Commits · eba76056a364d206de7f0ac944c8601c988b6405 · Roger Ferrer / llvm-epi · GitLab

Sep 27, 2021
- [FuncSpec] Don't specialise (or crash) on poison or constexpr values · eba76056
  Sjoerd Meijer authored Sep 27, 2021
  eba76056
- Revert "Recommit "Revert "[CVP] processSwitch: Remove default case when switch... · 3a998c06
  Jun Ma authored Sep 27, 2021
  3a998c06
- [X86][Costmodel] Load/store i16 Stride=2 VF=32 interleaving costs · 7424deb7
  Roman Lebedev authored Sep 27, 2021
  7424deb7
- [X86][Costmodel] Load/store i16 Stride=2 VF=16 interleaving costs · a5113e94
  Roman Lebedev authored Sep 27, 2021
  a5113e94
- [X86][Costmodel] Load/store i16 Stride=2 VF=8 interleaving costs · 70c90cc5
  Roman Lebedev authored Sep 27, 2021
  70c90cc5
- [X86][Costmodel] Load/store i16 Stride=2 VF=4 interleaving costs · 49e532aa
  Roman Lebedev authored Sep 27, 2021
  49e532aa
- [DAGCombiner][VP] Fold zero-length or false-masked VP ops · e2b46e33
  Fraser Cormack authored Sep 01, 2021
  e2b46e33
- [ARM] Improve detection of fallthough when aligning blocks · bb2d23dc
  David Green authored Sep 27, 2021
  bb2d23dc
- [X86] combineVectorHADDSUB - remove the broken HOP(x,x) merging code (PR51974) · 468ff703
  Simon Pilgrim authored Sep 27, 2021
  468ff703
- [RISCV] Create the correct mask type when lowering EXTRACT_VECTOR_ELT · d48f6df1
  Fraser Cormack authored Sep 24, 2021
  d48f6df1
- [X86][ISel] Lowering FROUND(f16) and FROUNDEVEN(f16) · 902ec614
  Freddy Ye authored Sep 27, 2021
  902ec614
- [ORC] Add missing lock to CompileOnDemandLayer::getPerDylibResources. · 1ea8d125
  Lang Hames authored Sep 26, 2021
  1ea8d125
- [ORC] Fix SimpleRemoteEPC data races. · 4b37462a
  Lang Hames authored Sep 26, 2021
  4b37462a
Sep 26, 2021
- [BasicAA] Don't check whether GEP is sized (NFC) · 7a855596
  Nikita Popov authored Sep 26, 2021
  7a855596
- [X86][SSE] combineMulToPMADDWD - enable sext_extend_vector_inreg(vXi16) ->... · c0eff50f
  Simon Pilgrim authored Sep 26, 2021
  c0eff50f
- [X86] Fold PACK(*_EXTEND_VECTOR_INREG, UNDEF) -> *_EXTEND_VECTOR_INREG · ed3e4917
  Simon Pilgrim authored Sep 26, 2021
  ed3e4917
- Reintroduce "[ORC] Introduce EPCGenericRTDyldMemoryManager." · 6498b0e9
  Lang Hames authored Sep 26, 2021
  6498b0e9
- [X86] Fold ADD(VPMADDWD(X,Y),VPMADDWD(Z,W)) -> VPMADDWD(SHUFFLE(X,Z), SHUFFLE(Y,W)) · 3fe97672
  Simon Pilgrim authored Sep 26, 2021
  3fe97672
- [ORC][llvm-jitlink] Add debugging output to SimpleRemoteEPC (and Server). · 175c1a39
  Lang Hames authored Sep 26, 2021
  175c1a39
- [RISCV] Remove redundant declaration RISCVMnemonicSpellCheck (NFC) · c4ae4a74
  Kazu Hirata authored Sep 26, 2021
  c4ae4a74
- [X86][Costmodel] Load/store i16 VF=2 interleaving costs · d9413f46
  Roman Lebedev authored Sep 26, 2021
  d9413f46
- [DSE] Don't check getUnderlyingObject() return value (NFC) · 14a49f58
  Nikita Popov authored Sep 26, 2021
  14a49f58
- [DSE] Make DSEState non-copyable (NFC) · f3c74b72
  Nikita Popov authored Sep 26, 2021
  f3c74b72
- [InstCombine] move add after min/max intrinsic · 6063e6b4
  Sanjay Patel authored Sep 26, 2021
  6063e6b4
- [CostModel][X86] Improve AVX1/AVX2 v16i32->v16i16/v16i8 truncation costs (PR51972) · 3538ee76
  Simon Pilgrim authored Sep 26, 2021
  3538ee76
- [ORC] Wait for handleDisconnect to complete in SimpleRemoteEPC::disconnect. · 320832cc
  Lang Hames authored Sep 26, 2021
  320832cc
- [ORC] Fix uninitialized variable. · 03710492
  Lang Hames authored Sep 26, 2021
  03710492
Sep 25, 2021
- [AA] Move earliest escape tracking from DSE to AA · ba664d90
  Nikita Popov authored Sep 23, 2021
  ba664d90
- [DSE] Make capture check more precise · 327bbbb1
  Nikita Popov authored Sep 25, 2021
  327bbbb1
- [BasicAA] Don't consider Argument as escape source (NFCI) · 1c3859f3
  Nikita Popov authored Sep 25, 2021
  1c3859f3
- Revert "[ORC] Introduce EPCGenericRTDyldMemoryManager." · 99951a56
  Lang Hames authored Sep 25, 2021
  99951a56
- [ORC] Introduce EPCGenericRTDyldMemoryManager. · bef55a2b
  Lang Hames authored Sep 24, 2021
  bef55a2b
- [DAG] ReduceLoadOpStoreWidth - replace getABITypeAlign with allowsMemoryAccess (PR45116) · 18c8ed54
  Simon Pilgrim authored Sep 25, 2021
  18c8ed54
- [CostModel][X86] Adjust vXi32 multiply costs if it can be performed using PMADDWD · 8c83bd3b
  Simon Pilgrim authored Sep 25, 2021
  8c83bd3b
- [X86][SSE] combineMulToPMADDWD - mask off upper bits of sign-extended vXi32 constants · eb7c78c2
  Simon Pilgrim authored Sep 25, 2021
  eb7c78c2
- [X86][SSE] combineMulToPMADDWD - enable sext(v8i16) -> zext(v8i16) fold on sub-128 bit vectors · 2a4fa0c2
  Simon Pilgrim authored Sep 25, 2021
  
  2a4fa0c2
- [Mips] Remove redundant declarations (NFC) · 44c401bd
  Kazu Hirata authored Sep 25, 2021
  44c401bd
- [X86][SSE] combineMulToPMADDWD - enable sext(v8i16) -> zext(v8i16) fold on pre-SSE41 targets · f5a26cca
  Simon Pilgrim authored Sep 25, 2021
  f5a26cca
- [X86] X86FastISel::fastMaterializeConstant - break if-else chain to fix... · 4c72b10f
  Simon Pilgrim authored Sep 25, 2021
  4c72b10f
- [X86] combineShiftToPMULH - relax from ISA from SSE41 to SSE2 · a25f25c3
  Simon Pilgrim authored Sep 24, 2021
  a25f25c3