Commit message (Collapse) | Author | Age | ||
---|---|---|---|---|
... | ||||
* | Update VectorwiseOp.h to allow Plugins similar to MatrixBase.h or ArrayBase.h | dlazenby | 2020-03-20 | |
| | ||||
* | Bug https://gitlab.com/libeigen/eigen/-/issues/1415: add missing ↵ | Masaki Murooka | 2020-03-20 | |
| | | | | EIGEN_DEVICE_FUNC to diagonal_product_evaluator_base. | |||
* | Remove reference to non-existent unary_op_base class. | Rasmus Munk Larsen | 2020-03-19 | |
| | ||||
* | Add missing arguments to numext::absdiff(). | Rasmus Munk Larsen | 2020-03-19 | |
| | ||||
* | Add absolute_difference coefficient-wise binary Array function | Joel Holdsworth | 2020-03-19 | |
| | ||||
* | Reenabling packetmath unsigned tests, adding dummy pabs for relevant unsigned | Everton Constantino | 2020-03-19 | |
| | | | | types. | |||
* | Add shift_left<N> and shift_right<N> coefficient-wise unary Array functions | Joel Holdsworth | 2020-03-19 | |
| | ||||
* | Implement integer square-root for NEON | Joel Holdsworth | 2020-03-19 | |
| | ||||
* | Update NullaryFunctors.h | Allan Leal | 2020-03-16 | |
| | ||||
* | Fixing HIP breakage caused by the recent commit that introduces Packet4h2 as ↵ | Deven Desai | 2020-03-12 | |
| | | | | the Eigen::Half packet type | |||
* | NEON: Added int64_t and uint64_t packet math | Joel Holdsworth | 2020-03-10 | |
| | ||||
* | NEON: Added int8_t and uint8_t packet math | Joel Holdsworth | 2020-03-10 | |
| | ||||
* | NEON: Added int16_t and uint16_t packet math | Joel Holdsworth | 2020-03-10 | |
| | ||||
* | NEON: Added uint32_t packet math | Joel Holdsworth | 2020-03-10 | |
| | ||||
* | NEON: Implemented half-size vectors | Joel Holdsworth | 2020-03-10 | |
| | ||||
* | NEON: Set packet_traits<double> flags | Joel Holdsworth | 2020-03-10 | |
| | ||||
* | test/packetmath: Add tests for all integer types | Joel Holdsworth | 2020-03-10 | |
| | ||||
* | test/packetmath: Made negate non-mandatory | Joel Holdsworth | 2020-03-10 | |
| | ||||
* | remove duplicate pset1 for half and add some comments about why we need ↵ | Sami Kama | 2020-03-10 | |
| | | | | expose pmul/add/div/min/max on host | |||
* | Don't restrict CMAKE_BUILD_TYPE | Ram-Z | 2020-02-28 | |
| | | | This prevents projects that add Eigen using `add_subdirectory` from using their own custom CMAKE_BUILD_TYPE and have Eigen respect the same custom flags. | |||
* | Update MarketIO.h | Cédric Hubert | 2020-02-28 | |
| | ||||
* | Revert "avoid selecting half-packets when unnecessary" | Rasmus Munk Larsen | 2020-02-25 | |
| | | | This reverts commit 5ca10480b0756e40b0723d90adeba8506291fc7c | |||
* | Revert "Pick full packet unconditionally when EIGEN_UNALIGNED_VECTORIZE" | Rasmus Munk Larsen | 2020-02-25 | |
| | | | This reverts commit 44df2109c8c700222643a9a45f144676348f4df1 | |||
* | Revert "do not pick full-packet if it'd result in more operations" | Rasmus Munk Larsen | 2020-02-25 | |
| | | | This reverts commit e9cc0cd353803a818204e48054bd89699b84e6c6 | |||
* | Revert "add some static checks for packet-picking logic" | Rasmus Munk Larsen | 2020-02-25 | |
| | | | This reverts commit 776960024585b907acc4abc3c59aef605941bb75 | |||
* | Revert "Disable test in test/vectorization_logic.cpp, which is currently ↵ | Rasmus Munk Larsen | 2020-02-25 | |
| | | | | | failing with AVX." This reverts commit b625adffd877639ff5cbe51ea154e1905a3b405c | |||
* | Disable test in test/vectorization_logic.cpp, which is currently failing ↵ | Rasmus Munk Larsen | 2020-02-24 | |
| | | | | with AVX. | |||
* | Include <sstream> explicitly, and don't rely on the implicit include via ↵ | Tobias Bosch | 2020-02-24 | |
| | | | | | <complex>. This implicit dependency does no longer exist in a recent llbm release (sha 78be61871704). | |||
* | Avoid a division in NonBlockingThreadPool::Steal. | Ilya Tokar | 2020-02-14 | |
| | | | | | | | Looking at profiles we spend ~10-20% of Steal on simply computing random % size. We can reduce random 32-bit int into [0, size) range with a single multiplication and shift. This transformation is described in https://lemire.me/blog/2016/06/27/a-fast-alternative-to-the-modulo-reduction/ | |||
* | add some static checks for packet-picking logic | Francesco Mazzoli | 2020-02-07 | |
| | ||||
* | do not pick full-packet if it'd result in more operations | Francesco Mazzoli | 2020-02-07 | |
| | | | | | See comment and <https://gitlab.com/libeigen/eigen/merge_requests/46#note_270622952>. | |||
* | Pick full packet unconditionally when EIGEN_UNALIGNED_VECTORIZE | Francesco Mazzoli | 2020-02-07 | |
| | | | | See comment for details. | |||
* | avoid selecting half-packets when unnecessary | Francesco Mazzoli | 2020-02-07 | |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | See <https://stackoverflow.com/questions/59709148/ensuring-that-eigen-uses-avx-vectorization-for-a-certain-operation> for an explanation of the problem this solves. In short, for some reason, before this commit the half-packet is selected when the array / matrix size is not a multiple of `unpacket_traits<PacketType>::size`, where `PacketType` starts out being the full Packet. For example, for some data of 100 `float`s, `Packet4f` will be selected rather than `Packet8f`, because 100 is not a multiple of 8, the size of `Packet8f`. This commit switches to selecting the half-packet if the size is less than the packet size, which seems to make more sense. As I stated in the SO post I'm not sure that I'm understanding the issue correctly, but this fix resolves the issue in my program. Moreover, `make check` passes, with the exception of line 614 and 616 in `test/packetmath.cpp`, which however also fail on master on my machine: CHECK_CWISE1_IF(PacketTraits::HasBessel, numext::bessel_i0, internal::pbessel_i0); ... CHECK_CWISE1_IF(PacketTraits::HasBessel, numext::bessel_i1, internal::pbessel_i1); | |||
* | Fail at compile time if default executor tries to use non-default device | Eugene Zhulenev | 2020-02-06 | |
| | ||||
* | Remove dead code from TensorReduction.h | Eugene Zhulenev | 2020-01-29 | |
| | ||||
* | fix hip-clang compilation due to new HIP scalar accessor | Jeff Daily | 2020-01-20 | |
| | ||||
* | Fix for HIP breakage - 200115. Adding a missing EIGEN_DEVICE_FUNC attr | Deven Desai | 2020-01-16 | |
| | ||||
* | Ensure Igamma does not NaN or Inf for large values. | Srinivas Vasudevan | 2020-01-14 | |
| | ||||
* | Remove rogue include in TypeCasting.h. Meta.h is already included by the ↵ | Rasmus Munk Larsen | 2020-01-14 | |
| | | | | top-level header in Eigen/Core. | |||
* | Convert StridedLinearBufferCopy::Kind to enum class | Eugene Zhulenev | 2020-01-13 | |
| | ||||
* | Switching unpacket_traits<Packet4i> to vectorizable=true. | Everton Constantino | 2020-01-13 | |
| | ||||
* | Adding correct cache sizes for PPC architecture. | Everton Constantino | 2020-01-13 | |
| | ||||
* | Removing executable bit from file mode | Christoph Hertzberg | 2020-01-11 | |
| | ||||
* | Bug #1790: Make `areApprox` check `numext::isnan` instead of bitwise ↵ | Christoph Hertzberg | 2020-01-11 | |
| | | | | equality (NaNs don't have to be bitwise equal). | |||
* | Added special_packetmath test and tweaked bounds on tests. | Srinivas Vasudevan | 2020-01-11 | |
| | | | | | Refactor shared packetmath code to header file. (Squashed from PR !38) | |||
* | call Explicitly ::rint and ::rintf for targets without c++11. Without this, ↵ | Rasmus Munk Larsen | 2020-01-10 | |
| | | | | the Windows build breaks when trying to compile numext::rint<double>. | |||
* | Improvements to the tidiness and completeness of the NEON implementation | Joel Holdsworth | 2020-01-10 | |
| | ||||
* | Fix for gcc build error when using Eigen headers with AVX512 | Anuj Rawat | 2020-01-10 | |
| | ||||
* | Adding RInt vector support for SYCL. | mehdi-goli | 2020-01-10 | |
| | ||||
* | Properly initialize b vector in SplineFitting | Matthew Powelson | 2020-01-09 | |
| | | | InterpolateWithDerivative does not initialize the be vector correctly. This issue is discussed In stackoverflow question 48382939. |