8afafd30

[InstCombine] Fold A == MIN_INT ? B != MIN_INT : A < B to A < B (#120177) ...

This PR folds:
 `A == MIN_INT ? B != MIN_INT : A < B` to `A < B`
 `A == MAX_INT ? B != MAX_INT : A > B` to `A > B`

Proof: https://alive2.llvm.org/ce/z/bR6E2s

This helps in optimizing comparison of optional unsigned non-zero types
in https://github.com/rust-lang/rust/issues/49892.

Rust compiler's current output: https://rust.godbolt.org/z/9fxfq3Gn8

2024-12-19 22:52:55 +08:00

..

CMakeLists.txt

…

InstCombineAddSub.cpp

PatternMatch: migrate to CmpPredicate (#118534)

2024-12-13 14:18:33 +00:00

InstCombineAndOrXor.cpp

[InstCombine] Drop samesign flags in foldLogOpOfMaskedICmps_NotAllZeros_BMask_Mixed (#120373)

2024-12-18 20:40:33 +08:00

InstCombineAtomicRMW.cpp

…

InstCombineCalls.cpp

[PatternMatch] Introduce m_c_Select (#114328)

2024-11-25 13:47:23 +00:00

InstCombineCasts.cpp

[InstCombine] Fold trunc nuw/nsw X to i1 -> true IFF X != 0 (#119131)

2024-12-08 22:28:16 +01:00

InstCombineCompares.cpp

PatternMatch: migrate to CmpPredicate (#118534)

2024-12-13 14:18:33 +00:00

InstCombineInternal.h

[InstCombine] Optimistically allow multiple shufflevector uses in foldOpPhi (#114278)

2024-12-12 17:20:48 -08:00

InstCombineLoadStoreAlloca.cpp

[Passes] Remove -enable-infer-alignment-pass flag (#111873)

2024-10-10 12:28:46 -07:00

InstCombineMulDivRem.cpp

[InstSimplify][InstCombine][ConstantFold] Move vector div/rem by zero fold to InstCombine (#114280)

2024-11-01 22:56:22 +08:00

InstCombineNegator.cpp

[InstCombine] Drop nsw in negation of select (#112893)

2024-11-08 16:20:04 +08:00

InstCombinePHI.cpp

[InstCombine] Intersect nowrap flags between geps while folding into phi

2024-11-13 19:42:06 +01:00

InstCombineSelect.cpp

[InstCombine] Fold A == MIN_INT ? B != MIN_INT : A < B to A < B (#120177)

2024-12-19 22:52:55 +08:00

InstCombineShifts.cpp

[InstCombine] Prevent infinite loop with two shifts (#118806)

2024-12-05 16:57:27 +01:00

InstCombineSimplifyDemanded.cpp

[InstCombine] Fold (X * (Y << K)) u>> K -> X * Y when highbits are not demanded (#111151)

2024-12-03 12:04:04 +08:00

InstCombineVectorOps.cpp

PatternMatch: migrate to CmpPredicate (#118534)

2024-12-13 14:18:33 +00:00

InstructionCombining.cpp

Revert "[InstCombine] Infer nuw for gep inbounds from base of object" (#120460)

2024-12-18 19:06:34 +01:00