Simon Pilgrim 5a7dfb4659 [CostModel][X86] Attempt to match v4f32 shuffles that map to MOVSS/INSERTPS instruction ...

improveShuffleKindFromMask matches this as a SK_InsertSubvector of a v1f32 (which legalises to f32) into a v4f32 base vector, making it easy to recognise. MOVSS is limited to index0.

2025-01-07 11:31:44 +00:00

..

AArch64

[VectorCombine] foldShuffleOfShuffles - fold "shuffle (shuffle x, y, m1), (shuffle y, x, m2)" -> "shuffle x, y, m3" (#120959)

2024-12-23 14:56:15 +00:00

AMDGPU

[VectorCombine] vectorizeLoadInsert - only fold when inserting into a poison vector (#119906)

2024-12-14 11:56:12 +00:00

Hexagon

…

RISCV

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

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

X86

[CostModel][X86] Attempt to match v4f32 shuffles that map to MOVSS/INSERTPS instruction

2025-01-07 11:31:44 +00:00

load-insert-store.ll

…

pr88796.ll

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

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