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

  • aee693ac52 Add log1p support for CUDA and half floats Igor Babuschkin 2016-08-08 20:24:59 +01:00
  • 72096f3bd4 Merged in suiyuan2009/eigen/fix_tanh_inconsistent_for_tensorflow (pull request PR-215) Benoit Steiner 2016-08-08 09:06:45 -07:00
  • 3e4a33d4ba bug #1272: Let CommaInitializer work for more border cases (enhances fix of bug #1242). The unit test tests all combinations of 2x2 block-sizes from 0 to 3. Christoph Hertzberg 2016-08-08 17:26:48 +02:00
  • 841e075154 Remove CUDA >= 300 checks and enable outer reductin for doubles Igor Babuschkin 2016-08-06 18:07:50 +01:00
  • 1031223c09 fix tanh inconsistent Ziming Dong 2016-08-06 19:48:50 +08:00
  • 5cf1e4c79b create fix_tanh_inconsistent branch Ziming Dong 2016-08-06 15:54:33 +08:00
  • 0425118e2a Merge upstream changes Igor Babuschkin 2016-08-05 14:34:57 +01:00
  • 9537e8b118 Make use of atomicExch for atomicExchCustom Igor Babuschkin 2016-08-05 14:29:58 +01:00
  • fe4b927e9c Add aliases Eigen_*_DIR to Eigen3_*_DIR This is to make configuring work again after project was renamed from Eigen to Eigen3 Christoph Hertzberg 2016-08-05 15:21:14 +02:00
  • fe778427f2 Fixed the constructors of the new half_base class. Benoit Steiner 2016-08-04 18:32:26 -07:00
  • 5eea1c7f97 Fixed cut and paste bug in debud message Benoit Steiner 2016-08-04 17:34:13 -07:00
  • 9506343349 Fixed the isnan, isfinite and isinf operations on GPU Benoit Steiner 2016-08-04 17:25:53 -07:00
  • b50d8f8c4a Extended a regression test to validate that we basic fp16 support works with cuda 7.0 Benoit Steiner 2016-08-03 16:50:13 -07:00
  • fad9828769 Deleted redundant regression test. Benoit Steiner 2016-08-03 16:08:37 -07:00
  • 373bb12dc6 Check that it's possible to forward declare the hlaf type. Benoit Steiner 2016-08-03 16:07:31 -07:00
  • 17b9a55d98 Move Eigen::half_impl::half to Eigen::half while preserving the free functions to the Eigen::half_impl namespace together with ADL Gael Guennebaud 2016-08-04 00:00:43 +02:00
  • ca2cee2739 Merged in ibab/eigen (pull request PR-206) Benoit Steiner 2016-08-03 11:53:04 -07:00
  • d92df04ce8 Cleaned up the new float16 test a bit Benoit Steiner 2016-08-03 11:50:07 -07:00
  • 81099ef482 Added a test for fp16 Benoit Steiner 2016-08-03 11:41:17 -07:00
  • a20b58845f CUDA_ARCH isn't always defined, so avoid relying on it too much when figuring out which implementation to use for reductions. Instead rely on the device to tell us on which hardware version we're running. Benoit Steiner 2016-08-03 10:00:43 -07:00
  • 819d0cea1b List PARDISO solver. Gael Guennebaud 2016-08-02 23:32:41 +02:00
  • f4404777ff Change project name to Eigen3, to be compatible with FindEigen3.cmake and Eigen3Config.cmake. This is related to pull-requests 214. Christoph Hertzberg 2016-08-02 17:08:57 +00:00
  • fd220dd8b0 Use numext::conj instead of std::conj Benoit Steiner 2016-08-01 18:16:16 -07:00
  • e256acec7c Avoid unecessary object copies Benoit Steiner 2016-08-01 17:03:39 -07:00
  • 7995cec90c Fix vectorization logic for coeff-based product for some corner cases. Gael Guennebaud 2016-07-31 15:20:22 +02:00
  • 02fe89f5ef half implementation has been moved to half_impl namespace Benoit Steiner 2016-07-29 15:09:34 -07:00
  • 2693fd54bf bug #1266: half implementation has been moved to half_impl namespace Benoit Steiner 2016-07-29 13:45:56 -07:00
  • c5b893f434 bug #1266: half implementation has been moved to half_impl namespace Christoph Hertzberg 2016-07-29 18:36:08 +02:00
  • 4057f9b1fc Enable slice-vectorization+inner-unrolling when unaligned vectorization is allowed. For instance, this permits to vectorize 5x5 matrices (including product) Gael Guennebaud 2016-07-28 13:47:33 +02:00
  • 5fbe7aa604 Update and fix Cholesky mini benchmark Gael Guennebaud 2016-07-28 11:26:30 +02:00
  • a72752caac Vectorize more small product expressions by letting the general assignement logic decides on the sizes that are OK for vectorization. Gael Guennebaud 2016-07-28 11:21:07 +02:00
  • cc2f6d68b1 bug #1264: fix compilation Gael Guennebaud 2016-07-27 23:30:47 +02:00
  • 188590db82 Add instructions for LAPACKE+Accelerate Gael Guennebaud 2016-07-27 15:07:35 +02:00
  • 8972323c08 Big 1261: add missing max(ADS,ADS) overload (same for min) Gael Guennebaud 2016-07-27 14:52:48 +02:00
  • 5d94dc85e5 bug #1260: add regression test Gael Guennebaud 2016-07-27 14:38:30 +02:00
  • 0d7039319c bug #1260: remove doubtful specializations of ScalarBinaryOpTraits Gael Guennebaud 2016-07-27 14:35:52 +02:00
  • d3d7c6245d Add brackets to block matrix and fixed some typos Christoph Hertzberg 2016-07-27 09:55:39 +02:00
  • 0eece608b4 Added tag 3.3-beta2 for changeset f6b3cf8de9 Gael Guennebaud 2016-07-26 23:52:14 +02:00
  • f6b3cf8de9 Bump to 3.3-beta2 3.3-beta2 Gael Guennebaud 2016-07-26 23:51:59 +02:00
  • 9d16b6e1cf Formatting Gael Guennebaud 2016-07-26 23:51:43 +02:00
  • fd2f989b1d Fix testing of nearly zero input matrices. Gael Guennebaud 2016-07-26 14:46:02 +02:00
  • c9e3e438eb Add more very small numbers in the list of nearly "zero" values when testing SVD and EVD algorithms Gael Guennebaud 2016-07-26 14:45:44 +02:00
  • 95113cb15c Improve robustness of 2x2 eigenvalue with shifting and scaling Gael Guennebaud 2016-07-26 14:43:54 +02:00
  • 7f7e84aa36 Fix compilation with MKL support Gael Guennebaud 2016-07-26 13:31:29 +02:00
  • 429028b652 Typo. Gael Guennebaud 2016-07-26 12:12:53 +02:00
  • 6b89fa802c Typos. Gael Guennebaud 2016-07-26 12:08:04 +02:00
  • c581c8fa79 Fix with expession template scalar types. Gael Guennebaud 2016-07-26 11:33:28 +02:00
  • 8021aed89e Split BLAS/LAPACK versus MKL documentation Gael Guennebaud 2016-07-26 11:11:59 +02:00
  • 757971e7ea bug #1258: fix compilation of Map<SparseMatrix>::coeffRef Gael Guennebaud 2016-07-26 09:40:19 +02:00
  • c9425492c8 Update doc. Gael Guennebaud 2016-07-25 18:41:26 +02:00
  • 0592b4cfbf merge Gael Guennebaud 2016-07-25 18:20:22 +02:00
  • 9c663e4ee8 Clean references to MKL in LAPACKe support. Gael Guennebaud 2016-07-25 18:20:08 +02:00
  • 0c06077efa Rename MKL files Gael Guennebaud 2016-07-25 18:00:47 +02:00
  • 4d54e3dd33 bug #173: remove dependency to MKL for LAPACKe backend. Gael Guennebaud 2016-07-25 17:55:07 +02:00
  • 3d3d34e442 Deleted dead code. Benoit Steiner 2016-07-25 08:53:37 -07:00
  • 34b483e25d bug #1249: enable use of __builtin_prefetch for GCC, clang, and ICC only. Gael Guennebaud 2016-07-25 15:17:45 +02:00
  • 6d5daf32f5 bug #1255: comment out broken and unsused line. Gael Guennebaud 2016-07-25 14:48:30 +02:00
  • f9598d73b5 bug #1250: fix pow() for AutoDiffScalar with custom nested scalar type. Gael Guennebaud 2016-07-25 14:42:19 +02:00
  • fd1117f2be Implement digits10 for mpreal Gael Guennebaud 2016-07-25 14:38:55 +02:00
  • 9908020d36 Add minimal support for Array<string>, and fix Tensor<string> Gael Guennebaud 2016-07-25 14:25:56 +02:00
  • 4184a3e544 Extend boost.multiprec unit test with ET on, complexes, and general/generalized eigenvalue solvers. Gael Guennebaud 2016-07-25 12:36:22 +02:00
  • 1b2049fbda Enforce scalar types in calls to max/min (helps with expression template scalar types) Gael Guennebaud 2016-07-25 12:35:10 +02:00
  • b118bc76eb Add digits10 overload for complex. Gael Guennebaud 2016-07-25 12:33:21 +02:00
  • c96af5381f Remove custom complex division function cdiv. Gael Guennebaud 2016-07-25 12:31:58 +02:00
  • e1c7c5968a Update doc. Gael Guennebaud 2016-07-25 11:18:04 +02:00
  • 8fffc81606 Add NumTraits::digits10() function based on numeric_limits::digits10 and make use of it for printing matrices. Gael Guennebaud 2016-07-25 11:13:01 +02:00
  • 5f03584752 merge Gael Guennebaud 2016-07-23 17:52:44 +02:00
  • 1b0353c659 Fix misuse of dummy_precesion in eigenvalues solvers Gael Guennebaud 2016-07-23 17:52:31 +02:00
  • c6b0de2c21 Improved partial reductions in more cases Benoit Steiner 2016-07-22 17:18:20 -07:00
  • 72744d93ef Allows the compiler to inline outer products (the change from default to dont-inline in changeset 737bed19c1 was not motivated) Gael Guennebaud 2016-07-22 17:02:28 +02:00
  • 32d95e86c9 merge Gael Guennebaud 2016-07-22 16:43:12 +02:00
  • 60d5980a41 add a note Gael Guennebaud 2016-07-22 15:46:23 +02:00
  • d7a0e52478 Fix testing of log nearby 1 Gael Guennebaud 2016-07-22 15:44:26 +02:00
  • 7acf23c14c Truely split unit test. Gael Guennebaud 2016-07-22 15:41:23 +02:00
  • 24af67a6cc Fix boostmultiprec for C++03 Gael Guennebaud 2016-07-22 15:30:54 +02:00
  • 395c835f4b Fix CUDA compilation Gael Guennebaud 2016-07-22 15:30:24 +02:00
  • d075d122ea Move half unit test from unsupported to main tests Gael Guennebaud 2016-07-22 14:34:19 +02:00
  • 47afc9a365 More cleaning in half: - put its definition and functions in its own half_impl namespace such that the free function does not polute the Eigen namespace while still making them visible for half through ADL. - expose Eigen::half throguh a using statement - move operator<< from std to half_float namespace Gael Guennebaud 2016-07-22 14:33:28 +02:00
  • 0f350a8b7e Fix CUDA compilation Gael Guennebaud 2016-07-21 18:47:07 +02:00
  • bf91a44f4a Use ADL and log10 for printing matrices. Gael Guennebaud 2016-07-21 15:48:24 +02:00
  • 82798162c0 Extend unit testing of half with ADL and arrays. Gael Guennebaud 2016-07-21 15:47:21 +02:00
  • 87fbda812f Add missing log10 and random generator for half. Gael Guennebaud 2016-07-21 15:46:45 +02:00
  • 01d12d3e82 Some cleanup in Halh: standard functions should be defined in the namespace of the class half to make ADL work, and thus the global is* functions can be removed. Gael Guennebaud 2016-07-21 15:10:48 +02:00
  • 007edee1ac Add a doc page summarizing the true speed of Eigen's decompositions. Gael Guennebaud 2016-07-21 12:32:02 +02:00
  • 9b76be9d21 Update benchmark for dense solver to stress least-squares pb, and to output a HTML table Gael Guennebaud 2016-07-21 12:30:53 +02:00
  • 72950effdf enable testing of Boost.Multiprecision with expression templates Gael Guennebaud 2016-07-20 18:21:30 +02:00
  • 7b4abc2b1d Fixed a code comment error Yi Lin 2016-07-20 22:28:54 +08:00
  • b64b9d0172 Add a unit test to stress our solvers with Boost.Multiprecision Gael Guennebaud 2016-07-20 15:20:14 +02:00
  • 5e4dda8a12 Enable custom scalar types in some unit tests. Gael Guennebaud 2016-07-20 15:19:17 +02:00
  • 87d480d785 Make use of EIGEN_TEST_MAX_SIZE Gael Guennebaud 2016-07-20 15:14:20 +02:00
  • 7722913475 Fix ambiguous specialization with custom scalar type Gael Guennebaud 2016-07-20 15:13:44 +02:00
  • fd057f86b3 Complete the coeff-wise math function table. Gael Guennebaud 2016-07-20 12:14:10 +02:00
  • 9e8476ef22 Add missing Eigen::rsqrt global function Gael Guennebaud 2016-07-20 11:59:49 +02:00
  • 4b4c296d6e Simplify ScalarBinaryOpTraits by removing the Defined enum, and extend its documentation. Gael Guennebaud 2016-07-20 09:56:39 +02:00
  • e3bf874c83 Workaround MSVC 2010 compilation issue. Gael Guennebaud 2016-07-18 15:17:25 +02:00
  • 0f89c6d6b5 Add a summary of possible values for EIGEN_COMP_MSVC Gael Guennebaud 2016-07-18 15:16:13 +02:00
  • 18884f17d7 Remove static constant declaration: this enforces compiler to generate costly code for thread safety. Gael Guennebaud 2016-07-18 15:05:17 +02:00
  • 79574e384e Make scalar_product_op the default (instead of void) Gael Guennebaud 2016-07-18 12:03:05 +02:00
  • 6a3c451c1c Permits call to explicit ctor. Gael Guennebaud 2016-07-18 12:02:20 +02:00
  • 0c3fe4aca5 merge Gael Guennebaud 2016-07-18 10:44:15 +02:00