devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • 562985bac4 bug #1646: fix false aliasing detection for A.row(0) = A.col(0); This changeset completely disable the detection for vectors for which are current mechanism cannot detect any positive aliasing anyway. Gael Guennebaud 2019-01-17 00:14:27 +01:00
  • 7401e2541d Fix compilation error for logical packet ops with older compilers. Rasmus Munk Larsen 2019-01-16 14:43:33 -08:00
  • ee550a2ac3 Fix flaky test for tensor fft. Rasmus Munk Larsen 2019-01-16 14:03:12 -08:00
  • 0f028f61cb GEBP: fix swapped kernel mode with AVX512 and complex scalars Gael Guennebaud 2019-01-16 22:26:38 +01:00
  • e118ce86fd GEBP: cleanup logic to choose between a 4 packets of 1 packet Gael Guennebaud 2019-01-16 21:47:42 +01:00
  • 70e133333d bug #1661: fix regression in GEBP and AVX512 Gael Guennebaud 2019-01-16 21:22:20 +01:00
  • ce88e297dc Add a comment stating this doc page is partly obsolete. Gael Guennebaud 2019-01-16 16:29:02 +01:00
  • 729d1291c2 bug #1585: update doc on lazy-evaluation Gael Guennebaud 2019-01-16 16:28:17 +01:00
  • c8e40edac9 Remove Eigen2ToEigen3 migration page (obsolete since 3.3) Gael Guennebaud 2019-01-16 16:27:00 +01:00
  • aeffdf909e bug #1617: add unit tests for empty triangular solve. Gael Guennebaud 2019-01-16 15:24:59 +01:00
  • 502f717980 bug #1646: disable aliasing detection for empty and 1x1 expression Gael Guennebaud 2019-01-16 14:33:45 +01:00
  • 0b466b6933 bug #1633: use proper type for madd temporaries, factorize RhsPacketx4. Gael Guennebaud 2019-01-16 13:50:13 +01:00
  • dbfcceabf5 Bug: 1633: refactor gebp kernel and optimize for neon Renjie Liu 2019-01-16 12:51:36 +08:00
  • 2b70b2f570 Make Transform::rotation() an alias to Transform::linear() in the case of an Isometry Gael Guennebaud 2019-01-15 22:50:42 +01:00
  • 2c2c114995 Silent maybe-uninitialized warnings by gcc Gael Guennebaud 2019-01-15 16:53:15 +01:00
  • 6ec6bf0b0d Enable visitor on empty matrices (the visitor is left unchanged), and protect min/maxCoeff(Index*,Index*) on empty matrices by an assertion (+ doc & unit tests) Gael Guennebaud 2019-01-15 15:21:14 +01:00
  • 027e44ed24 bug #1592: makes partial min/max reductions trigger an assertion on inputs with a zero reduction length (+doc and tests) Gael Guennebaud 2019-01-15 15:13:24 +01:00
  • f8bc5cb39e Fix detection of vector-at-time: use Rows/Cols instead of MaxRow/MaxCols. This fix VectorXd(n).middleCol(0,0).outerSize() which was equal to 1. Gael Guennebaud 2019-01-15 15:09:49 +01:00
  • 32d7232aec fix always true warning with gcc 4.7 Gael Guennebaud 2019-01-15 11:18:48 +01:00
  • 6cf7afa3d9 Typo Gael Guennebaud 2019-01-15 11:04:37 +01:00
  • e7d4d4f192 cleanup Gael Guennebaud 2019-01-15 10:51:03 +01:00
  • 7b3aab0936 Merged in rmlarsen/eigen (pull request PR-570) Rasmus Larsen 2019-01-14 21:31:33 +00:00
  • 8bf00c2baf Remove extra <tr>. Rasmus Munk Larsen 2019-01-14 13:29:29 -08:00
  • ec7fe83554 Merge. Rasmus Munk Larsen 2019-01-14 13:26:58 -08:00
  • 2ea4efc0c3 Merge. Rasmus Munk Larsen 2019-01-14 13:26:58 -08:00
  • 2c5843dbbb Update documentation. Rasmus Munk Larsen 2019-01-14 13:26:34 -08:00
  • 250dcd1fdb bug #1652: fix position of EIGEN_ALIGN16 attributes in Neon and Altivec Gael Guennebaud 2019-01-14 21:45:56 +01:00
  • 5a59452aae Merged eigen/eigen into default Rasmus Larsen 2019-01-14 10:23:23 -08:00
  • 3c9e6d206d AVX512: fix pgather/pscatter for Packet4cd and unaligned pointers Gael Guennebaud 2019-01-14 17:57:28 +01:00
  • 61b6eb05fe AVX512 (r)sqrt(double) was mistakenly disabled with clang and others Gael Guennebaud 2019-01-14 17:28:47 +01:00
  • ccddeaad90 fix warning Gael Guennebaud 2019-01-14 16:51:16 +01:00
  • d4881751d3 Doc: add Isometry in the list of supported Mode of Transform<> Gael Guennebaud 2019-01-14 16:38:26 +01:00
  • 3d18879fc4 Initialize isometric transforms like affine transforms. Greg Coombe 2019-01-11 23:14:35 -08:00
  • cd12bf6317 bug #1761: Manually grafted changes in Meta.h 5937c4ae32 Fall back is_integral to std::is_integral in c++11 f0862b062f Fix internal::is_integral<size_t/ptrdiff_t> with MSVC 2013 and older. Christoph Hertzberg 2019-10-24 13:29:35 +02:00
  • 1760432f6e Provide numext::[u]int{32,64}_t. Manually grafted from 4356a55a61 Christoph Hertzberg 2019-10-11 16:51:08 +02:00
  • a7d6a65d49 Implement c++03 compatible fix for changeset 7a43af1a33 Gael Guennebaud 2019-10-09 16:00:57 +02:00
  • 74d474e7df Resolve merge conflict Christoph Hertzberg 2019-10-09 15:52:12 +02:00
  • 47e2f8a42c PR 719: fix real/imag namespace conflict (grafted from 87427d2eaa ) Gael Guennebaud 2019-10-08 09:15:17 +02:00
  • 3cf273591a Fix compilation of FFTW unit test (grafted from 7a43af1a33 ) Gael Guennebaud 2019-10-08 08:58:35 +02:00
  • 8ee2e10af4 Fix perf issue in SimplicialLDLT::solve for complexes (again, m_diag is real) (grafted from 9549ba8313 ) Gael Guennebaud 2019-10-01 12:54:25 +02:00
  • 165db26dc0 Fix speed issue with SimplicialLDLT for complexes: the diagonal is real! (grafted from c8b2c603b0 ) Gael Guennebaud 2019-09-30 16:14:34 +02:00
  • 52d159c19f bug #1741: fix self-adjoint*matrix, triangular*matrix, and triangular^1*matrix with a destination having a non-trivial inner-stride Gael Guennebaud 2019-09-11 15:04:25 +02:00
  • 6abc9e5379 Fix compilation of BLAS backend and frontend Gael Guennebaud 2019-09-11 10:02:37 +02:00
  • f722e43770 bug #1741: fix SelfAdjointView::rankUpdate and product to triangular part for destination with non-trivial inner stride (grafted from c06e6fd115 ) Gael Guennebaud 2019-09-10 23:29:52 +02:00
  • f483c7ea8a bug #1741: fix C.noalias() = A*C; with C.innerStride()!=1 (grafted from ea0d5dc956 ) Gael Guennebaud 2019-09-10 16:25:24 +02:00
  • 9d988a1e1a Initialize isometric transforms like affine transforms. Greg Coombe 2019-01-11 23:14:35 -08:00
  • 4356a55a61 PR 571: Implements an accurate argument reduction algorithm for huge inputs of sin/cos and call it instead of falling back to std::sin/std::cos. This makes both the small and huge argument cases faster because: - for small inputs this removes the last pselect - for large inputs only the reduction part follows a scalar path, the rest use the same SIMD path as the small-argument case. Gael Guennebaud 2019-01-14 13:54:01 +01:00
  • f566724023 Fix StorageIndex FIXME in dense LU solvers Gael Guennebaud 2019-01-13 17:54:30 +01:00
  • 1c6e6e2c3f Merge. Rasmus Munk Larsen 2019-01-11 17:47:11 -08:00
  • 0ba3b45419 Merged eigen/eigen into default Rasmus Larsen 2019-01-11 17:46:04 -08:00
  • 28ba1b2c32 Add support for inverse hyperbolic functions. Fix cost of division. Rasmus Munk Larsen 2019-01-11 17:45:37 -08:00
  • a49d01edba Fix warnings in ptrue for complex and half types. Rasmus Munk Larsen 2019-01-11 13:18:17 -08:00
  • 9005f0111f Replace compiler's alignas/alignof extension by respective c++11 keywords when available. This also fix a compilation issue with gcc-4.7. Gael Guennebaud 2019-01-11 17:10:54 +01:00
  • 89c4001d6f Fix warnings in ptrue for complex and half types. Rasmus Munk Larsen 2019-01-11 14:10:57 -08:00
  • 1e6d15b55b Fix shorten-64-to-32 warning in TensorContractionThreadPool Eugene Zhulenev 2019-01-11 11:41:53 -08:00
  • df29511ac0 Fix merge. Rasmus Munk Larsen 2019-01-11 10:36:36 -08:00
  • 8e71ed4cc9 Merge. Rasmus Munk Larsen 2019-01-11 10:35:07 -08:00
  • fff5a5b579 Resolve. Rasmus Munk Larsen 2019-01-11 10:28:52 -08:00
  • 9396ace46b Merge. Rasmus Munk Larsen 2019-01-11 10:28:52 -08:00
  • 74882471d0 Merged eigen/eigen into default Rasmus Larsen 2019-01-11 10:20:55 -08:00
  • e9936cf2b9 Merge. Rasmus Munk Larsen 2019-01-11 09:58:33 -08:00
  • 3c9add6598 Remove reinterpret_cast from AVX512 complex implementation Mark D Ryan 2019-01-11 14:02:09 +01:00
  • 0522460a0d bug #1656: Enable failtests only if BUILD_TESTING is enabled Christoph Hertzberg 2019-01-11 11:07:56 +01:00
  • 0abe03764c Fix shorten-64-to-32 warning in TensorContractionThreadPool Eugene Zhulenev 2019-01-10 10:27:55 -08:00
  • fcfced13ed Rename pones -> ptrue. Use _CMP_TRUE_UQ where appropriate. Rasmus Munk Larsen 2019-01-09 17:20:33 -08:00
  • ce38c342c3 merge. Rasmus Munk Larsen 2019-01-09 17:20:33 -08:00
  • a05ec7993e merge Rasmus Munk Larsen 2019-01-09 17:17:30 -08:00
  • e15bb785ad Collapsed revision * Add packet up "pones". Write pnot(a) as pxor(pones(a), a). * Collapsed revision * Simplify a bit. * Undo useless diffs. * Fix typo. Rasmus Munk Larsen 2019-01-09 16:34:23 -08:00
  • f6ba6071c5 Fix typo. Rasmus Munk Larsen 2019-01-09 16:34:23 -08:00
  • 8f04442526 Collapsed revision * Collapsed revision * Add packet up "pones". Write pnot(a) as pxor(pones(a), a). * Collapsed revision * Simplify a bit. * Undo useless diffs. * Fix typo. Rasmus Munk Larsen 2019-01-09 16:34:23 -08:00
  • 8f178429b9 Collapsed revision * Collapsed revision * Add packet up "pones". Write pnot(a) as pxor(pones(a), a). * Collapsed revision * Simplify a bit. * Undo useless diffs. * Fix typo. Rasmus Munk Larsen 2019-01-09 16:34:23 -08:00
  • 1119c73d22 Collapsed revision * Add packet up "pones". Write pnot(a) as pxor(pones(a), a). * Collapsed revision * Simplify a bit. * Undo useless diffs. * Fix typo. Rasmus Munk Larsen 2019-01-09 16:34:23 -08:00
  • e00521b514 Undo useless diffs. Rasmus Munk Larsen 2019-01-09 16:32:53 -08:00
  • f2767112c8 Simplify a bit. Rasmus Munk Larsen 2019-01-09 16:29:18 -08:00
  • cb955df9a6 Add packet up "pones". Write pnot(a) as pxor(pones(a), a). Rasmus Munk Larsen 2019-01-09 16:17:08 -08:00
  • cb3c059fa4 Merged eigen/eigen into default Rasmus Larsen 2019-01-09 15:04:17 -08:00
  • d812f411c3 bug #1654: fix compilation with cuda and no c++11 Gael Guennebaud 2019-01-09 18:00:05 +01:00
  • 3492a1ca74 fix plog(+inf) with AVX512 Gael Guennebaud 2019-01-09 16:53:37 +01:00
  • 47810cf5b7 Add dedicated implementations of predux_any for AVX512, NEON, and Altivec/VSE Gael Guennebaud 2019-01-09 16:40:42 +01:00
  • 3f14e0d19e fix warning Gael Guennebaud 2019-01-09 15:45:21 +01:00
  • aeec68f77b Add missing pcmp_lt and others for AVX512 Gael Guennebaud 2019-01-09 15:36:41 +01:00
  • e6b217b8dd bug #1652: implements a much more accurate version of vectorized sin/cos. This new version achieve same speed for SSE/AVX, and is slightly faster with FMA. Guarantees are as follows: - no FMA: 1ULP up to 3pi, 2ULP up to sin(25966) and cos(18838), fallback to std::sin/cos for larger inputs - FMA: 1ULP up to sin(117435.992) and cos(71476.0625), fallback to std::sin/cos for larger inputs Gael Guennebaud 2019-01-09 15:25:17 +01:00
  • e70ffef967 Optimize evalShardedByInnerDim Eugene Zhulenev 2019-01-08 16:26:31 -08:00
  • 055f0b73db Add support for pcmp_eq and pnot, including for complex types. Rasmus Munk Larsen 2019-01-07 16:53:36 -08:00
  • 190d053e41 Explicitly set fill character when printing aligned data to ostream Eugene Zhulenev 2019-01-03 14:55:28 -08:00
  • bc5dd4cafd PR560: Fix the AVX512f only builds Mark D Ryan 2019-01-03 14:33:04 +01:00
  • 697fba3bb0 Fix unit test Gael Guennebaud 2018-12-27 11:20:47 +01:00
  • 60d3fe9a89 One more stupid AVX 512 fix (I don't have direct access to AVX512 machines) Gael Guennebaud 2018-12-24 13:05:03 +01:00
  • 4aa667b510 Add EIGEN_STRONG_INLINE where required Gael Guennebaud 2018-12-24 10:45:01 +01:00
  • 961ff567e8 Add missing pcmp_lt_or_nan for AVX512 Gael Guennebaud 2018-12-23 22:13:29 +01:00
  • 0f6f75bd8a Implement a faster fix for sin/cos of large entries that also correctly handle INF input. Gael Guennebaud 2018-12-23 17:26:21 +01:00
  • 38d704def8 Make sure that psin/pcos return number in [-1,1] for large inputs (though sin/cos on large entries is quite useless because it's inaccurate) Gael Guennebaud 2018-12-23 16:13:24 +01:00
  • 5713fb7feb Fix plog(+INF): it returned ~87 instead of +INF Gael Guennebaud 2018-12-23 15:40:52 +01:00
  • 6dd93f7e3b Make code compile again for older compilers. See https://stackoverflow.com/questions/7411515/ Christoph Hertzberg 2018-12-22 13:09:07 +01:00
  • 1024a70e82 gebp: Add new ½ and ¼ packet rows per (peeling) round on the lhs MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Gustavo Lima Chaves 2018-12-21 11:03:18 -08:00
  • e763fcd09e Introducing "vectorized" byte on unpacket_traits structs Gustavo Lima Chaves 2018-12-19 14:24:44 -08:00
  • efa4c9c40f bug #1615: slightly increase the default unrolling limit to compensate for changeset 101ea26f5e . This solves a performance regression with clang and 3x3 matrix products. Gael Guennebaud 2018-12-13 10:42:39 +01:00
  • f20c991679 add changesets related to matrix product perf. Gael Guennebaud 2018-12-13 10:33:29 +01:00
  • dd6d65898a Fix shorten-64-to-32 warning. Use regular memcpy if num_threads==0. Rasmus Munk Larsen 2018-12-12 14:45:31 -08:00
  • f582ea3579 Fix compilation with expression template scalar type. Gael Guennebaud 2018-12-12 22:47:00 +01:00