Commits · d3a52089eb3e9d8b41dd8ca08d3d532dee78c4e8 · Roger Ferrer / llvm-epi · GitLab

Mar 04, 2022
- Reapply [MergeICmps] Don't require GEP · d3a52089
  Nikita Popov authored Feb 25, 2022
  d3a52089
- [MergeICmp] Make instruction move robust against empty block (NFCI) · 53abe3ff
  Nikita Popov authored Mar 04, 2022
  53abe3ff
- [AArch64] Improve access to fixed-width object when stack has SVE. · 7c65d228
  Sander de Smalen authored Mar 03, 2022
  7c65d228
- [AArch64] NFC: Add test for access to fixed-width stack object when stack has SVE. · d363bdda
  Sander de Smalen authored Mar 01, 2022
  d363bdda
- [Bitcode] Move x86_intrcc upgrade to bitcode reader · 7a258c6a
  Nikita Popov authored Mar 04, 2022
  7a258c6a
- [Bitcode] Fully support opaque pointer auto upgrade · e3a9f68e
  Nikita Popov authored Feb 23, 2022
  e3a9f68e
- [AArch64] Additional insert-subvector codegen tests. NFC · 04661a4d
  David Green authored Mar 04, 2022
  
  04661a4d
- [LTO] Initialize canAutoHide() using canBeOmittedFromSymbolTable() · dd29597e
  Jez Ng authored Mar 03, 2022
  dd29597e
- [RISCV] With Zbb, fold (sext_inreg (abs X)) -> (max X, (negw X)) · 3d4e83f1
  Craig Topper authored Mar 03, 2022
  3d4e83f1
- Revert "[MergeICmps] Don't require GEP" · bc1574b4
  Arthur Eubanks authored Mar 03, 2022
  bc1574b4
Mar 03, 2022
- [Mips] support "sp" named register · e0adc3be
  Nick Desaulniers authored Mar 03, 2022
  e0adc3be
- Revert "Inliner: Correctly merge amdgpu-unsafe-fp-atomics attribute" · 27712243
  Matt Arsenault authored Mar 03, 2022
  27712243
- [x86] add tests for vector rotate; NFC · 5f944aac
  Sanjay Patel authored Mar 03, 2022
  
  5f944aac
- [AArch64][x86] add tests for rotate/funnel combines; NFC · 74a65e38
  Sanjay Patel authored Mar 03, 2022
  
  74a65e38
- Revert "Reapply "[SLP] Schedule only sub-graph of vectorizable instructions""" · deae979a
  Philip Reames authored Mar 03, 2022
  deae979a
- [AMDGPU] Added hsa-gfx90a-v3.s test. NFC. · 13256caf
  Stanislav Mekhanoshin authored Mar 03, 2022
  
  13256caf
- Revert "[GlobalOpt] Don't replace alias with aliasee if either alias/aliasee may be preemptible" · f0b61f79
  Arthur Eubanks authored Mar 03, 2022
  f0b61f79
- [AMDGPU] Add test for instructions unsupported on gfx940. NFC. · f7698991
  Stanislav Mekhanoshin authored Mar 03, 2022
  
  f7698991
- [IVDescriptor] Find original 'Previous' for first-order recurrences. · 139215af
  Florian Hahn authored Mar 03, 2022
  139215af
- [UpdateTestChecks] Don't skip first line with --filter · 473efae3
  Sebastian Neubauer authored Feb 14, 2022
  473efae3
- [Attributor] Regenerate test checks (NFC) · 82a4b0b9
  Nikita Popov authored Mar 03, 2022
  
  82a4b0b9
- [SDAG] try harder to remove a rotate from X == 0 · e9302bf7
  Sanjay Patel authored Mar 03, 2022
  e9302bf7
- [x86] add tests for setcc of rotate; NFC · d3c16be9
  Sanjay Patel authored Mar 02, 2022
  
  d3c16be9
- [AArch64] Use first op of FADDPv* instead of implicit def. · 0f261256
  Florian Hahn authored Mar 03, 2022
  0f261256
- [RISCV] Add cost modelling for masked memory op · 89f15fc6
  Alex Tsao authored Mar 03, 2022
  89f15fc6
- [X86][XOP] Tidyup VPHADD/VPHSUB unary horizontal ops default schedule class · 0c9c92ff
  Simon Pilgrim authored Mar 03, 2022
  0c9c92ff
- [NFC] TypePromotion test · a8b4f5bb
  Sam Parker authored Mar 03, 2022
  
  a8b4f5bb
- Reapply [lit] Read command stdout/stderr as text on Windows · 9e2236fb
  Martin Storsjö authored Feb 27, 2022
  9e2236fb
- [AArch64] Update costs for some fp16 converts · 47f4cd9c
  David Green authored Mar 03, 2022
  47f4cd9c
- [AArch64] Vector shifts cost 1 · 65c0e45a
  David Green authored Mar 03, 2022
  65c0e45a
- [InstCombine] Support opaque pointers in callee bitcast fold · c1b96671
  Nikita Popov authored Mar 03, 2022
  c1b96671
- [InstCombine] Remove unnecessary byval check in callee cast fold · 6c8adc50
  Nikita Popov authored Mar 03, 2022
  6c8adc50
- [InstCombine] Add callee bitcast test with byval on callee (NFC) · 2555ed55
  Nikita Popov authored Mar 03, 2022
  2555ed55
- [AArch64] Allow copying of SVE registers in Streaming SVE · e4fa8291
  Cullen Rhodes authored Mar 03, 2022
  e4fa8291
- [AArch64] Add legal types for Streaming SVE · 61658679
  Cullen Rhodes authored Mar 03, 2022
  61658679
- [RISCV] Fold store of vmv.f.s to a vse with VL=1. · 44a43035
  Jianjian GUAN authored Feb 21, 2022
  44a43035
- [PowerPC][atomics] Precommit test cases for i128 cmpxchg. NFC. · 1cfcbf19
  Kai Luo authored Mar 03, 2022
  
  1cfcbf19
- [AMDGPU] Removed XFAIL from hsa-gfx940-v3.s. NFC. · b05918f2
  Stanislav Mekhanoshin authored Mar 02, 2022
  b05918f2
- [AMDGPU] Add gfx1036 target · 84069581
  Aakanksha authored Mar 02, 2022
  84069581
Mar 02, 2022
- [AMDGPU] Fix 3 tests with expensive checks. NFC. · ad786f5a
  Stanislav Mekhanoshin authored Mar 02, 2022
  ad786f5a