Commits

Breadcrumbs

History for

onmaster

User selector

All users

Datepicker

All time

Commit History

Commits on Aug 11, 2022

• Support in-place in CCpuMathEngine::VectorPowerDiffOp for X86 (neoml-lib#747)

  

  Valeriy Fedyunin

  authoredAug 11, 2022
  e8aacaa

Commits on Aug 10, 2022

• Optimize LoadSse and LoadIntSse (neoml-lib#743)

  

  Valeriy Fedyunin

  authoredAug 10, 2022
  99ff652

Commits on Aug 8, 2022

• Remove unused LoadFromVector functions (neoml-lib#741)

  

  Valeriy Fedyunin

  authoredAug 8, 2022
  0229e3c

Commits on Aug 5, 2022

• Fix memory leaks in sparse matrix tests (neoml-lib#740)

  

  Valeriy Fedyunin

  authoredAug 5, 2022
  a0e4705

Commits on Aug 4, 2022

• Don't use vceqzq_s32 instruction (it's not available on ARM v7) (neoml-lib#735)

  

  Valeriy Fedyunin

  authoredAug 4, 2022
  94bd254

• Add IMathEngine::VectorEltwiseWhere (neoml-lib#720)

  

  Valeriy Fedyunin

  and

  SAngeliuk

  authoredAug 4, 2022
  0a28dbc

• Add IMathEngine::ScatterND (neoml-lib#719)

  

  Valeriy Fedyunin

  and

  SAngeliuk

  authoredAug 4, 2022
  1f650b1

Commits on Aug 3, 2022

• Don't use vceqzq_s32 instruction (it's not available on ARM v7) (neoml-lib#732)

  

  Valeriy Fedyunin

  authoredAug 3, 2022
  1f16097

• Fix signed/unsigned comparison warning (neoml-lib#731)

  

  Valeriy Fedyunin

  authoredAug 3, 2022
  bba40ed

• Fix encoding in CpuFunctorCommon.h (neoml-lib#727)

  

  Valeriy Fedyunin

  authoredAug 3, 2022
  4baeb1c

Commits on Aug 2, 2022

• Switch to own thread pool in distributed learning (neoml-lib#723)

  

  SAngeliuk

  and

  Valeriy Fedyunin

  authoredAug 2, 2022
  d605e57

Commits on Aug 1, 2022

• Add integer version of IMathEngine::SumMatrixRows (neoml-lib#716)

  

  Valeriy Fedyunin

  authoredAug 1, 2022
  a6eb044

• Add IMathEngine::VectorEltwiseEqual (neoml-lib#714)

  

  Valeriy Fedyunin

  and

  SAngeliuk

  authoredAug 1, 2022
  9dc80ee

Commits on Jul 25, 2022

• Extend VectorCumSumAlongDimension (neoml-lib#703)

  

  Valeriy Fedyunin

  authoredJul 25, 2022
  7f5f784

• Add CErfLayer (neoml-lib#700)

  

  Valeriy Fedyunin

  authoredJul 25, 2022
  762c151

Commits on Jul 22, 2022

• Add IMathEngine::VectorErf (neoml-lib#699)

  

  Valeriy Fedyunin

  authoredJul 22, 2022
  18d5b0c

• Add more versions of IMathEngine::VectorEltwiseLess (neoml-lib#698)

  

  Valeriy Fedyunin

  authoredJul 22, 2022
  d858fac

• Implement IMathEngine::VectorEltwiseNot (neoml-lib#697)

  

  Valeriy Fedyunin

  authoredJul 22, 2022
  5f74bbd

• Add integer version of IMathEngine::VectorMultiply (neoml-lib#694)

  

  Valeriy Fedyunin

  and

  SAngeliuk

  authoredJul 22, 2022
  df1a512

Commits on Jul 18, 2022

• Add CLogLayer (neoml-lib#688)

  

  Valeriy Fedyunin

  authoredJul 18, 2022
  0c54725

Commits on Jul 12, 2022

• only utf only hardcode (neoml-lib#672)

  

  voropz

  and

  Valeriy Fedyunin

  authoredJul 12, 2022
  ee937fc

Commits on Jul 7, 2022

• Clean memory as often as possible in distributed CPU (neoml-lib#677)

  

  Valeriy Fedyunin

  authoredJul 7, 2022
  7b4910c

Commits on Jul 6, 2022

• Disable memory pools for distributed CCpuMathEngine (temporarily) (neoml-lib#676)

  

  Valeriy Fedyunin

  authoredJul 6, 2022
  b9c4222

Commits on Jul 5, 2022

• Check for AVX2 instructions instead of AVX (neoml-lib#675)

  

  Valeriy Fedyunin

  authoredJul 5, 2022
  73bc34c

Commits on Jun 29, 2022

• [arm64] exp, log and tanh fix (neoml-lib#673)

  SanSanch5

  authoredJun 29, 2022
  905cf72

Commits on Jun 22, 2022

• Use unity build in OSS tests (neoml-lib#653)

  

  Valeriy Fedyunin

  authoredJun 22, 2022
  0ce77e9

Commits on Jun 21, 2022

• Use CSimdMathEngine::Sigmoid when possible (neoml-lib#664)

  

  Valeriy Fedyunin

  and

  SAngeliuk

  authoredJun 21, 2022
  5854899

• Enable AvxDll (except for convolution) when AVX512 is available (neoml-lib#663)

  

  Valeriy Fedyunin

  authoredJun 21, 2022
  cd849a2

• Fix infinite type conversion in CTypedMemoryHandle (neoml-lib#661)

  

  Valeriy Fedyunin

  and

  SAngeliuk

  authoredJun 21, 2022
  f3a1073

• Optimize CPU BroadcastCopy when additionalWidth == 1 (neoml-lib#658)

  

  Valeriy Fedyunin

  and

  SAngeliuk

  authoredJun 21, 2022
  a047c9b

Commits on Jun 17, 2022

• Optimize dataCopy for x86 CPU (neoml-lib#660)

  

  Valeriy Fedyunin

  authoredJun 17, 2022
  e10507f

Commits on Jun 16, 2022

• Optimize IMathEngine::TransposeMatrix (neoml-lib#655)

  

  Valeriy Fedyunin

  authoredJun 16, 2022
  662b1a9

Commits on Jun 7, 2022

• Fix TLS leak in case of multiple load/free cycles of NeoMathEngine.dll (neoml-lib#650)

  

  Valeriy Fedyunin

  authoredJun 7, 2022
  59e96d7

Commits on Jun 3, 2022

• Turn off metal by default (neoml-lib#647)

  SAngeliuk

  authoredJun 3, 2022
  5d60a3c

Commits on Jun 1, 2022

• Update CExpNeon algorithm, make it work with FPE traps enabled (neoml-lib#634)

  SanSanch5

  authoredJun 1, 2022
  f4a0c93

Pagination

PreviousNext