mirror/eigen
2
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-01-06 14:14:46 +08:00
Code Issues Packages Projects Releases Wiki Activity
5,412 Commits 11 Branches 89 Tags 168 MiB
75e574275c
Commit Graph

258 Commits

Author SHA1 Message Date
Benoit Jacob
11402edfd3 with old gcc (bug #89), only include iostream in debug mode 2011-02-16 12:01:47 -05:00
Benoit Jacob
59596efdf7 Fix bug #89: on GCC <= 4.3, use a custom assert implementation to work around a compiler bug 2011-02-16 08:50:19 -05:00
Benoit Jacob
63626bb966 remove debug #error 2011-02-09 14:37:52 -05:00
Benoit Jacob
85f9fab003 back out changeset efdf2e4056 
. It turns out that the SSE3 header is always included, even without any SSE enabled, so it was making us wrongly use SSE3 paths. Backing this out fixes msvc related crashes, at least bug #165.
 2011-02-09 14:01:26 -05:00
Benoit Jacob
ba9f6a2c3b now random<integer types> spans over 0..RAND_MAX, or -RAND_MAX/2..RAND_MAX/2 for signed types, or the most significant bits for smaller integer types. 2011-02-07 10:55:41 -05:00
Benoit Jacob
dc22ae101f kill stage 15, it's useless 2011-01-31 09:18:49 -05:00
Benoit Jacob
fd400ffffb reverse order of testing for eigen2 support stages. Higher stages now have priority. So if your whole project builds with say stage 10, you can manually enable stage 20 for selected files. 2011-01-27 10:34:44 -05:00
Benoit Jacob
8df5bca979 rename build stages to multiples of 10; old stage 2 becomes stage 15, while stage 20 generates errors (instead of warnings) on conflicting API. 2011-01-23 18:22:18 -05:00
Benoit Jacob
30de1651d3 relax Map const correctness in eigen2 support stages <= 3 
introduce new 'strict' stage 4
 2011-01-21 10:42:19 -05:00
Benoit Jacob
cc2b7a5397 introduce the 3 stages of eigen2 support, writing to the mailing list about that in Eigen2 to Eigen3 Migration Path thread 2011-01-21 09:51:03 -05:00
Jitse Niesen
8db9acbc16 Move doxygen comments for EIGEN_NO_DEBUG from source to I14. 
This reverts changeset 76fbe94279
. Benoit and I agree that my
approach there (to use doxygen comments) pollutes the code too much.
 2010-12-27 15:07:11 +00:00
Hauke Heibel
efdf2e4056 Added automatic SSE3/4.1/4.2 support for MSVC. 2010-12-16 20:08:22 +01:00
Hauke Heibel
6f5c45ceff Fixed ctor from const raw data for Matrices and added the missing implementation for Arrays. 
Fixed a warning regarding the conversion from int to bool in MapBase.
 2010-12-15 15:19:51 +01:00
Benoit Jacob
819bcbed19 fix comment 2010-12-07 02:17:15 -05:00
Eamon Nerbonne
7a7ca99a31 [mq]: Mingw32 fix 
intrin.h is not required nor supported by mingw32.  It is present (and supported) on mingw-w64 builds, even those for 32-bit systems, but here too it's not required on 32-bit systems.  So if we're on mingw, and it's 64-bit, then and only then is the intrin.h inclusion necessary.
 2010-12-03 23:24:06 +01:00
Benoit Jacob
a3f214ade9 holy crap, i had disabled all static asserts in 71f023de3e 2010-11-22 08:21:30 -05:00
Gael Guennebaud
2577ef90c0 generalize our internal rank K update routine to support more general A*B product while evaluating only one triangular part and make it available via, e.g.: 
R.triangularView<Lower>() += s * A * B;
 2010-11-10 18:58:55 +01:00
Gael Guennebaud
0e6c1170ab trsv: add support for inner-stride!=1, reduce code instanciation, move implementation to a new products/XX.h file 2010-11-05 12:43:14 +01:00
Hauke Heibel
5d4ff3f99c Fixed bug #95 by changing _M_IX64 to _M_X64 as proposed by Jan Schlicht. 2010-10-27 11:07:38 +02:00
Benoit Jacob
8c17fab8f5 renaming: ei_matrix_storage -> DenseStorage 
DenseStorageBase  -> PlainObjectBase
 2010-10-20 09:34:13 -04:00
Gael Guennebaud
a76ce042e6 MSVC for windows mobile does not have the errno.h file 2010-10-07 18:09:15 +02:00
Gael Guennebaud
af22364988 an attempt to fix compilation on windows mobile 2010-10-07 17:54:46 +02:00
Benoit Jacob
71f023de3e fix compilation on ubuntu 9.04's version of gcc 4.3 (yes, wtf) 2010-09-27 09:57:57 -04:00
Eamon Nerbonne
d17bb02ccd Fixes mingw32 compile issues 2010-09-02 10:38:23 +02:00
Jitse Niesen
76fbe94279 Document EIGEN_NO_DEBUG macro. 
I needed some doxygen tricks to get this to work, so it may not be worth it.
 2010-08-10 11:37:23 +01:00
Gael Guennebaud
76eb9c9fd9 fix compilation by including file in correct order 2010-07-19 23:32:13 +02:00
Konstantinos Margaritis
6ad3f1ab1f Added NEON/Complex.h, ~3.5x faster than scalar std::complex<float> 
minor fix in AltiVec Complex.h
 2010-07-10 00:09:29 +03:00
Konstantinos Margaritis
f6bd508351 forgot to add the Complex.h include for AltiVec. 2010-07-09 17:56:53 +03:00
Gael Guennebaud
e1eccfad3f add intitial support for the vectorization of complex<float> 2010-07-05 16:18:09 +02:00
Gael Guennebaud
28e64b0da3 email change 2010-06-24 23:21:58 +02:00
Hauke Heibel
4bac6fbe1e The intrin.h header needs to be included after cmath in order to prevent warnigns. 
Fixed (hopefully) final Index realted warnings.
 2010-06-21 18:39:24 +02:00
Hauke Heibel
80b6e5f278 Added include reuqired for __cpuid. 2010-06-21 16:43:31 +02:00
Gael Guennebaud
7fd8418b19 finish to merge Array into Core: 
- mv Array/* into Core/
- merge Functors.h files, and move Norms.h into Dot.h
 2010-06-19 23:36:38 +02:00
Gael Guennebaud
bfeba41174 Add a Transpositions class to ease the representation and 
manipulation of permutations as a sequence of transpositions.
Make LDLT use it.
 2010-06-04 23:17:57 +02:00
Thomas Capricelli
742bbdfa57 clang/llvm is now good enough. I can compile a project using those (one of 
the binary segfaults though, and i think it's related..)
 2010-05-21 02:03:43 +02:00
Benoit Jacob
0928c40f68 rename Coeffs.h -> DenseCoeffsBase.h 2010-05-08 16:02:13 -04:00
Benoit Jacob
bfdc1c4973 introduce DenseCoeffsBase: this is where the coeff / coeffRef / etc... methods go. 
Rationale: coeffRef() methods should only exist when we have DirectAccess. So a natural thing to do would have been to use enable_if, but since there are many methods it made more sense to do the "enable_if" for the whole group by introducing a new class. And that also that the benefit of not changing method prototypes.
 2010-05-08 13:45:31 -04:00
Carlos Becker
34b3cdb82c Added EIGEN_DONT_PARALLELIZE preprocessor directive 2010-04-26 17:08:54 +02:00
Benoit Jacob
58e7297859 * remove class DenseDirectAccessBase 
* remove member XprBase typedefs, use ei_dense_xpr_base
* remove member _HasDirectAccess typedefs, use ei_has_direct_access
 2010-04-23 10:27:15 -04:00
Benoit Jacob
a4f9ca44ab add minor to Eigen2Support 2010-04-22 20:49:01 -04:00
Benoit Jacob
2362eadcdd remove Minor 
adapt 3x3 and 4x4 (non-SSE) inverse paths
 2010-04-22 20:40:31 -04:00
Benoit Jacob
ff6a46105d * Refactoring of the class hierarchy: introduction of DenseDirectAccessBase, removal of extra _Base/_Options template parameters. 
* Introduction of strides-at-compile-time so for example the optimized code really knows when it needs to evaluate to a temporary
* StorageKind / XprKind
* Quaternion::setFromTwoVectors: use JacobiSVD instead of SVD
* ComplexSchur: support the 1x1 case
 2010-04-16 10:13:32 -04:00
Benoit Jacob
c4f8afdf49 #undef minor at the right place 2010-03-06 14:44:57 -05:00
Benoit Jacob
2bd31d3fbc * include Macros.h much earlier: since it takes care of the alignment platform detection, it is needed before we do the vectorization stuff in Eigen/Core !! 
* kill EIGEN_DONT_ALIGN_HEAP option (one should use EIGEN_DONT_ALIGN)
* rename EIGEN_DONT_ALIGN_STACK to EIGEN_DONT_ALIGN_STATICALLY. hope it's a better name.
 2010-03-06 09:05:15 -05:00
Gael Guennebaud
cefd9b8888 merge with default branch 2010-03-04 18:47:52 +01:00
Konstantinos Margaritis
710bc073a7 arm_neon.h is a standard header file, fixed 2010-03-03 12:15:34 -06:00
Gael Guennebaud
7dd81aad74 factorize default performance related settings to a single file 
included after the architecture specific files such that they
can be adapted by each platform.
 2010-03-03 18:47:58 +01:00
Konstantinos Margaritis
112c550b4a Added initial NEON support, most tests pass however we had to use some hackish workarounds 
as gcc on ARM (both CodeSourcery 4.4.1 used and experimental 4.5) fail to
ensure proper alignment with __attribute__((aligned(16))). This has to be
fixed upstream to remove the workarounds.
 2010-03-03 11:25:41 -06:00
Hauke Heibel
ff8c2149c1 Added a generic reallocation implementation based on ei_aligned_malloc/_free. 
Rewrote ei_handmade_aligned_realloc such that it is now using std::realloc.
Reorganized functions in Memory.h for better readability.
Add missing <cerrno> include to Core (it's now required in Memory.h).
 2010-02-28 14:32:57 +01:00
Benoit Jacob
27f5250258 Only include <iosfwd> unless either EIGEN_DEBUG_ASSIGN is defined or we're in eigen2 support mode 2010-02-27 19:04:22 -05:00
Benoit Jacob
b1c6c215a4 merge 2010-02-25 21:07:30 -05:00
Benoit Jacob
77c922bf05 * move the 's': InstructionsSet ---> InstructionSets 
* proper capitalization: SSE, AltiVec
 2010-02-25 06:43:45 -05:00
Thomas Capricelli
50a5ac3c4b oops, fix typo 2010-02-25 05:31:22 +01:00
Thomas Capricelli
00bc535b66 provide a static method to describe which SIMD instructions are used 2010-02-24 21:52:08 +01:00
Thomas Capricelli
0f3d69b65e Provide "eigen" defines to decide which instruction set is used 
(sse3, ssse3 and sse4), independantly from the compiler.
Only those defines should be used in other places, and the user can
rely on those to know which sets are used.
 2010-02-24 21:43:30 +01:00
Gael Guennebaud
3e62fafce8 clean a bit the parallelizer 2010-02-22 11:08:37 +01:00
Hauke Heibel
f0c8dcf1e2 Renamed AnyMatrixBase to EigenBase. 2010-02-20 15:26:02 +01:00
Benoit Jacob
b73e22905d miserable half-working state, commiting to a fork just in case, just to perfect 
my day, my hard disk would die.
Will write a more detailed commit message once it's working.
 2010-02-18 20:42:38 -05:00
Benoit Jacob
9251cfed9b this had to be done here, not at the end. 2010-02-12 09:03:16 -05:00
Benoit Jacob
37ca4200b2 Piotr's patch was missing many occurences of size_t. So, 
using std::size_t;
This is the only way that we can ensure QCC support in the long term without having to think about it everytime.
 2010-02-12 08:58:29 -05:00
Gael Guennebaud
5686eca7b1 * fix multiple temporary copies for coeff based products 
* introduce a lazy product version of the coefficient based implementation
  => flagged is not used anymore
  => small outer product are now lazy by default (aliasing is really unlikely for outer products)
 2010-02-09 11:05:39 +01:00
Gael Guennebaud
73eb0e633c * resurected Flagged from Eigen2Support 
* reimplement .diagonal() for ProductBase to make (A*B).diagonal() more efficient!
 2010-02-04 18:28:09 +01:00
Gael Guennebaud
0ce5bc0d14 add support for global math function for array 2010-01-27 23:23:59 +01:00
Hauke Heibel
5d48cc1f5b Moved the Array module to Core. 2010-01-20 20:51:01 +01:00
Gael Guennebaud
c5d7c9f0de remove the Triangular suffix to Upper, Lower, UnitLower, etc, 
and remove the respective bit flags
 2010-01-07 21:15:32 +01:00
Gael Guennebaud
9d9e00b608 merge and add start/end to Eigen2Support 2010-01-05 13:07:32 +01:00
Thomas Capricelli
57275b2b8c make some changes to please clang, fix some warnings too. 2010-01-04 23:21:04 +01:00
Gael Guennebaud
eaaba30cac merge with default branch 2009-12-22 22:51:08 +01:00
Gael Guennebaud
ebb2878829 finally add a Array class with storage via the introduction of a DenseStorageBase 
base class shared by both Matrix and Array
 2009-12-17 13:37:00 +01:00
Gael Guennebaud
8e05f9cfa1 add a DenseBase class for MAtrixBase and ArrayBase and more code factorisation 2009-12-04 23:17:14 +01:00
Mark Borgerding
ff1e9542f6 added comments to help vim understand the header files are c++. 2009-12-01 18:00:29 -05:00
Benoit Jacob
684d76eba3 add SSE4 support, start with integer multiplication 2009-11-24 15:12:43 -05:00
Gael Guennebaud
4af1753b6f * remove EnforceAlignedAccess option to Block, VectorBlock, Map and MapBase 
because thanks to the previous commit this is not needed anymore
* add a more general ForceAlignedAccess expression which can be used for any expression.
  It is already used by StableNorm.h.
 2009-11-20 16:30:14 +01:00
Gael Guennebaud
eb8f450071 Hey, finally the copyCoeff stuff is not only used to implement swap anymore :) 
Add an internal pseudo expression allowing to optimize operators like +=, *= using
the copyCoeff stuff.
This allows to easily enforce aligned load for the destination matrix everywhere.
 2009-11-20 15:39:38 +01:00
Gael Guennebaud
e3d890bc5a Another big refactoring change: 
* add a new Eigen2Support module including Cwise, Flagged, and some other deprecated stuff
* add a few cwiseXxx functions
* adapt a few modules to use cwiseXxx instead of the .cwise() prefix
 2009-11-18 18:15:19 +01:00
Gael Guennebaud
1e62e0b0d8 more ET refactoring: 
* extend Cwise for multiple storage base class
* a lot of cleaning in the Sparse module
 2009-11-17 16:04:19 +01:00
Benoit Jacob
955cd7f884 * add PermutationMatrix 
* DiagonalMatrix:
   - add MaxSizeAtCompileTime parameter
   - DiagonalOnTheLeft ---> OnTheLeft
   - fix bug in DiagonalMatrix::setIdentity()
 2009-11-15 21:12:15 -05:00
Gael Guennebaud
7b0c4102fa * add a Make* expression type builder to allow the 
construction of generic expressions working
  for both dense and sparse matrix. A nicer solution
  would be to use CwiseBinaryOp for any kind of matrix.
  To this end we either need to change the overall design
  so that the base class(es) depends on the kind of matrix,
  or we could add a template parameter to each expression
  type (e.g., int Kind = ei_traits<MatrixType>::Kind)
  allowing to specialize each expression for each kind of matrix.
* Extend AutoDiffScalar to work with sparse vector expression
  for the derivatives.
 2009-10-16 13:22:38 +02:00
Benoit Jacob
6b5f96cb03 undef B0 2009-09-19 19:14:28 -04:00
Gael Guennebaud
49dd5d7847 * add a HouseholderSequence class (not good enough yet for Triadiagonalization and HessenbergDecomposition) 
* rework a bit AnyMatrixBase, and mobe it to a separate file
 2009-09-16 14:35:42 +02:00
Gael Guennebaud
239ada95b7 add overloads of lazyAssign to detect common aliasing issue with 
transpose and adjoint
 2009-08-15 22:19:29 +02:00
Gael Guennebaud
50c703f0c7 As proposed on the list: 
- rename EvalBeforeAssignBit to MayAliasBit
- make .lazy() remove the MayAliasBit only, and mark it as deprecated
- add a NoAlias pseudo expression, and MatrixBase::noalias() function
Todo:
- we have to decide whether += and -= assume no aliasing by default ?
- once we agree on the API: update the Sparse module and the unit tests respectively.
 2009-08-15 18:35:51 +02:00
Gael Guennebaud
8abec72259 oops forgot to remove the #include in Core 2009-08-14 09:49:33 +02:00
Gael Guennebaud
1b257a7620 add an optimized "apply in place a rotation in the plane", 
and make Jacobi and SelfAdjointEigenSolver use it
=> ~ x1.75 speedup for JacobiSVD and x2 for SelfAdjointEigenSolver
 2009-08-13 11:42:02 +02:00
Gael Guennebaud
56d00779db more product refactoring 2009-08-06 12:20:02 +02:00
Gael Guennebaud
7d607048a9 implement a ProductBase class and, as a proof of concept, update TriangularProduct 
and SelfAdjointMatrixProduct to take advantage of it => fewer LOC
 2009-08-04 16:54:17 +02:00
Gael Guennebaud
54804eb626 synch with main branch 2009-07-28 17:35:07 +02:00
Gael Guennebaud
6aba84719d trmm is now working in all storage order configurations 2009-07-27 10:27:01 +02:00
Gael Guennebaud
f4112dcff3 The new trsm is working very very well (read very fast) for 
lower triangular matrix and row or col major lhs.
TODO: handle upper triangular and row major rhs cases
 2009-07-25 21:41:01 +02:00
Gael Guennebaud
a81388fae9 Implement efficient sefladjoint product (aka SYRK) : C += alpha * U U^T 
It is currently available via SelfAdjointView::rankKupdate.
TODO: allows to write SelfAdjointView += u * u.adjoint()
 2009-07-23 19:01:20 +02:00
Gael Guennebaud
d6475ea390 more refactoring in the level3 products 2009-07-22 11:54:58 +02:00
Gael Guennebaud
d6627d540e * refactoring of the matrix product into multiple small kernels 
* started an efficient selfadjoint matrix * general matrix product
  based on the generic kernels ( => need a very little LOC)
 2009-07-21 16:58:35 +02:00
Gael Guennebaud
32b08ac971 re-implement stableNorm using a homemade blocky and 
vectorization friendly algorithm (slow if no vectorization)
 2009-07-17 16:22:39 +02:00
Gael Guennebaud
587029a612 started an implementation of BandMatrix: at least the read/write access 
to the main/sub/super diagonals seems to work well.
 2009-07-14 23:27:37 +02:00
Gael Guennebaud
8120a5cecd synch with main devel branch 2009-07-14 23:06:25 +02:00
Gael Guennebaud
279cedc1ce some cleaning/renaming is Triangular/SelfadjointView 2009-07-14 22:38:21 +02:00
Gael Guennebaud
a2cf7ba955 add triangular * vector product 2009-07-13 13:17:55 +02:00
Gael Guennebaud
a2087cd7a3 Add an efficient rank2 update function (like the level2 blas xSYR2 routine). 
Note that it is already used in Tridiagonalization.
 2009-07-11 21:14:59 +02:00
Gael Guennebaud
ec5c608aa3 Set of fixes and workaround to make sun studio more happy. 
Still remains the problem of alignment and vectorization.
 2009-07-10 16:10:03 +02:00
Gael Guennebaud
1aea45335f * bybye Part, welcome TriangularView and SelfAdjointView. 
* move solveTriangular*() to TriangularView::solve*()
* move .llt() to SelfAdjointView
* add a high level wrapper to the efficient selfadjoint * vector product
* improve LLT so that we can specify which triangular part is meaningless
=> there are still many things to do (doc, cleaning, improve the matrix products, etc.)
 2009-07-06 23:43:20 +02:00