Commits · 7d028670f6cdfb368041ae4e0ad3f4e81000cffb · Roger Ferrer / llvm-epi-0.8 · GitLab

This project is archived. Its data is read-only.

Nov 14, 2018
- [InstSimplify] add tests for funnel shift with select; NFC · 7d028670
  Sanjay Patel authored Nov 14, 2018
  7d028670
- [InstCombine] Remove a couple of asserts based on incorrect assumptions · 0905fc77
  Mandeep Singh Grang authored Nov 14, 2018
  0905fc77
- [SimplifyCFG] Regenerate preserve-branchweights.ll test. NFC · 9fd8c20c
  John Brawn authored Nov 14, 2018
  9fd8c20c
- Recommit r346483: [CallSiteSplitting] Only record conditions up to the IDom(call site). · 505091a8
  Florian Hahn authored Nov 14, 2018
  505091a8
- Revert r346810 "Preserve loop metadata when splitting exit blocks" · 41390b47
  Reid Kleckner authored Nov 14, 2018
  41390b47
- [InstCombine] fold funnel shift amount based on demanded bits · a1395648
  Sanjay Patel authored Nov 13, 2018
  a1395648
- Preserve loop metadata when splitting exit blocks · 3c87c2a3
  Craig Topper authored Nov 13, 2018
  3c87c2a3
Nov 13, 2018
- [InstCombine] canonicalize rotate patterns with cmp/select · f8f12272
  Sanjay Patel authored Nov 13, 2018
  f8f12272
- [IR] Add a dedicated FNeg IR Instruction · cbde0d9c
  Cameron McInally authored Nov 13, 2018
  cbde0d9c
- [CSP, Cloning] Update DuplicateInstructionsInSplitBetween to use DomTreeUpdater. · 107d0a87
  Florian Hahn authored Nov 13, 2018
  107d0a87
- [InstCombine] add tests for funnel shift demanded bits; NFC · bcc5a742
  Sanjay Patel authored Nov 13, 2018
  bcc5a742
- [InstCombine] add rotate variants that include select; NFC · 02f289e5
  Sanjay Patel authored Nov 12, 2018
  02f289e5
Nov 12, 2018
- [InstCombine] narrow width of rotate patterns, part 3 · 35b1c2d1
  Sanjay Patel authored Nov 12, 2018
  35b1c2d1
- [InstCombine] narrow width of rotate patterns, part 2 (PR39624) · 98e427cc
  Sanjay Patel authored Nov 12, 2018
  98e427cc
- [InstCombine] add more tests for rotate narrowing; NFC · b32d03df
  Sanjay Patel authored Nov 12, 2018
  b32d03df
- [InstCombine] regenerate checks; NFC · 8512e590
  Sanjay Patel authored Nov 12, 2018
  8512e590
- [CostModel] Add more realistic SK_InsertSubvector generic costs. · 47d38198
  Simon Pilgrim authored Nov 12, 2018
  47d38198
- [VectorUtils] add funnel-shifts to the list of vectorizable intrinsics · 1456fd76
  Sanjay Patel authored Nov 12, 2018
  1456fd76
- [LoopVectorize] add tests for funnel shifts; NFC · 75120dcb
  Sanjay Patel authored Nov 12, 2018
  75120dcb
- [LICM] Hoist guards from non-header blocks · 7d49a3a8
  Max Kazantsev authored Nov 12, 2018
  7d49a3a8
Nov 11, 2018
- [IPSCCP,PM] Preserve PDT in the new pass manager. · 9026d4ee
  Florian Hahn authored Nov 11, 2018
  9026d4ee
Nov 10, 2018
- [InstCombine] auto-generate full checks; NFC · 3482801d
  Sanjay Patel authored Nov 10, 2018
  3482801d
Nov 09, 2018
- [JumpThreading] Fix exponential time algorithm computing known values. · 15930bf3
  Eli Friedman authored Nov 09, 2018
  15930bf3
- [CostModel][X86] SK_ExtractSubvector is free if the subvector is at the start of the source vector · fc8f1d7d
  Simon Pilgrim authored Nov 09, 2018
  fc8f1d7d
- Revert r346483: [CallSiteSplitting] Only record conditions up to the IDom(call site). · 9f878e9b
  Florian Hahn authored Nov 09, 2018
  9f878e9b
- [IPSCCP,PM] Preserve DT in the new pass manager. · a1062f4b
  Florian Hahn authored Nov 09, 2018
  a1062f4b
- [CallSiteSplitting] Only record conditions up to the IDom(call site). · 52578f95
  Florian Hahn authored Nov 09, 2018
  52578f95
Nov 08, 2018
- [LoopInterchange] Support reductions across inner and outer loop. · a684a994
  Florian Hahn authored Nov 08, 2018
  a684a994
- [LTO] Drop non-prevailing definitions only if linkage is not local or appending · e61652a3
  Pirama Arumuga Nainar authored Nov 08, 2018
  e61652a3
- InstCombine: Avoid introducing poison values when lowering llvm.amdgcn.[us]bfe · 28d66216
  Tom Stellard authored Nov 08, 2018
  28d66216
- [CodeExtractor] Mark functions noreturn when applicable · d6699423
  Vedant Kumar authored Nov 08, 2018
  d6699423
- Return "[IndVars] Smart hard uses detection" · 266c087b
  Max Kazantsev authored Nov 08, 2018
  266c087b
- [LSR] Combine unfolded offset into invariant register · 7b88bab3
  Gil Rapaport authored Nov 08, 2018
  7b88bab3
- [MergeFuncs] Improve ordering of equal functions · 73cb9784
  whitequark authored Nov 08, 2018
  73cb9784
- [MergeFuncs] Call removeUsers() prior to unnamed_addr RAUW · 3580ac61
  whitequark authored Nov 08, 2018
  3580ac61
- [PGO] Exit early if all count values are zero · fb4bcc45
  Rong Xu authored Nov 07, 2018
  fb4bcc45
Nov 07, 2018
- [SimpleLoopUnswitch] partial unswitch needs to be careful when replacing invariants with constants · f9a02a70
  Fedor Sergeev authored Nov 07, 2018
  f9a02a70
- [LoopSink] Do not sink instructions into non-cold blocks · d47d188b
  Mandeep Singh Grang authored Nov 07, 2018
  d47d188b
- [NewGVN] Make sure we do not add a user to itself. · ac86038b
  Florian Hahn authored Nov 07, 2018
  ac86038b
- [InstCombine] propagate FMF for fcmp+fabs folds · 57a08b33
  Sanjay Patel authored Nov 07, 2018
  57a08b33