[all-commits] [llvm/llvm-project] 3097c6: [LoopVectorize][NFC] Rewrite tests to check output...

[David Sherwood via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 3097c60928c773d8c7f97432491c0e4367b6215c
      https://github.com/llvm/llvm-project/commit/3097c60928c773d8c7f97432491c0e4367b6215c
  Author: David Sherwood <david.sherwood at arm.com>
  Date:   2024-11-19 (Tue, 19 Nov 2024)

  Changed paths:
    M llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
    M llvm/test/Transforms/LoopVectorize/AArch64/aarch64-predication.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/extractvalue-no-scalarization-required.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/intrinsiccost.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/masked-op-cost.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/maximize-bandwidth-invalidate.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/no_vector_instructions.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/scalable-fp-ext-trunc-illegal-type.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/scalable-vectorization-cost-tuning.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/select-costs.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/sve-inductions-unusual-types.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/sve-strict-fadd-cost.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/type-shrinkage-zext-costs.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-icmpcost.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-saddsatcost.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-selectandorcost.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-shiftcost.ll
    M llvm/test/Transforms/LoopVectorize/RISCV/force-vect-msg.ll
    M llvm/test/Transforms/LoopVectorize/SystemZ/mem-interleaving-costs-02.ll
    M llvm/test/Transforms/LoopVectorize/X86/fneg-cost.ll
    M llvm/test/Transforms/LoopVectorize/X86/mul_slm_16bit.ll
    M llvm/test/Transforms/LoopVectorize/X86/reduction-small-size.ll
    M llvm/test/Transforms/LoopVectorize/X86/redundant-vf2-cost.ll
    M llvm/test/Transforms/LoopVectorize/X86/uint64_to_fp64-cost-model.ll
    M llvm/test/Transforms/LoopVectorize/X86/uniformshift.ll
    M llvm/test/Transforms/LoopVectorize/X86/vector-scalar-select-cost.ll

  Log Message:
  -----------
  [LoopVectorize][NFC] Rewrite tests to check output of vplan cost model (#113697)

Currently it's very difficult to improve the cost model for tail-folded
loops because as soon as you add a VPInstruction::computeCost function
that adds the costs of instructions such as
VPInstruction::ActiveLaneMask
and VPInstruction::ExplicitVectorLength the assert in
LoopVectorizationPlanner::computeBestVF fails for some tests. This is
because the VF chosen by the legacy cost model doesn't match the vplan
cost model. See PR #90191. This assert is currently making it difficult
to improve the cost model.

Hopefully we will be in a position to remove the assert soon, however
in order to do that we have to fix up a whole bunch of tests that rely
upon the legacy cost model output. I've tried my best to update
these tests to use vplan output instead.

There is still work needed for the VF=1 case because the vplan cost
model is not printed out in this case. I've not attempted to fix those
in this patch.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications