GitHub-Proxy/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-10-14 00:51:28 +08:00
Code Issues Packages Projects Releases Wiki Activity
12,636 Commits 13 Branches 92 Tags
Commit Graph

287 Commits

Author SHA1 Message Date
Rasmus Munk Larsen
b6fdf7468c Rename inverse -> pseudoInverse. 2016-02-10 13:03:07 -08:00
Rasmus Munk Larsen
bb8811c655 Enable inverse() method for computing pseudo-inverse. 2016-02-09 20:35:20 -08:00
Rasmus Munk Larsen
53f60e0afc Make applyZAdjointOnTheLeftInPlace protected. 2016-02-08 09:01:43 -08:00
Rasmus Munk Larsen
414efa47d3 Add missing calls to tests of COD. 
Fix a few mistakes in 3.2 -> 3.3 port.
 2016-02-08 08:50:34 -08:00
Rasmus Munk Larsen
16ec450ca1 Nevermind. 2016-02-06 17:54:01 -08:00
Rasmus Munk Larsen
019fff9a00 Add my name to copyright notice in ColPivHouseholder.h, mostly for previous work on stable norm downdate formula. 2016-02-06 17:48:42 -08:00
Rasmus Munk Larsen
d904c8ac8f Implement complete orthogonal decomposition in Eigen. 2016-02-06 16:32:00 -08:00
Rasmus Munk Larsen
5fb04ab2da Fix bad line break. Don't repeat Kahan matrix test since it is deterministic. 2016-02-03 10:12:10 -08:00
Rasmus Munk Larsen
d9a6f86cc0 Make the array of directly compute column norms a member to avoid allocation in computeInPlace. 2016-02-03 09:55:30 -08:00
Rasmus Munk Larsen
acce4dd050 Change Eigen's ColPivHouseholderQR to use the numerically stable norm downdate formula from http://www.netlib.org/lapack/lawnspdf/lawn176.pdf, which has been used in LAPACK's xGEQPF and xGEQP3 since 2006. With the old formula, the code chooses the wrong pivots and fails to correctly determine rank on graded matrices. 
This change also adds additional checks for non-increasing diagonal in R11 to existing unit tests, and adds a new unit test with the Kahan matrix, which consistently fails for the original code.

Benchmark timings on Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz. Code compiled with AVX & FMA. I just ran on square matrices of 3 difference sizes.

Benchmark               Time(ns)     CPU(ns) Iterations
-------------------------------------------------------
Before:
BM_EigencolPivQR/64        53677       53627      12890
BM_EigencolPivQR/512    15265408    15250784         46
BM_EigencolPivQR/4k  15403556228 15388788368          2

After (non-vectorized version):
Benchmark               Time(ns)     CPU(ns) Iterations  Degradation
--------------------------------------------------------------------
BM_EigencolPivQR/64        63736       63669      10844         18.5%
BM_EigencolPivQR/512    16052546    16037381         43          5.1%
BM_EigencolPivQR/4k  15149263620 15132025316          2         -2.0%

Performance-wise there seems to be a ~18.5% degradation for small (64x64) matrices, probably due to the cost of more O(min(m,n)^2) sqrt operations that are not needed for the unstable formula.
 2016-01-28 15:07:26 -08:00
Gael Guennebaud
8b0d1eb0f7 Fix numerous doxygen shortcomings, and workaround some clang -Wdocumentation warnings 2016-01-01 21:45:06 +01:00
Gael Guennebaud
4519fd5d40 Fix MKL compilation issue 2015-12-11 11:11:38 +01:00
Gael Guennebaud
79c1e6d0a6 Fix compilation of MKL support. 2015-12-11 10:55:07 +01:00
Gael Guennebaud
7031a851d4 Generalize matrix ctor and compute() method of dense decomposition to 1) limit temporaries, 2) forward expressions to nested decompositions, 3) fix ambiguous ctor instanciation for square decomposition 2015-09-07 10:42:04 +02:00
Gael Guennebaud
8580eb6808 bug #949: add static assertion for incompatible scalar types in dense end-user decompositions. 2015-03-13 21:06:20 +01:00
Marc Glisse
37a93c4263 New scoring functor to select the pivot. 
This is can be useful for non-floating point scalars, where choosing the biggest element is generally not the best choice.
 2015-03-03 17:08:28 +01:00
Gael Guennebaud
98604576d1 Merged in chtz/eigen-indexconversion (pull request PR-92) 
bug #877, bug #572: Get rid of Index conversion warnings, summary of changes:

- Introduce a global typedef Eigen::Index making Eigen::DenseIndex and AnyExpr<>::Index deprecated (default is std::ptrdiff_t).

 - Eigen::Index is used throughout the API to represent indices, offsets, and sizes.

 - Classes storing an array of indices uses the type StorageIndex to store them. This is a template parameter of the class. Default is int.

 - Methods that *explicitly* set or return an element of such an array take or return a StorageIndex type. In all other cases, the Index type is used.
 2015-02-16 15:29:00 +01:00
Gael Guennebaud
cc641aabb7 Remove deprecated usage of expr::Index. 2015-02-16 14:46:51 +01:00
Christoph Hertzberg
bd511dde9d bug #952: Missing \endcode made doxygen fail to build ColPivHouseholderQR 2015-02-15 06:08:25 +01:00
Gael Guennebaud
fc202bab39 Index refactoring: StorageIndex must be used for storage only (and locally when it make sense). In all other cases use the global Index type. 2015-02-13 18:57:41 +01:00
Gael Guennebaud
fe51319980 Merge Index-refactoring branch with default, fix PastixSupport, remove some useless typedefs 2015-02-13 10:03:53 +01:00
Gael Guennebaud
f1092d2f73 bug #941: fix accuracy issue in ColPivHouseholderQR, do not stop decomposition on a small pivot 2015-01-30 19:04:04 +01:00
Christoph Hertzberg
e8cdbedefb bug #877, bug #572: Introduce a global Index typedef. Rename Sparse*::Index to StorageIndex, make Dense*::StorageIndex an alias to DenseIndex. Overall this commit gets rid of all Index conversion warnings. 2014-12-04 22:48:53 +01:00
Mark Borgerding
880e72c130 quieted more g++ warnings of the form: warning: typedef XXX locally defined but not used [-Wunused-local-typedefs] 2014-10-16 09:19:32 -04:00
Christoph Hertzberg
36448c9e28 Make constructors explicit if they could lead to unintended implicit conversion 2014-09-23 14:28:23 +02:00
Gael Guennebaud
0ca43f7e9a Remove deprecated code not used by evaluators 2014-09-18 15:15:27 +02:00
Gael Guennebaud
8846aa6d1b Optimization: enable cache-efficient application of HouseholderSequence. 2014-09-04 09:15:59 +02:00
Gael Guennebaud
85c7659574 Refactoring of sparse solvers through a SparseSolverBase class and usage of the Solve<> expression. Introduce a SolveWithGuess expression on top of Solve. 2014-09-01 15:00:19 +02:00
Gael Guennebaud
74b1d79d77 merge default and evaluator branches 2014-03-12 16:24:25 +01:00
Gael Guennebaud
9be72cda2a Port QR module to Solve/Inverse 2014-03-11 11:47:32 +01:00
Jitse Niesen
a58325ac2f Minor corrections in QR docs. 2013-12-31 18:06:28 +00:00
Christoph Hertzberg
bbf373bbe9 Applied patch from Richard JW Roberts, resolving bug #704 2013-12-21 22:14:03 +01:00
Christoph Hertzberg
6b6071866b Make pivoting HouseholderQR compatible with custom scalar types 2014-02-25 18:55:16 +01:00
Gael Guennebaud
28b2abdbea Fix FullPivHouseholderQR ctors for non squared fixed size matrix types 2013-11-19 12:53:46 +01:00
Gael Guennebaud
1bb1a57ef7 merge with default branch 2013-11-05 10:31:59 +01:00
Gael Guennebaud
bbd49d194a Add a rank method with threshold control to JacobiSVD, and make solve uses it to return the minimal norm solution for rank-deficient problems 2013-11-01 18:21:46 +01:00
Gael Guennebaud
8f496cd3a3 Fix changeset 2702788da71303afa71e53dfa6542a560115425a 
for fixed size matrices
 2013-11-01 18:17:55 +01:00
Gael Guennebaud
2702788da7 Fix bug #678: vectors of row and columns transpositions were not properly resized in FullPivQR 2013-10-29 18:02:18 +01:00
Gael Guennebaud
150c9fe536 Make FullPivHouseholderQR::solve returns the least-square solution instead of aborting if no exact solution exist 2013-08-20 11:52:48 +02:00
Gael Guennebaud
55b57fcba6 Disable some shortcuts with nvcc 2013-07-31 16:56:31 +02:00
Gael Guennebaud
7ee378d89d Fix various scalar type conversion warnings. 2013-07-12 16:40:02 +02:00
Gael Guennebaud
55365566b2 Fix HouseholderSequence::conjugate() and ::adjoint() and add respective unit tests. 2013-06-17 00:14:42 +02:00
Gael Guennebaud
62670c83a0 Fix bug #314: move remaining math functions from internal to numext namespace 2013-06-10 23:40:56 +02:00
Gael Guennebaud
0525874a03 Fix bug #599: add missing documentation of some members in QR module. 2013-06-10 11:58:28 +02:00
Gael Guennebaud
899c0c2b6c Clean source code and unit tests with respect to -Wunused-local-typedefs 2013-04-10 22:27:35 +02:00
Gael Guennebaud
5dda7842ca Add assertion on the input matrix size in factorizations relying on permutations of 32bits int 2013-02-26 11:42:32 +01:00
Gael Guennebaud
08388cc712 Remove superfluous cast. 2013-02-24 20:37:52 +01:00
Desire NUENTSA
59f9400420 Clarify the doc for column-pivoting QR 2013-02-21 13:33:31 +01:00
Desire NUENTSA
bc18e06fe3 Add matrixR() to get the triangular factor from the Householder QR 2013-02-20 13:58:26 +01:00
Gael Guennebaud
f8407742c1 Fix some implicit int64 to int conversion warnings. However, the real issue 
is that PermutationMatrix mixes the type of the stored indices and the "Index"
type used for the sizes, coeff indices, etc., which should be DenseIndex.
(transplanted from 66cbfd4d39efbec240fe90b2dd77293e71f060e8
)
 2013-02-14 18:16:51 +01:00