llvm-project

History

[NVPTX] Fix NaN + overflow semantics of f2ll/d2i (#159530 )

Fix the NaN-handling semantics of various NVVM intrinsics converting
from fp types to integer types.

Previously in ConstantFolding, NaN inputs would be constant-folded to 0.
However, v9.0 of the PTX spec states that:

In float-to-integer conversions, depending upon conversion types, NaN
input results in following value:
 * Zero if source is not `.f64` and destination is not `.s64`, .`u64`.
* Otherwise `1 << (BitWidth(dst) - 1)` corresponding to the value of
`(MAXINT >> 1) + 1` for unsigned type or `MININT` for signed type.

Also, support for constant-folding +/-Inf and values which
overflow/underflow the integer output type has been added (they clamp to
min/max int).

Because of this NaN-handling semantic difference, we also need to
disable transforming several intrinsics to FPToSI/FPToUI, as the LLVM
intstruction will return poison, but the intrinsics have defined
behaviour for these edge-cases like NaN/Inf/overflow.

2025-09-25 16:19:11 +01:00

ConstProp

[ConstantFolding] Avoid use of isNonIntegralPointerType()

2025-09-23 15:29:33 -07:00

X86

…

2010-12-20-Boolean.ll

…

2011-01-14-Thread.ll

…

2011-02-01-Vector.ll

…

2011-09-05-InsertExtractValue.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

2011-10-27-BinOpCrash.ll

[IR] Remove support for lshr/ashr constant expressions (#71955 )

2023-11-14 09:25:14 +01:00

2011-11-23-MaskedBitsCrash.ll

…

2013-04-19-ConstantFoldingCrash.ll

…

abs_intrinsic.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

add_vp.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

add-mask.ll

…

add.ll

…

addsub.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

and-icmps-same-ops.ll

…

and-or-icmp-ctpop.ll

Reapply [InstCombine] Simplify and/or of icmp eq with op replacement (#70335 )

2023-11-03 10:16:15 +01:00

and-or-icmp-min-max.ll

Reapply [InstCombine] Simplify and/or of icmp eq with op replacement (#70335 )

2023-11-03 10:16:15 +01:00

and-or-icmp-nullptr.ll

…

and-or-icmp-zero.ll

…

and-or-implied-cond.ll

[InstSimplify] Fix simplifyAndOrWithICmpEq with undef refinement (#98898 )

2024-07-16 11:40:04 +02:00

and.ll

…

AndOrXor.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

assume_icmp.ll

…

assume-fcmp-constant-implies-class.ll

Reapply "ValueTracking: Identify implied fp classes by general fcmp (#66505 )"

2024-01-25 13:38:23 +05:30

assume-non-zero.ll

[ValueTracking] Precommit test for #118406

2025-02-04 22:53:33 +01:00

bitcast-vector-fold.ll

[LLVM][IR] Add support for vector ConstantInt/FP to ConstandFolding:FoldBitCast. (#117163 )

2024-12-10 15:42:05 +00:00

bitreverse-fold.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

bitreverse.ll

…

bswap.ll

…

call.ll

[InstCombine] Propagate poison pow[i], [us]add, [us]sub and [us]mul (#146750 )

2025-07-04 22:55:07 +01:00

canonicalize.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

cast-unsigned-icmp-cmp-0.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

cast.ll

…

cmp_ext.ll

…

cmp_of_min_max.ll

…

cmp-alloca-offsets.ll

[InstSimplify] Correctly handle comparison with zero-size allocs (#115728 )

2024-11-14 11:55:19 +01:00

cmp-vec-fast-path.ll

[LLVM][IR] Use splat syntax when printing ConstantExpr based splats. (#116856 )

2024-11-21 11:21:12 +00:00

compare.ll

[ConstantFold] Constant fold icmp of boolean scalable vectors (#142528 )

2025-06-03 19:51:21 +08:00

const-fold-nvvm-add.ll

[NVPTX] Constant fold NVVM add/mul/div/fma (#152544 )

2025-08-22 17:33:00 +01:00

const-fold-nvvm-div.ll

[NVPTX] Constant fold NVVM add/mul/div/fma (#152544 )

2025-08-22 17:33:00 +01:00

const-fold-nvvm-f2i-d2i.ll

[NVPTX] Fix NaN + overflow semantics of f2ll/d2i (#159530 )

2025-09-25 16:19:11 +01:00

const-fold-nvvm-f2ll-d2ll.ll

[NVPTX] Fix NaN + overflow semantics of f2ll/d2i (#159530 )

2025-09-25 16:19:11 +01:00

const-fold-nvvm-fma.ll

[NVPTX] Constant fold NVVM add/mul/div/fma (#152544 )

2025-08-22 17:33:00 +01:00

const-fold-nvvm-fmin-fmax.ll

[NVPTX] Constant fold NVVM fmin and fmax (#121966 )

2025-01-16 14:38:51 +00:00

const-fold-nvvm-mul.ll

[NVPTX] Constant fold NVVM add/mul/div/fma (#152544 )

2025-08-22 17:33:00 +01:00

const-fold-nvvm-unary-arithmetic.ll

[ConstantFolding] Fix incorrect nvvm_round folding (#151563 )

2025-08-01 10:31:43 +01:00

constant-fold-fp-denormal.ll

[llvm-project] Fix typo "seperate" (#95373 )

2024-06-13 20:20:27 +01:00

constant-fold-inttoptr-add.ll

[Value] Look through inttoptr (add ..) in accumulateConstantOffsets (#124981 )

2025-01-30 20:58:38 +00:00

constantfold-add-nuw-allones-to-allones.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

constantfold-shl-nuw-C-to-C.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

constfold-constrained.ll

…

ctpop-pow2.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

dead-code-removal.ll

[llvm] Remove br i1 undef from some regression tests [NFC] (#116161 )

2024-11-15 14:19:50 +00:00

disable_folding.ll

[ConstantFolding] Add flag to disable call folding (#140270 )

2025-05-30 11:27:18 +01:00

distribute.ll

…

div-by-0-guard-before-smul_ov-not.ll

Reapply [InstCombine] Simplify and/or of icmp eq with op replacement (#70335 )

2023-11-03 10:16:15 +01:00

div-by-0-guard-before-smul_ov.ll

Reapply [InstCombine] Simplify and/or of icmp eq with op replacement (#70335 )

2023-11-03 10:16:15 +01:00

div-by-0-guard-before-umul_ov-not.ll

Reapply [InstCombine] Simplify and/or of icmp eq with op replacement (#70335 )

2023-11-03 10:16:15 +01:00

div-by-0-guard-before-umul_ov.ll

Reapply [InstCombine] Simplify and/or of icmp eq with op replacement (#70335 )

2023-11-03 10:16:15 +01:00

div.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

domcondition.ll

…

exact-nsw-nuw.ll

…

exp10.ll

[InstSimplify] Add poison propagation for trivially vectorizable intrinsics (#149243 )

2025-07-19 19:37:21 -07:00

extract-element.ll

[LLVM][IR] Teach extractelement folds about constant ConstantInt/FP. (#116793 )

2024-11-21 12:39:53 +00:00

extract-vector.ll

…

fast-math-strictfp.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

fast-math.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

fcmp-select.ll

…

fcmp.ll

[ValueTracking] Bail out on x86_fp80 when computing fpclass with knownbits (#130477 )

2025-03-09 21:10:35 +08:00

fdiv-strictfp.ll

…

fdiv.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

floating-point-arithmetic-strictfp.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

floating-point-arithmetic.ll

[IR] Handle fabs LHS in fcmpImpliesClass (#152913 )

2025-08-10 22:55:04 +08:00

floating-point-compare.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

fminmax-folds.ll

[NFC][InstSimplify] Refactor fminmax-folds.ll test (#160504 )

2025-09-25 11:11:21 +01:00

fold-intrinsics.ll

[InstSimplify] Simplify extractvalue (umul_with_overflow(x, 1)). (#157307 )

2025-09-07 18:32:40 +01:00

fold-vscale.ll

…

fp-nan.ll

[LLVM][IR] Use splat syntax when printing ConstantExpr based splats. (#116856 )

2024-11-21 11:21:12 +00:00

fp-undef-poison-strictfp.ll

…

fp-undef-poison.ll

[ConstantFolding] Fold sqrt poison -> poison (#141821 )

2025-06-11 13:27:14 +02:00

fptoi-range.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

fptoi-sat.ll

…

freeze-noundef.ll

[IR] Support or disjoint in hasPoisonGeneratingFlags()

2023-11-30 17:26:23 +01:00

freeze.ll

…

gc_relocate.ll

[InstSimplify] Add trivial simplifications for gc.relocate intrinsic (#81639 )

2024-02-14 02:16:32 +03:00

gep.ll

Revert "[InstSimplify] Fold getelementptr inbounds null, idx -> null (#130742 )" (#138168 )

2025-05-01 14:21:59 -07:00

get_active_lane_mask.ll

[InstSimplify] Consider vscale_range for get active lane mask (#160073 )

2025-09-24 11:35:15 +01:00

gv-alloca-cmp.ll

[IR] Don't allow values of opaque type (#137625 )

2025-04-30 15:01:00 +02:00

icmp-abs-nabs.ll

…

icmp-bool-constant.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

icmp-constant.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

icmp-monotonic.ll

[InstSimplify] Fold X * C >= X to true (#129352 )

2025-03-01 12:02:57 -05:00

icmp-not-bool-constant.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

icmp-ranges.ll

…

icmp.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

implies.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

insert-vector.ll

…

insertelement.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

insertvalue.ll

…

invalid-load-operand-infinite-loop.ll

…

invariant.group-load.ll

…

is_fpclass.ll

…

known-never-infinity.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

known-never-nan.ll

[InstSimplify] fold uno/ord comparison if fpclass is always NaN (#97763 )

2024-07-09 14:44:48 -07:00

known-non-zero-opaque-ptrs.ll

[ValueTracking] Handle nonnull attributes at callsite (#124908 )

2025-01-29 23:14:36 +08:00

known-non-zero.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

ldexp.ll

[PatternMatch] Do not accept undef elements in m_AllOnes() and friends (#88217 )

2024-04-17 18:22:05 +09:00

load-relative-32.ll

…

load-relative.ll

…

load.ll

…

log2-pow2-intrinsic.ll

…

log10-pow10-intrinsic.ll

…

log-exp-intrinsic.ll

…

logic-of-fcmps.ll

[InstSimplify] Generalize simplifyAndOrOfFCmps to handle fabs (#116590 )

2024-11-19 20:10:40 +08:00

maxmin_intrinsics.ll

[InstSimplify] Fold {u,s}{min,max} x, poison -> poison (#138166 )

2025-05-02 07:49:27 +08:00

maxmin.ll

…

mul.ll

[PatternMatch] Do not accept undef elements in m_AllOnes() and friends (#88217 )

2024-04-17 18:22:05 +09:00

named-vector-shuffle-reverse.ll

Move several vector intrinsics out of experimental namespace (#88748 )

2024-04-29 10:16:45 +01:00

negate.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

noalias-ptr.ll

…

null-ptr-is-valid-attribute.ll

…

null-ptr-is-valid.ll

…

opaque-ptr.ll

…

or-icmps-same-ops.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

or.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

past-the-end.ll

[ConstantFolding] Infer getelementptr nuw flag (#119214 )

2024-12-09 16:44:05 +01:00

phi-cse.ll

…

phi.ll

[InstSimplify] Fix incorrect poison propagation when folding phi (#96631 )

2024-11-07 14:09:45 +01:00

po2-shift-add-and-to-zero.ll

[InstSimplify] Fold converted urem to 0 if there's no overlapping bits (#71528 )

2023-11-20 10:27:16 +00:00

pr28725.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

pr33957.ll

…

pr49495.ll

…

pr87042.ll

[InstSimplify] Make sure the simplified value doesn't generate poison in threadBinOpOverSelect (#87075 )

2024-04-11 12:48:52 +08:00

pr122582.ll

[ConstantFolding] Add ilogb in isMathLibCallNoop (#122582 )

2025-01-20 12:19:07 -05:00

ptr_diff.ll

…

ptrmask.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

ptrtoint.ll

[InstSimplify] Fold ptrtoint(ptradd(P,X-ptrtoint(P))) to X (#98649 )

2024-07-15 09:26:03 +02:00

reassociate.ll

…

redundant-null-check-in-uadd_with_overflow-of-nonnull-ptr.ll

…

rem.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

require-dominator.ll

[llvm] Remove br i1 undef from some regression tests [NFC] (#116161 )

2024-11-15 14:19:50 +00:00

result-of-add-of-negative-is-non-zero-and-no-underflow.ll

Reapply [InstCombine] Simplify and/or of icmp eq with op replacement (#70335 )

2023-11-03 10:16:15 +01:00

result-of-usub-by-nonzero-is-non-zero-and-no-overflow.ll

…

result-of-usub-is-non-zero-and-no-overflow.ll

…

returned.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

round-intrinsics.ll

…

saturating-add-sub.ll

[InstCombine] Propagate poison pow[i], [us]add, [us]sub and [us]mul (#146750 )

2025-07-04 22:55:07 +01:00

sdiv.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

select_or_and.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

select-abs.ll

ValueTracking: handle more ops in isNotCrossLaneOperation (#112183 )

2024-10-14 14:08:12 +01:00

select-equivalence-fp.ll

[InstSimplify] Treat float binop with identity as refining (#120098 )

2024-12-17 10:58:52 +01:00

select-icmp.ll

[InstSimplify] Add additional checks when substituting pointers (#125385 )

2025-02-02 19:04:23 +08:00

select-implied.ll

…

select-inseltpoison.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

select-logical.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

select-maxmin.ll

…

select.ll

[InstSimplify] Handle trunc to i1 in Select with bit test folds. (#122944 )

2025-02-01 08:48:32 +01:00

shift-128-kb.ll

…

shift-knownbits.ll

[Reland][ValueTracking] Improve Bitcast handling to match SDAG (#145223 )

2025-08-04 14:51:03 +05:30

shift.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

shr-nop.ll

[ValueTracking] Handle assume(trunc x to i1) in ComputeKnownBits (#118406 )

2025-04-10 19:29:28 +02:00

shr-scalar-vector-consistency.ll

…

shufflevector-inseltpoison.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

shufflevector.ll

[ConstantFolding] Fold scalable shufflevector of poison/undef. (#143475 )

2025-06-10 12:50:16 -07:00

signed-div-rem.ll

…

simplify-nested-bitcast.ll

[ConstantFolding] Infer getelementptr nuw flag (#119214 )

2024-12-09 16:44:05 +01:00

sincos.ll

[InstSimplify] Fold trig functions call of poison to poison (#148969 )

2025-07-16 08:35:13 -07:00

smul_fix.ll

…

srem.ll

[PatternMatch] Do not accept undef elements in m_AllOnes() and friends (#88217 )

2024-04-17 18:22:05 +09:00

strictfp-fadd.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

strictfp-fsub.ll

…

strictfp-sqrt-nonneg.ll

…

sub.ll

[PatternMatch] Do not accept undef elements in m_AllOnes() and friends (#88217 )

2024-04-17 18:22:05 +09:00

subnuw-with-xor.ll

[InstSimpify] Simplifying (xor (sub C_Mask, X), C_Mask) -> X (#122552 )

2025-01-11 15:10:42 -06:00

umax-1.ll

[InstSimplify] Simplify 'x u>= 1' to true when x is known non-zero (#145204 )

2025-06-22 13:32:19 +08:00

undef.ll

…

uscmp.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

vec-cmp.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

vec-icmp-of-cast.ll

Reapply "[InstCombine] Match scalable splats in m_ImmConstant (#132522 )" (#134262 )

2025-04-03 18:03:16 +01:00

vector_gep.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00

vector_ptr_bitcast.ll

…

vector-inseltextelt.ll

…

vp-reverse.ll

[InstSimplify] Add basic simplifications for vp.reverse (#144112 )

2025-06-16 10:07:56 -07:00

vscale-inseltpoison.ll

[ConstantFold] Support scalable constant splats in ConstantFoldCastInstruction (#133207 )

2025-04-03 16:24:56 +01:00

vscale.ll

[ConstantFold] Support scalable constant splats in ConstantFoldCastInstruction (#133207 )

2025-04-03 16:24:56 +01:00

xor.ll

[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548 )

2024-11-06 11:53:33 +00:00