Commits · master · Arthur / eigen · GitLab

Aug 23, 2022
- Sparse Core: Replace malloc/free with conditional_aligned · b2a13c9d
  Charles Schlosser authored Aug 23, 2022 and Rasmus Munk Larsen committed Aug 23, 2022
  
  b2a13c9d
Aug 22, 2022
- Fix psign for unsigned integer types, such as bool. · 6aad0f82
  Rasmus Munk Larsen authored Aug 22, 2022
  
  6aad0f82
- Protect new pblend implementation with EIGEN_VECTORIZE_AVX2 · 1a09defc
  Rasmus Munk Larsen authored Aug 22, 2022
  
  1a09defc
Aug 17, 2022
- Use proper double word division algorithm for pow<double>. Gives 11-15% speedup. · 7c67dc67
  Rasmus Munk Larsen authored Aug 17, 2022
  
  7c67dc67
- Add support for AVX512-FP16 for vectorizing half precision math · 7a3b667c
  Matthew Sterrett authored Aug 17, 2022 and Rasmus Munk Larsen committed Aug 17, 2022
  
  7a3b667c
Aug 16, 2022
- add fixed power unary operation · 76a669fb
  Charles Schlosser authored Aug 16, 2022 and Rasmus Munk Larsen committed Aug 16, 2022
  
  76a669fb
- Removed unnecessary checks for FP16C · 39fcc897
  Matthew Sterrett authored Aug 16, 2022
  
  39fcc897
- [SYCL] Fix some SYCL tests · 2f7cce2d
  Romain Biessy authored Aug 16, 2022 and Rasmus Munk Larsen committed Aug 16, 2022
  
  2f7cce2d
Aug 11, 2022
- Disable bad "deprecated warning" edge-case in BDCSVD · 27367017
  Arthur authored Aug 11, 2022 and Rasmus Munk Larsen committed Aug 11, 2022
  
  27367017
Aug 09, 2022
- Eliminate bool bitwise warnings. · b8e93bf5
  Antonio Sánchez authored Aug 09, 2022 and Rasmus Munk Larsen committed Aug 09, 2022
  
  b8e93bf5
- Don't double-define Half functions on aarch64 · 66ea0c09
  Lexi Bromfield authored Aug 09, 2022 and Antonio Sánchez committed Aug 09, 2022
  
  66ea0c09
- Vectorize the sign operator in Eigen. · 97e0784d
  Rasmus Munk Larsen authored Aug 09, 2022
  
  97e0784d
Aug 08, 2022
- Fix vectorized Jacobi Rotation · be20207d
  Arthur authored Aug 08, 2022 and Rasmus Munk Larsen committed Aug 08, 2022
  
  be20207d
- Fix code and unit test for a few corner cases in vectorized pow() · 7a87ed1b
  Rasmus Munk Larsen authored Aug 08, 2022
  
  7a87ed1b
- Fix non-VSX PowerPC build · 9e0afe0f
  Chip Kerchner authored Aug 08, 2022 and Rasmus Munk Larsen committed Aug 08, 2022
  
  9e0afe0f
Aug 03, 2022
- Fix use of Packet2d type for non-VSX. · 84a9d6fa
  Chip Kerchner authored Aug 03, 2022 and Rasmus Munk Larsen committed Aug 03, 2022
  
  84a9d6fa
- Partial Packet support for GEMM real-only (PowerPC). Also fix compilation... · ce60a7be
  Chip Kerchner authored Aug 03, 2022 and Rasmus Munk Larsen committed Aug 03, 2022
  ce60a7be
- Fix inner iterator for sparse block. · 5a1c7807
  Antonio Sánchez authored Aug 03, 2022 and Rasmus Munk Larsen committed Aug 03, 2022
  
  5a1c7807
Aug 02, 2022
- Fix flaky packetmath_1 test. · 39d22ef4
  Antonio Sánchez authored Aug 02, 2022 and Rasmus Munk Larsen committed Aug 02, 2022
  
  39d22ef4
Jul 29, 2022
- Use numext::sqrt in ConjugateGradient. · 7896c7dc
  Antonio Sánchez authored Jul 29, 2022 and Rasmus Munk Larsen committed Jul 29, 2022
  
  7896c7dc
- Improve pblend AVX implementation · e618c4a5
  Ilya Tokar authored Jul 29, 2022 and Rasmus Munk Larsen committed Jul 29, 2022
  
  e618c4a5
- Add true determinant to QR and it's variants · ef4654ba
  sjusju authored Jul 29, 2022 and Rasmus Munk Larsen committed Jul 29, 2022
  
  ef4654ba
- Avoid including <sstream> with EIGEN_NO_IO · b7668c03
  Alexander Richardson authored Jul 29, 2022 and Rasmus Munk Larsen committed Jul 29, 2022
  
  b7668c03
- Updated AccelerateSupport documentation after PR 966. · 7dd3dda3
  John Mather authored Jul 29, 2022 and Rasmus Munk Larsen committed Jul 29, 2022
  
  7dd3dda3
Jul 28, 2022
- Add Sparse Subset of Matrix Inverse · 69714ff6
  Julian Kent authored Jul 28, 2022 and Rasmus Munk Larsen committed Jul 28, 2022
  
  69714ff6
Jul 22, 2022
- Include immintrin.h header for enscripten. · 34780d8b
  Antonio Sánchez authored Jul 22, 2022
  
  34780d8b
Jul 20, 2022
- Disable AVX512 GEMM kernels by default. · 2cf4d18c
  Antonio Sánchez authored Jul 20, 2022
  
  2cf4d18c
Jul 19, 2022
- Fix aligned_realloc to call check_that_malloc_is_allowed() if ptr == 0 · a678a3e0
  Charles Schlosser authored Jul 19, 2022 and Antonio Sánchez committed Jul 19, 2022
  
  a678a3e0
Jul 18, 2022
- Add option to disable avx512 GEBP kernels · 4a563594
  b-shi authored Jul 18, 2022 and Antonio Sánchez committed Jul 18, 2022
  
  4a563594
Jul 12, 2022
- Fix wrong doxygen group usage · 1092574b
  Mathieu Westphal authored Jul 12, 2022
  
  1092574b
Jul 09, 2022
- AutoDiff depends on Core, so include appropriate header. · e1165dbf
  Antonio Sánchez authored Jul 09, 2022
  
  e1165dbf
- Fix ODR violations. · bb51d9f4
  Antonio Sánchez authored Jul 09, 2022
  
  bb51d9f4
Jun 29, 2022
- Enable subtests which use device side malloc since this has been fixed in ROCm 5.2. · 06a458a1
  Rohit Santhanam authored Jun 29, 2022
  
  06a458a1
Jun 27, 2022
- Add pload_partial, pstore_partial (and unaligned versions), pgather_partial,... · 84cf3ff1
  Chip Kerchner authored Jun 27, 2022 and Rasmus Munk Larsen committed Jun 27, 2022
  84cf3ff1
- Better performance for Power10 using more load and store vector pairs for GEMV · c603275d
  Chip Kerchner authored Jun 27, 2022 and Rasmus Munk Larsen committed Jun 27, 2022
  
  c603275d
Jun 24, 2022
- Fix clang-tidy warnings about function definitions in headers. · 0e187141
  Antonio Sánchez authored Jun 23, 2022 and Antonio Sánchez committed Jun 24, 2022
  
  0e187141
- Skip f16/bf16 bessel specializations on AVX512 if unavailable. · 8ed3b9dc
  Antonio Sánchez authored Jun 24, 2022
  
  8ed3b9dc
- Eliminate undef warnings when not compiling for AVX512. · bc2ab816
  Antonio Sánchez authored Jun 24, 2022
  
  bc2ab816
Jun 21, 2022
- Use numext::sqrt in Householder.h. · 0e083b17
  Antonio Sánchez authored Jun 21, 2022 and Rasmus Munk Larsen committed Jun 21, 2022
  
  0e083b17
Jun 17, 2022
- AVX512 TRSM kernels use alloca if EIGEN_NO_MALLOC requested · 37673ca1
  b-shi authored Jun 17, 2022 and Rasmus Munk Larsen committed Jun 17, 2022
  
  37673ca1