llvm/llvm-project

Pull Requests Commits

WebAssembly: Use LastSubtargetCache in getSubtargetImpl

arsenm committed 53 minutes ago

457d2417

[ARM] Prevent debug values from causing extra Thumb1 spills (#237017)

yqtian-se committed 1 hour ago

Verified 3ad0e40f

[LoongArch] Avoid left shift of negative immediates in foldLargeOffset() (#242761)

MaskRay committed 1 hour ago

Verified 22ce9d7d

RISCV: Migrate to LastSubtargetCache in getSubtargetImpl (#242559)

arsenm committed 2 hours ago

Verified f3681f20

[libc][mathvec] Cover the non-FMA expf rounding correction (#241112)

sriramshastry committed 2 hours ago

Verified cbfac21c

[mlir][gpu] Let gpu::TargetOptions carry LLVM library options (#242734)

MaskRay committed 3 hours ago

Verified 0d99dea8

[clang] Preserve ext_vector_type locations during template substitution (#229328)

ChenMiaoi committed 4 hours ago

Verified 7075e15f

[libc] Simplify execve syscall error handling (#242601)

c8ef committed 5 hours ago

Verified 64be9daf

[libc] Simplify getrandom syscall error handling (#242602)

c8ef committed 5 hours ago

Verified b75c8062

[Mips] Use a proper boolean for the select condition in lowerShift*Parts (#241685)

brad0 committed 6 hours ago

Verified 0212d77f

[GlobalISel] Ignore debug uses in direct single-use combines (#237643)

yqtian-se committed 6 hours ago

Verified 4ab698df

[gn build] Port f2368ada4905 (#242724)

nico committed 7 hours ago

Verified 53e35685

[gn build] Port d2c9cc6e795c (#242723)

nico committed 7 hours ago

Verified 9666ecc1

[SPIR-V] Legalize G_PHI of a vector of pointers (#241166)

gonsolo committed 7 hours ago

Verified e5009d8a

[GVN] Fold equivalent `bitextract` and `bitinsert` (#242716)

pedroclobo committed 7 hours ago

Verified 7df78681

[VPlan] Remove dead hasScalarVFOnly check in handleFindLastRed (NFC). (#242714)

fhahn committed 7 hours ago

Verified f247014d

[VPlan] Keep NaN exit when using active-lane-mask for control flow. (#241543)

fhahn committed 8 hours ago

Verified f8bac503

[LV] Add more tests for FP min/max select rdx without fast-math (NFC) (#242708)

fhahn committed 8 hours ago

Verified b345f6ca

[ConstantFolding] Fold loads from `ConstantByte` initializers (#229137)

pedroclobo committed 8 hours ago

Verified 2b0ae1a5

[Support] Reject library options in cl::ParseCommandLineOptions (#242689)

MaskRay committed 9 hours ago

Verified 77949b7c

[VPlan] Don't skip the cost of replicating recipes at scalable VFs. (#242701)

fhahn committed 9 hours ago

Verified 92c12a31

[VPlan] Use store's mask for histogram recipe (NFC). (#242702)

fhahn committed 9 hours ago

Verified 3eab741c

[mlir-translate] Parse LLVM library options into context (#242707)

MaskRay committed 9 hours ago

Verified 8377e5c1

[DAGCombiner] Don't fold fmul/fdiv by a power of 2 that converts to inf (#242493)

Joel-Wwalker committed 10 hours ago

Verified 852cf83e

[Passes] Remove -enable-loopinterchange (#238899)

MaskRay committed 10 hours ago

Verified a10c1cac

[APInt] Add a single-word fast path to smul_ov (#242542)

cjsrxzdyzds committed 10 hours ago

Verified a054f4fb

AMDGPU: Copy 16-bit arguments in 32-bit registers as i32 and truncate (#242681)

arsenm committed 10 hours ago

Verified d682f147

DAG: Allow narrowing uint_to_fp source to the zext source type (#242668)

arsenm committed 10 hours ago

Verified 241c94e3

AMDGPU: Add more uitofp of zext from i32 tests (#242667)

arsenm committed 10 hours ago

Verified 5143d1e8

[Offload] Handle the unchecked error properly (#242521)

Kewen12 committed 10 hours ago

Verified 89a999fe

Older