Search results

LLVM broke frontends relying on documented NaN behavior for minnum/maxnum

floating-pointFloating-point mathFloating-point math

llvm:instcombineCovers the InstCombine, InstSimplify and AggressiveInstCombine passesCovers the InstCombine, InstSimplify and AggressiveInstCombine passes

llvm:transforms

Status: Open.

#170082 In llvm/llvm-project;

· RalfJung opened on Dec 1, 2025

[clang] Make fp contract(on) not overwrite earlier contract(fast) and CLI --ffp-contract=fast(-honor-pragmas)

clangClang issues not falling into any other categoryClang issues not falling into any other category

clang:frontendLanguage frontend issues, e.g. anything involving "Sema"Language frontend issues, e.g. anything involving "Sema"

floating-pointFloating-point mathFloating-point math

Status: Open (in progress).

llvm/llvm-projectnumber 169859

#169859 In llvm/llvm-project;

· mikolaj-pirog opened on Nov 27, 2025

[Clang][Codegen] Rint always lowered to llvm intrinsic

clang:codegenIR generation bugs: mangling, exceptions, etc.IR generation bugs: mangling, exceptions, etc.

floating-pointFloating-point mathFloating-point math

Status: Open.

#168139 In llvm/llvm-project;

· HendrikHuebner opened on Nov 14, 2025

[DAGCombiner] Honor rewrite semantics of fast-math flags in fdiv combine

backend:AArch64

backend:AMDGPU

backend:NVPTX

backend:PowerPC

backend:X86

floating-pointFloating-point mathFloating-point math

llvm:SelectionDAGSelectionDAGISel as wellSelectionDAGISel as well

Status: Open (in progress).

llvm/llvm-projectnumber 167595

#167595 In llvm/llvm-project;

· mikolaj-pirog opened on Nov 11, 2025

Initialization of union not raise expected FPU flag

clang:codegenIR generation bugs: mangling, exceptions, etc.IR generation bugs: mangling, exceptions, etc.

floating-pointFloating-point mathFloating-point math

Status: Open.

#166866 In llvm/llvm-project;

· tydeman opened on Nov 6, 2025

[Flang][AArch64] Missed optimization opportunities with <code>unsafe-math-optimizations</code>

floating-pointFloating-point mathFloating-point math

missed-optimization

Status: Open.

#165809 In llvm/llvm-project;

· yus3710-fj opened on Oct 31, 2025

[Host-dependent][ARM vs x86] ConstantFoldBinaryFP affects DCE consistency

floating-pointFloating-point mathFloating-point math

llvm:transforms

Status: Open.

#165644 In llvm/llvm-project;

· guoxin049 opened on Oct 30, 2025

[X86] Miscompile when using <code>-ftrapping-math</code>

backend:X86

floating-pointFloating-point mathFloating-point math

miscompilation

Status: Open.

#164802 In llvm/llvm-project;

· abhishek-kaushik22 opened on Oct 23, 2025

clang++ miscompiles std::pow with mixed types when using -O3 -fno-math-errno -fveclib=libmvec

floating-pointFloating-point mathFloating-point math

llvm:optimizations

miscompilation

Status: Open.

#164642 In llvm/llvm-project;

· davidbatetc opened on Oct 22, 2025

[DAGCombiner] Relax nsz constraint with fp-&gt;int-&gt;fp optimizations

backend:AArch64

backend:AMDGPU

backend:X86

floating-pointFloating-point mathFloating-point math

llvm:SelectionDAGSelectionDAGISel as wellSelectionDAGISel as well

Status: Open (in progress).

llvm/llvm-projectnumber 164503

#164503 In llvm/llvm-project;

· guy-david opened on Oct 21, 2025

Mathematical operations on constant zero are not folded away

floating-pointFloating-point mathFloating-point math

llvm:optimizations

missed-optimization

Status: Open.

#164023 In llvm/llvm-project;

· mqudsi opened on Oct 17, 2025

bool bitfield += complex gets infinite loop

clang:frontendLanguage frontend issues, e.g. anything involving "Sema"Language frontend issues, e.g. anything involving "Sema"

floating-pointFloating-point mathFloating-point math

Status: Open.

#163140 In llvm/llvm-project;

· tydeman opened on Oct 13, 2025