caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
c41b41eb11b4e6debf2f1bfb2b1d637b209a14c2
clang-p2996/llvm/test/Transforms/LoopVectorize/AArch64
History
Sander de Smalen c41b41eb11 [LoopVectorize] Use overflow-check analysis to improve tail-folding. 
This work follows on from D142109 and addresses a possible regression
when we know the loop iteration counter cannot overflow.

When we know the overflow-check always evaluates to false, it's better to
use the other style of tail folding where it assumes a runtime check was
added, because that avoids having to calculate a modified trip-count.

Reviewed By: paulwalker-arm

Differential Revision: https://reviews.llvm.org/D142894
2023-03-01 14:17:58 +00:00
..
aarch64-predication.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
aarch64-unroll.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
arbitrary-induction-step.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2023-01-04 17:25:42 +01:00
arm64-unroll.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
backedge-overflow.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
deterministic-type-shrinkage.ll
[AsmParser] Remove typed pointer auto-detection
2023-01-18 09:58:32 +01:00
eliminate-tail-predication.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
epilog-vectorization-widen-inductions.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
extend-vectorization-factor-for-unprofitable-memops.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
extractvalue-no-scalarization-required.ll
[AsmParser] Remove typed pointer auto-detection
2023-01-18 09:58:32 +01:00
first-order-recurrence-fold-tail.ll
[LoopVectorize] Use overflow-check analysis to improve tail-folding.
2023-03-01 14:17:58 +00:00
first-order-recurrence.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
fixed-order-recurrence.ll
[NFC] Port all runlines for LoopVectorize pass tests to -passes syntax
2022-12-05 22:17:30 +03:00
gather-cost.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
gather-do-not-vectorize-addressing.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
i1-reg-usage.ll
[NFC] Port all runlines for LoopVectorize pass tests to -passes syntax
2022-12-05 22:17:30 +03:00
induction-trunc.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
interleave-allocsize-not-equal-typesize.ll
[VectorUtils] Skip interleave members with diff type and alloca sizes.
2022-11-13 22:06:20 +00:00
interleaved_cost.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
interleaved-store-of-first-order-recurrence.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
interleaved-vs-scalar.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
interleaving-load-store.ll
[AArch64] Set MaxInterleaveFactor for Apple A14, A15, A16.
2023-01-11 18:52:51 +00:00
interleaving-reduction.ll
[AArch64] Set MaxInterleaveFactor for Apple A14, A15, A16.
2023-01-11 18:52:51 +00:00
intrinsiccost.ll
[InstCombine][LV] Fold (add (zext (add X, -1)), 1) -> (zext X) if X is non-zero.
2023-01-30 17:45:01 -08:00
lit.local.cfg
…
loop-vectorization-factors.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
loopvectorize_pr33804_double.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
masked-call.ll
[LV] Synthesize all true masks for masked vector function variants
2023-02-14 14:33:18 +00:00
masked-op-cost.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
max-vf-for-interleaved.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
maximize-bandwidth-invalidate.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2023-01-04 17:25:42 +01:00
no_vector_instructions.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
nontemporal-load-store.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2023-01-04 17:25:42 +01:00
outer_loop_test1_no_explicit_vect_width.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
Oz-and-forced-vectorize.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
pr31900.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
pr33053.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
pr36032.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
pr46950-load-cast-context-crash.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
pr56627.ll
…
predication_costs.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
reduction-small-size.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
reg-usage.ll
[LV] Reland testcase in 0ec4cae
2023-02-28 18:14:18 +05:30
runtime-check-size-based-threshold.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
scalable-alloca.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
scalable-avoid-scalarization.ll
[LV] Add users for loads to make tests more robust.
2023-02-04 20:42:57 +00:00
scalable-call.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
scalable-predicate-instruction.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
scalable-reduction-inloop-cond.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
scalable-reductions-tf.ll
[LoopVectorize] Remove runtime check and scalar tail loop when tail-folding.
2023-03-01 09:01:19 +00:00
scalable-reductions.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
scalable-strict-fadd.ll
[AsmParser] Remove typed pointer auto-detection
2023-01-18 09:58:32 +01:00
scalable-vectorization-cost-tuning.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
scalable-vectorization.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
scalable-vf-hint.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
scalar_interleave.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2023-01-04 17:25:42 +01:00
scalarize-store-with-predication.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
sdiv-pow2.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
select-costs.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
sleef-calls-aarch64.ll
[AArch64] Enable libm vectorized functions via SLEEF
2023-01-20 18:52:38 +01:00
smallest-and-widest-types.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
sme-vectorize.ll
[NFC] Port all runlines for LoopVectorize pass tests to -passes syntax
2022-12-05 22:17:30 +03:00
strict-fadd-cost.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
strict-fadd-vf1.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
strict-fadd.ll
[AsmParser] Remove typed pointer auto-detection
2023-01-18 09:58:32 +01:00
sve-basic-vec.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
sve-cond-inv-loads.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
sve-epilog-vect-inloop-reductions.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
sve-epilog-vect-reductions.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
sve-epilog-vect-strict-reductions.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
sve-epilog-vect.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
sve-extract-last-veclane.ll
[AsmParser] Remove typed pointer auto-detection
2023-01-18 09:58:32 +01:00
sve-fneg.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
sve-gather-scatter-cost.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
sve-gather-scatter.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
sve-illegal-type.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2023-01-04 17:25:42 +01:00
sve-inductions-unusual-types.ll
[AsmParser] Remove typed pointer auto-detection
2023-01-18 09:58:32 +01:00
sve-inductions.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
sve-inv-loads.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
sve-inv-store.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
sve-large-strides.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
sve-live-out-pointer-induction.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
sve-low-trip-count.ll
[LoopVectorize] Use overflow-check analysis to improve tail-folding.
2023-03-01 14:17:58 +00:00
sve-masked-loadstore.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
sve-runtime-check-size-based-threshold.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
sve-scalable-load-in-loop.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
sve-select-cmp.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
sve-strict-fadd-cost.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
sve-tail-folding-cost.ll
[NFC] Port all runlines for LoopVectorize pass tests to -passes syntax
2022-12-05 22:17:30 +03:00
sve-tail-folding-forced.ll
[LoopVectorize] Remove runtime check and scalar tail loop when tail-folding.
2023-03-01 09:01:19 +00:00
sve-tail-folding-option.ll
[AsmParser] Remove typed pointer auto-detection
2023-01-18 09:58:32 +01:00
sve-tail-folding-optsize.ll
[LoopVectorize] Use overflow-check analysis to improve tail-folding.
2023-03-01 14:17:58 +00:00
sve-tail-folding-overflow-checks.ll
[LoopVectorize] Use overflow-check analysis to improve tail-folding.
2023-03-01 14:17:58 +00:00
sve-tail-folding-reductions.ll
[LoopVectorize] Remove runtime check and scalar tail loop when tail-folding.
2023-03-01 09:01:19 +00:00
sve-tail-folding-unroll.ll
[LoopVectorize] Remove runtime check and scalar tail loop when tail-folding.
2023-03-01 09:01:19 +00:00
sve-tail-folding.ll
[LoopVectorize] Remove runtime check and scalar tail loop when tail-folding.
2023-03-01 09:01:19 +00:00
sve-type-conv.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
sve-vector-reverse-mask4.ll
[InstCombine] Bubble vector.reverse of compare operands to their result.
2022-12-21 15:53:14 +00:00
sve-vector-reverse.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
sve-widen-extractvalue.ll
[IRBuilder] Use canonical i64 type for insertelement index used by vector splats.
2023-01-11 14:08:06 +00:00
sve-widen-gep.ll
[AsmParser] Remove typed pointer auto-detection
2023-01-18 09:58:32 +01:00
sve-widen-phi.ll
Reland D143267: [LoopVectorize] Use DataLayout::getIndexType instead of i32 for non-constant GEP indices.
2023-02-09 09:42:29 +00:00
synthesize-mask-for-call.ll
[LV] Synthesize all true masks for masked vector function variants
2023-02-14 14:33:18 +00:00
tail-fold-uniform-memops.ll
[LoopVectorize] Remove runtime check and scalar tail loop when tail-folding.
2023-03-01 09:01:19 +00:00
tail-folding-styles.ll
[LoopVectorize] Remove runtime check and scalar tail loop when tail-folding.
2023-03-01 09:01:19 +00:00
type-shrinkage-insertelt.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
unsafe-vf-hint-remark.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
veclib-calls-libsystem-darwin.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
vector-reverse-mask4.ll
[LoopVectorize] Convert some tests to opaque pointers (NFC)
2022-12-14 15:16:59 +01:00
vector-reverse.ll
[AsmParser] Remove typed pointer auto-detection
2023-01-18 09:58:32 +01:00
widen-call-with-intrinsic-or-libfunc.ll
[LV] Synthesize all true masks for masked vector function variants
2023-02-14 14:33:18 +00:00