Commits · e745277012ec05d1e6f980e05f2a3ea7c827eeec · Roger Ferrer / llvm-epi · GitLab

Jul 26, 2021
- [AArch64] NFC: Make some AArch64-SVE LoopVectorize tests generic. · e7452770
  Sander de Smalen authored Jul 26, 2021
  e7452770
- [LV] Remove assert that VF cannot be scalable in setCostBasedWideningDecision. · b9051ba8
  Sander de Smalen authored Jul 26, 2021
  b9051ba8
- [LV] Don't assume isScalarAfterVectorization if one of the uses needs widening. · 981e9dce
  Sander de Smalen authored Jul 26, 2021
  981e9dce
- Recommit "[VPlan] Add recipe for first-order rec phis, make splicing explicit." · 7a1e73f0
  Florian Hahn authored Jul 26, 2021
  7a1e73f0
- [LV] Add test to store a first-order rec via interleave group. · 93664503
  Florian Hahn authored Jul 26, 2021
  93664503
- [SVE] Add support for folding for select + masked loads · 20b0fa91
  Dylan Fleming authored Jul 26, 2021
  20b0fa91
- Fix test failures caused by 0aff1798 · b2a5f002
  David Sherwood authored Jul 26, 2021
  
  b2a5f002
- [Analysis] Add simple cost model for strict (in-order) reductions · 0aff1798
  David Sherwood authored Jul 07, 2021
  0aff1798
Jul 25, 2021
- Revert "[VPlan] Add recipe for first-order rec phis, make splicing explicit." · b1777b04
  Nico Weber authored Jul 25, 2021
  b1777b04
Jul 22, 2021
- [LoopVectorize] Fix crash for predicated instruction with scalable VF · 5a4de84d
  Caroline Concatto authored May 05, 2021
  5a4de84d
- [InstCombine] Fold (gep (oneuse(gep Ptr, Idx0)), Idx1) -> (gep Ptr, (add Idx0, Idx1)) (PR51069) · 1c9bec72
  Simon Pilgrim authored Jul 22, 2021
  1c9bec72
Jul 21, 2021
- Revert "[LV] Use lookThroughAnd with logical reductions" · be753b20
  Kerry McLaughlin authored Jul 21, 2021
  be753b20
- [LoopVectorize] Regenerate sve-vector-reverse.ll test checks · ca9b60f9
  Simon Pilgrim authored Jul 21, 2021
  
  ca9b60f9
- [LV] Use lookThroughAnd with logical reductions · e22a5996
  Kerry McLaughlin authored Jul 19, 2021
  e22a5996
Jul 20, 2021
- [VPlan] Add recipe for first-order rec phis, make splicing explicit. · d2a73fb4
  Florian Hahn authored Jul 20, 2021
  d2a73fb4
Jul 19, 2021
- [LoopUtils] Fix incorrect RT check bounds of loop-invariant mem accesses · e908e063
  Mindong Chen authored Jul 19, 2021
  e908e063
- [LV] Re-generate check lines of some fragile tests (NFC) · f3814ed3
  Mindong Chen authored Jul 19, 2021
  f3814ed3
- [LV] Add test with ptr induction used as scalar and vector. · 156b431c
  Florian Hahn authored Jul 06, 2021
  156b431c
Jul 16, 2021
- [LV] Avoid scalable vectorization for loops containing alloca · 49d73130
  Kerry McLaughlin authored Jul 16, 2021
  49d73130
- Reland "[LV] Print remark when loop cannot be vectorized due to invalid costs." · 239d01fa
  Sander de Smalen authored Jul 15, 2021
  239d01fa
Jul 15, 2021
- [LV] Enable vectorization of multiple exit loops w/computable exit counts · 95346ba8
  Philip Reames authored Jul 15, 2021
  95346ba8
- Revert "[LV] Print remark when loop cannot be vectorized due to invalid costs." · a607f641
  Sander de Smalen authored Jul 15, 2021
  a607f641
- [LV] Fix determinism for failing scalable-call.ll test. · dc7bdc1e
  Sander de Smalen authored Jul 15, 2021
  dc7bdc1e
Jul 14, 2021
- [LV] Print remark when loop cannot be vectorized due to invalid costs. · efaf3099
  Sander de Smalen authored Jul 14, 2021
  efaf3099
- [CostModel][AArch64] Make loads/stores of <vscale x 1 x eltty> invalid. · eac16707
  Sander de Smalen authored Jul 14, 2021
  eac16707
Jul 12, 2021
- [CostModel][X86] Adjust fptosi/fptoui SSE/AVX legalized costs based on llvm-mca reports. · ae0d73ac
  Simon Pilgrim authored Jul 12, 2021
  ae0d73ac
- [LV] Ignore candidate VFs with invalid costs. · d2e4ccc7
  Sander de Smalen authored Jul 12, 2021
  d2e4ccc7
Jul 08, 2021
- [SLP][COST][X86]Improve cost model for masked gather. · 0d74fd3f
  Alexey Bataev authored Jun 28, 2021
  0d74fd3f
Jul 07, 2021
- [LV] Unconditionally branch from middle to scalar preheader if the scalar loop must execute (try 4) · 72314466
  Philip Reames authored Jul 07, 2021
  72314466
- [SVE] Fix ShuffleVector cast<FixedVectorType> in truncateToMinimalBitwidths · 7215dcfe
  Dylan Fleming authored Jul 07, 2021
  7215dcfe
- [SVE] Fix cast<FixedVectorType> in truncateToMinimalBitwidths · 7586b47f
  Dylan Fleming authored Jul 07, 2021
  7586b47f
Jul 06, 2021
- Recommit "[VPlan] Add VPReductionPHIRecipe (NFC)." and follow-ups. · ef0d147c
  Florian Hahn authored Jul 06, 2021
  ef0d147c
- [LV] Prevent vectorization with unsupported element types. · a7512401
  Kerry McLaughlin authored Jul 06, 2021
  a7512401
- Revert "[VPlan] Add VPReductionPHIRecipe (NFC)." and follow-ups · 706bbfb3
  Florian Hahn authored Jul 06, 2021
  706bbfb3
- [VPlan] Add VPReductionPHIRecipe (NFC). · 6c3451cd
  Florian Hahn authored Jul 06, 2021
  6c3451cd
Jul 05, 2021
- [LV] Extend FIXME in test add in 91ee1e37 . · 0bb9c05b
  Florian Hahn authored Jul 05, 2021
  
  0bb9c05b
- [LV] Add initial test cases with small clamped indices. · 91ee1e37
  Florian Hahn authored Jul 01, 2021
  
  91ee1e37
- [AArch64] Cost-model i8 vector loads/stores · ee752134
  Sjoerd Meijer authored Jul 01, 2021
  ee752134
Jul 02, 2021
- [CostModel][X86] Adjust uitofp(vXi64) SSE/AVX legalized costs based on llvm-mca reports. · cdca1785
  Simon Pilgrim authored Jul 02, 2021
  cdca1785
Jul 01, 2021
- [LoopVectorize][X86] Regenerate conversion-cost.ll tests · 0af9b25a
  Simon Pilgrim authored Jul 01, 2021
  
  0af9b25a