Simon Pilgrim 89ca3e72ca [CostModel][X86] Reduce worst case v8i16/v16i8 SSE2 shuffle costs (#124789) ...

These were based off instruction count, not throughput - we can probably improve these further, but these throughput numbers match the worse expanded shuffles we see in the vector-shuffle-128-v* codegen tests.

2025-01-29 10:23:09 +00:00

..

concat-boolmasks.ll

[VectorCombine] eraseInstruction - ensure we reattempt to fold other users of an erased instruction's operands (REAPPLIED)

2025-01-02 18:19:02 +00:00

extract-binop-inseltpoison.ll

[CostModel][X86] getShuffleCost - shuffles with only one defined element are always cheap (#124412)

2025-01-27 15:56:22 +00:00

extract-binop.ll

[CostModel][X86] getShuffleCost - shuffles with only one defined element are always cheap (#124412)

2025-01-27 15:56:22 +00:00

extract-cmp-binop.ll

[CostModel][X86] Reduce worst case v8i16/v16i8 SSE2 shuffle costs (#124789)

2025-01-29 10:23:09 +00:00

extract-cmp.ll

[VectorCombine] Combine extract/insert from vector into a shuffle (#115213)

2024-11-13 11:16:09 +00:00

extract-fneg-insert.ll

[CostModel][X86] getShuffleCost - use processShuffleMasks for all shuffle kinds to legal types (#120599) (#121760)

2025-01-06 13:32:55 +00:00

insert-binop-inseltpoison.ll

…

insert-binop-with-constant-inseltpoison.ll

…

insert-binop-with-constant.ll

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

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

insert-binop.ll

…

lit.local.cfg

…

load-extractelement-scalarization.ll

[VectorCombine] Add scalarizeLoadExtract infinite loop test from #120984 regression

2025-01-02 17:23:12 +00:00

load-inseltpoison.ll

[CostModel][X86] getShuffleCost - shuffles with only one defined element are always cheap (#124412)

2025-01-27 15:56:22 +00:00

load-widening.ll

…

load.ll

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

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

no-sse-inseltpoison.ll

…

no-sse.ll

…

permute-of-binops.ll

[VectorCombine] foldPermuteOfBinops - match identity shuffles only if they match the destination type

2025-01-14 16:09:50 +00:00

pr67803.ll

…

pr114901.ll

[VectorCombine] foldExtractedCmps - (re-)enable fold on non-commutative binops

2024-11-06 12:10:31 +00:00

pr115575.ll

[VectorCombine] scalarizeBinopOrCmp - check for out of bounds element indices

2024-11-09 16:00:03 +00:00

reduction-of-truncations.ll

…

reduction-two-vecs-combine.ll

…

scalarize-cmp-inseltpoison.ll

…

scalarize-cmp.ll

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

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

scalarize-vector-gep.ll

…

select-shuffle.ll

…

shuffle-inseltpoison.ll

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

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

shuffle-of-binops.ll

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

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

shuffle-of-casts.ll

[CostModel][X86] getShuffleCost - use processShuffleMasks to split SK_PermuteTwoSrc shuffles to legal types (#120599)

2024-12-20 10:39:45 +00:00

shuffle-of-cmps.ll

[VectorCombine][X86] shuffle-of-cmps.ll - tweak shuf_fcmp_oeq_v4i32 shuffle to be not so cheap

2025-01-07 11:07:48 +00:00

shuffle-of-intrinsics.ll

[CostModel][X86] Improve cost estimation of insert_subvector shuffle patterns of legalized types (#119363)

2024-12-10 16:28:56 +00:00

shuffle-of-shuffles.ll

[CostModel][X86] getShuffleCost - use processShuffleMasks to split SK_PermuteTwoSrc shuffles to legal types (#120599)

2024-12-20 10:39:45 +00:00

shuffle.ll

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

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