llvm-project

History

[aarch64] Add hyperbolic and arc trig intrinsic lowering (#98937 )

## The change(s)
- `VecFuncs.def`: define intrinsic to  sleef/armpl mapping
- `LegalizerHelper.cpp`: add missing `fewerElementsVector` handling for
the new trig intrinsics
- `AArch64ISelLowering.cpp`: Add arch64 specializations for lowering
like neon instructions
- `AArch64LegalizerInfo.cpp`: Legalize the new trig intrinsics. aarch64
has specail legalization requirments in `AArch64LegalizerInfo.cpp`. If
we redirect the clang builtin without handling this we will break the
aarch64 compiler

## History
This change is part of an implementation of
https://github.com/llvm/llvm-project/issues/87367's investigation on
supporting IEEE math operations as intrinsics.
Which was discussed in this RFC:
https://discourse.llvm.org/t/rfc-all-the-math-intrinsics/78294

This change adds wasm lowering cases for `acos`, `asin`, `atan`, `cosh`,
`sinh`, and `tanh`.

https://github.com/llvm/llvm-project/issues/70079
https://github.com/llvm/llvm-project/issues/70080
https://github.com/llvm/llvm-project/issues/70081
https://github.com/llvm/llvm-project/issues/70083
https://github.com/llvm/llvm-project/issues/70084
https://github.com/llvm/llvm-project/issues/95966

## Why is aarch64 needed
The last step is to redirect the `acos`, `asin`, `atan`, `cosh`, `sinh`,
and `tanh` to emit the intrinsic. We can't emit the intrinsic without
the intrinsics becoming legal for aarch64 in `AArch64LegalizerInfo.cpp`

2024-07-19 10:18:23 -04:00

aarch64-predication.ll

…

aarch64-unroll.ll

…

arbitrary-induction-step.ll

[test] Replace aarch64-*-{eabi,gnueabi}{,hf} with aarch64

2024-02-12 18:29:55 -08:00

arm64-unroll.ll

…

backedge-overflow.ll

[test] Replace aarch64-*-{eabi,gnueabi}{,hf} with aarch64

2024-02-12 18:29:55 -08:00

clamped-trip-count.ll

[LV] Remove unnecessary getRuntimeVF call when computing vector TC.

2024-06-12 14:35:37 +01:00

conditional-branches-cost.ll

[LV] Add additional test coverage for cost modeling.

2024-06-26 10:18:01 +01:00

cost-no-valid-vplans-built.ll

[VPlan] Return std::nullopt early if plans are empty.

2024-06-27 12:25:59 +01:00

deterministic-type-shrinkage.ll

[InstCombine] Infer nsw/nuw for trunc (#87910 )

2024-04-11 19:10:53 +08:00

divs-with-scalable-vfs.ll

[LV] Check isPredInst instead of isScalarWithPred in uniform analysis. (#98892 )

2024-07-19 12:02:25 +01:00

eliminate-tail-predication.ll

[VPlan] Compute scalable VF in preheader for induction increment. (#74762 )

2023-12-08 12:18:31 +00:00

epilog-vectorization-widen-inductions.ll

[VPlan] Explicitly handle scalar pointer inductions. (#83068 )

2024-03-26 16:01:57 +01:00

extend-vectorization-factor-for-unprofitable-memops.ll

…

extractvalue-no-scalarization-required.ll

…

first-order-recurrence-fold-tail.ll

[VPlan] Introduce ResumePhi VPInstruction, use to create phi for FOR. (#94760 )

2024-07-11 16:08:04 +01:00

first-order-recurrence.ll

Move several vector intrinsics out of experimental namespace (#88748 )

2024-04-29 10:16:45 +01:00

fixed-order-recurrence.ll

[VPlan] Introduce ResumePhi VPInstruction, use to create phi for FOR. (#94760 )

2024-07-11 16:08:04 +01:00

gather-cost.ll

…

gather-do-not-vectorize-addressing.ll

[VPlan] Compute scalable VF in preheader for induction increment. (#74762 )

2023-12-08 12:18:31 +00:00

i1-reg-usage.ll

…

induction-costs-sve.ll

[LV] Remove unnecessary getRuntimeVF call when computing vector TC.

2024-06-12 14:35:37 +01:00

induction-costs.ll

[VPlan] Introduce ResumePhi VPInstruction, use to create phi for FOR. (#94760 )

2024-07-11 16:08:04 +01:00

induction-trunc.ll

…

interleave_count_for_estimated_tc.ll

[LV] Remove loop trip count threshold for deciding whether to interleave a loop (#67725 )

2024-02-05 17:23:58 -08:00

interleave_count_for_known_tc.ll

[LV] Remove loop trip count threshold for deciding whether to interleave a loop (#67725 )

2024-02-05 17:23:58 -08:00

interleave-allocsize-not-equal-typesize.ll

…

interleaved_cost.ll

[test] Replace aarch64-*-{eabi,gnueabi}{,hf} with aarch64

2024-02-12 18:29:55 -08:00

interleaved-store-of-first-order-recurrence.ll

…

interleaved-vs-scalar.ll

…

interleaving-load-store.ll

[VPlan] Model address separately. (#72164 )

2024-01-01 19:51:15 +00:00

interleaving-reduction.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

intrinsiccost.ll

[InstCombine] Infer nsw/nuw for trunc (#87910 )

2024-04-11 19:10:53 +08:00

lit.local.cfg

…

loop-vectorization-factors.ll

[VPlan] Introduce ResumePhi VPInstruction, use to create phi for FOR. (#94760 )

2024-07-11 16:08:04 +01:00

loopvectorize_pr33804_double.ll

[VPlan] Add a test for testing unused interleave recipes (#75026 )

2023-12-14 21:16:11 +08:00

low_trip_memcheck_cost.ll

[LoopVectorize] Fix divide-by-zero bug (#80836 ) (#81721 )

2024-02-14 16:07:51 +00:00

masked-call.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

masked-op-cost.ll

[TTI] getCommonMaskedMemoryOpCost - use the target getMemoryOpCost/getCFInstrCost implementations.

2024-05-11 12:50:26 +01:00

max-vf-for-interleaved.ll

…

maximize-bandwidth-invalidate.ll

…

neon-gather-cost.ll

…

no_vector_instructions.ll

…

nontemporal-load-store.ll

…

outer_loop_prefer_scalable.ll

[VPlan] Update HCFG builder after 72937203dd3b to fix leak.

2024-07-09 15:28:43 +01:00

outer_loop_test1_no_explicit_vect_width.ll

…

Oz-and-forced-vectorize.ll

…

pr31900.ll

…

pr33053.ll

…

pr36032.ll

…

pr46950-load-cast-context-crash.ll

…

pr56627.ll

…

pr60831-sve-inv-store-crash.ll

[VPlan] Compute scalable VF in preheader for induction increment. (#74762 )

2023-12-08 12:18:31 +00:00

pr73894.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

predication_costs.ll

…

prefer-fixed-if-equal-to-scalable.ll

[LV][AArch64] Prefer Fixed over Scalable if cost-model is equal (Neoverse V2) (#95819 )

2024-07-17 10:46:28 +01:00

reduction-recurrence-costs-sve.ll

[VPlan] Introduce ResumePhi VPInstruction, use to create phi for FOR. (#94760 )

2024-07-11 16:08:04 +01:00

reduction-small-size.ll

…

reg-usage.ll

…

runtime-check-size-based-threshold.ll

…

runtime-check-trip-count-decisions.ll

[LoopVectorize] Enhance Vectorization decisions for predicate tail-folded loops with low trip counts (#69588 )

2023-10-30 13:43:26 +00:00

scalable-alloca.ll

…

scalable-avoid-scalarization.ll

[VPlan] Initial modeling of VF * UF as VPValue. (#74761 )

2023-12-08 18:30:30 +00:00

scalable-call.ll

[VFABI] Move the Vector ABI demangling utility to LLVMCore. (#77513 )

2024-01-17 09:55:30 +00:00

scalable-predicate-instruction.ll

…

scalable-reduction-inloop-cond.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

scalable-reductions-tf.ll

…

scalable-reductions.ll

[LoopVectorize][AArch64] Add limited support for scalable vectorisation of i1 types (#95920 )

2024-06-25 15:04:24 +01:00

scalable-strict-fadd.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

scalable-vectorization-cost-tuning.ll

…

scalable-vectorization.ll

…

scalable-vf-hint.ll

…

scalar_interleave.ll

…

scalarize-store-with-predication.ll

…

sdiv-pow2.ll

…

select-costs.ll

[LoopVectorize] LLVM fails to vectorise loops with multi-bool varables (#89226 )

2024-07-15 20:21:50 +01:00

smallest-and-widest-types.ll

…

sme-vectorize.ll

[AArch64] Consider streaming mode in TTI interfaces for vectorization. (#96305 )

2024-06-24 11:06:16 +01:00

store-costs-sve.ll

[LV] Remove unnecessary getRuntimeVF call when computing vector TC.

2024-06-12 14:35:37 +01:00

streaming-compatible-sve-no-maximize-bandwidth.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

strict-fadd-cost.ll

…

strict-fadd-vf1.ll

…

strict-fadd.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

sve-basic-vec.ll

…

sve-cond-inv-loads.ll

[InstCombine] Canonicalize constant GEPs to i8 source element type (#68882 )

2024-01-24 15:25:29 +01:00

sve-epilog-vect-inloop-reductions.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

sve-epilog-vect-reductions.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

sve-epilog-vect-strict-reductions.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

sve-epilog-vect-vscale-tune.ll

…

sve-epilog-vect.ll

[VPlan] Explicitly handle scalar pointer inductions. (#83068 )

2024-03-26 16:01:57 +01:00

sve-extract-last-veclane.ll

…

sve-fneg.ll

[VPlan] Model address separately. (#72164 )

2024-01-01 19:51:15 +00:00

sve-gather-scatter-cost.ll

…

sve-gather-scatter.ll

[InstCombine] Canonicalize non-i8 gep of mul to i8 (#96606 )

2024-06-26 14:25:54 +01:00

sve-illegal-type.ll

[LoopVectorize][AArch64] Add limited support for scalable vectorisation of i1 types (#95920 )

2024-06-25 15:04:24 +01:00

sve-inductions-unusual-types.ll

[LV] Add missing check lines in vector.ph in tests.

2024-07-16 10:45:53 +01:00

sve-inductions.ll

[VPlan] Initial modeling of VF * UF as VPValue. (#74761 )

2023-12-08 18:30:30 +00:00

sve-interleaved-accesses.ll

[VPlan] Introduce ResumePhi VPInstruction, use to create phi for FOR. (#94760 )

2024-07-11 16:08:04 +01:00

sve-interleaved-masked-accesses.ll

Move several vector intrinsics out of experimental namespace (#88748 )

2024-04-29 10:16:45 +01:00

sve-inv-loads.ll

…

sve-inv-store.ll

[VPlan] Initial modeling of VF * UF as VPValue. (#74761 )

2023-12-08 18:30:30 +00:00

sve-invalid-costs.ll

[LoopVectorize] Fix build issue caused by #95920 (#96647 )

2024-06-25 15:51:32 +01:00

sve-large-strides.ll

…

sve-live-out-pointer-induction.ll

[VPlan] Don't add live-outs for IV phis.

2024-07-14 20:49:03 +01:00

sve-low-trip-count.ll

[LV] Remove unnecessary getRuntimeVF call when computing vector TC.

2024-06-12 14:35:37 +01:00

sve-masked-loadstore.ll

…

sve-multiexit.ll

[VPlan] Model address separately. (#72164 )

2024-01-01 19:51:15 +00:00

sve-runtime-check-size-based-threshold.ll

[VPlan] Model address separately. (#72164 )

2024-01-01 19:51:15 +00:00

sve-scalable-load-in-loop.ll

…

sve-select-cmp.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

sve-strict-fadd-cost.ll

…

sve-tail-folding-cost.ll

…

sve-tail-folding-forced.ll

[LV] Remove unnecessary getRuntimeVF call when computing vector TC.

2024-06-12 14:35:37 +01:00

sve-tail-folding-option.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

sve-tail-folding-optsize.ll

[LV] Remove unnecessary getRuntimeVF call when computing vector TC.

2024-06-12 14:35:37 +01:00

sve-tail-folding-overflow-checks.ll

[VPlan] Initial modeling of VF * UF as VPValue. (#74761 )

2023-12-08 18:30:30 +00:00

sve-tail-folding-reductions.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

sve-tail-folding-unroll.ll

[LV] Remove unnecessary getRuntimeVF call when computing vector TC.

2024-06-12 14:35:37 +01:00

sve-tail-folding.ll

[LV] Remove unnecessary getRuntimeVF call when computing vector TC.

2024-06-12 14:35:37 +01:00

sve-type-conv.ll

…

sve-vector-reverse-mask4.ll

Move several vector intrinsics out of experimental namespace (#88748 )

2024-04-29 10:16:45 +01:00

sve-vector-reverse.ll

Move several vector intrinsics out of experimental namespace (#88748 )

2024-04-29 10:16:45 +01:00

sve-vfabi.ll

[VPlan] Initial modeling of VF * UF as VPValue. (#74761 )

2023-12-08 18:30:30 +00:00

sve-widen-extractvalue.ll

…

sve-widen-gep.ll

[VPlan] Explicitly handle scalar pointer inductions. (#83068 )

2024-03-26 16:01:57 +01:00

sve-widen-phi.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

synthesize-mask-for-call.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

tail-fold-uniform-memops.ll

…

tail-folding-styles.ll

[LV] Remove unnecessary getRuntimeVF call when computing vector TC.

2024-06-12 14:35:37 +01:00

type-shrinkage-insertelt.ll

[VPlan] Replace IR based truncateToMinimalBitwidths with VPlan version.

2023-12-02 16:12:38 +00:00

type-shrinkage-zext-costs.ll

[VPlan] Compute scalable VF in preheader for induction increment. (#74762 )

2023-12-08 12:18:31 +00:00

uniform-args-call-variants.ll

[InstCombine] Canonicalize non-i8 gep of mul to i8 (#96606 )

2024-06-26 14:25:54 +01:00

unsafe-vf-hint-remark.ll

…

veclib-calls-libsystem-darwin.ll

[aarch64] Add hyperbolic and arc trig intrinsic lowering (#98937 )

2024-07-19 10:18:23 -04:00

veclib-function-calls.ll

[AArch64][TLI] Add TLI mappings for ArmPL modf, sincos, sincospi (#83143 )

2024-04-10 09:34:46 +01:00

veclib-intrinsic-calls.ll

[aarch64] Add hyperbolic and arc trig intrinsic lowering (#98937 )

2024-07-19 10:18:23 -04:00

vector-call-linear-args.ll

[LV] Fix handling of interleaving linear args (#78725 )

2024-01-26 11:30:35 +00:00

vector-reverse-mask4.ll

[VPlan] Split VPWidenMemoryInstructionRecipe (NFCI). (#87411 )

2024-04-17 11:00:58 +01:00

vector-reverse.ll

[VPlan] Model address separately. (#72164 )

2024-01-01 19:51:15 +00:00

widen-call-with-intrinsic-or-libfunc.ll

[VPlan] Model branch cond to enter scalar epilogue in VPlan. (#92651 )

2024-07-05 10:08:42 +01:00

wider-VF-for-callinst.ll

[VPlan] Compute scalable VF in preheader for induction increment. (#74762 )

2023-12-08 12:18:31 +00:00