Simon Pilgrim 94df95de6b [TTI][X86] getShuffleCosts - for SK_PermuteTwoSrc, if the masks are known to be "inlane" no need to scale the costs by worst-case legalization (#117999) ...

SK_PermuteTwoSrc legalization has to assume any of the legalised source registers could be referenced in split shuffles, but if we already know that each 128-bit lane only references elements from the same lane of the source operands, then this scaling won't occur.

Hopefully this can help with #113356 without us having to get full processShuffleMasks canonicalization finished first.

2024-12-01 12:01:47 +00:00

..

addsub-inseltpoison.ll

…

addsub.ll

…

blendv-select.ll

…

ctlz-loop.ll

…

earlycse-after-simplifycfg-two-entry-phi-node-folding.ll

…

excessive-unrolling.ll

…

hoist-load-of-baseptr.ll

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

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

horiz-math-inseltpoison.ll

[CostModel][X86] Add initial costs for non-lane-crossing one/two input shuffles (#114680)

2024-11-04 10:19:02 +00:00

horiz-math.ll

[CostModel][X86] Add initial costs for non-lane-crossing one/two input shuffles (#114680)

2024-11-04 10:19:02 +00:00

lit.local.cfg

…

loadcombine.ll

…

loop-idiom-vs-indvars.ll

…

loop-vectorizer-noalias.ll

…

masked-memory-ops-with-cf.ll

…

masked-memory-ops.ll

…

merge-functions2.ll

…

merge-functions3.ll

…

merge-functions.ll

…

nancvt.ll

…

peel-before-lv-to-enable-vectorization.ll

…

pixel-splat.ll

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

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

pr38280.ll

[SCEV] Collect and merge loop guards through PHI nodes with multiple incoming values (#113915)

2024-11-15 10:03:08 +00:00

pr48844-br-to-switch-vectorization.ll

[LV] Vectorize Epilogues for loops with small VF but high IC (#108190)

2024-11-17 19:35:32 +00:00

pr50392.ll

[VectorCombine] isExtractExtractCheap - specify the extract/insert shuffle mask to improve shuffle costs (#114780)

2024-11-13 12:31:39 +00:00

pr50555.ll

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

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

pr52078.ll

…

pr52253.ll

…

pr52289.ll

…

pr61061.ll

…

pr67803.ll

…

pr88239.ll

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

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

pr94546.ll

[TTI][X86] getShuffleCosts - for SK_PermuteTwoSrc, if the masks are known to be "inlane" no need to scale the costs by worst-case legalization (#117999)

2024-12-01 12:01:47 +00:00

preserve-access-group.ll

[Local] Only intersect tbaa metadata if instr moves. (#116682)

2024-11-20 19:31:16 +00:00

scalarization-inseltpoison.ll

…

scalarization.ll

…

shuffle-inseltpoison.ll

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

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

shuffle.ll

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

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

simplifycfg-late.ll

…

speculation-vs-tbaa.ll

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

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

spurious-peeling.ll

…

SROA-after-final-loop-unrolling-2.ll

…

SROA-after-final-loop-unrolling.ll

…

SROA-after-loop-unrolling.ll

…

store-constant-merge.ll

…

unroll-vectorizer.ll

Add the initializes attribute inference (#117104)

2024-11-20 19:15:23 -08:00

unsigned-multiply-overflow-check.ll

…

vdiv-nounroll.ll

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

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

vdiv.ll

[LV] Vectorize Epilogues for loops with small VF but high IC (#108190)

2024-11-17 19:35:32 +00:00

vec-load-combine.ll

…

vec-shift.ll

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

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

vector-math.ll

…

vector-reduction-known-first-value.ll

…

vector-reductions-expanded.ll

…

vector-reductions-logical.ll

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

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

vector-reductions.ll

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

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