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
6,705 Commits 15 Branches 93 Tags
2afdef6a54e7fd09a4a6d5e933cf15ffa129beb6
Commit Graph

64 Commits

Author SHA1 Message Date
Gael Guennebaud
2afdef6a54 Generalize first_aligned to take the requested alignment as a template parameter, and add a first_default_aligned variante calling first_aligned with the requirement of the largest packet for the given scalar type. 2015-08-06 17:52:01 +02:00
Gael Guennebaud
1f5024332e First part of a big refactoring of alignment control to enable the handling of arbitrarily aligned buffers. It includes: 
- AlignedBit flag is deprecated. Alignment is now specified by the evaluator through the 'Alignment' enum, e.g., evaluator<Xpr>::Alignment. Its value is in Bytes.
 - Add several enums to specify alignment: Aligned8, Aligned16, Aligned32, Aligned64, Aligned128. AlignedMax corresponds to EIGEN_MAX_ALIGN_BYTES. Such enums are used to define the above Alignment value, and as the 'Options' template parameter of Map<> and Ref<>.
 - The Aligned enum is now deprecated. It is now an alias for Aligned16.
 - Currently, traits<Matrix<>>, traits<Array<>>, traits<Ref<>>, traits<Map<>>, and traits<Block<>> also expose the Alignment enum.
 2015-08-06 15:31:07 +02:00
Christoph Hertzberg
ea87561564 bug #1039: Redefining EIGEN_DEFAULT_DENSE_INDEX_TYPE may lead to errors 2015-07-13 16:08:25 +02:00
Gael Guennebaud
36643eec0c Add a call_assignment_no_alias_no_transpose shortcut 2015-06-24 17:50:43 +02:00
Gael Guennebaud
972a535288 Remove aligned-on-scalar assert and fallback to non vectorized path at runtime (first_aligned already had this runtime guard) 2015-06-14 15:04:07 +02:00
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
Gael Guennebaud
ee06f78679 Introduce unified macros to identify compiler, OS, and architecture. They are all defined in util/Macros.h and prefixed with EIGEN_COMP_, EIGEN_OS_, and EIGEN_ARCH_ respectively. 2014-11-04 21:58:52 +01:00
Christoph Hertzberg
d3f52debc6 Make cuda_basic test compile again by adding lots of EIGEN_DEVICE_FUNC. 
Although the test passes now, there might still be some missing.
 2014-10-13 17:18:26 +02:00
Georg Drenkhahn
edaefeb978 Using Kernel::Index type instead of int to prevent possible implicit conversion from long to int. 2014-09-21 10:01:12 +02:00
Gael Guennebaud
0ca43f7e9a Remove deprecated code not used by evaluators 2014-09-18 15:15:27 +02:00
Gael Guennebaud
c83e01f2d6 Favor column major storage for inner products 2014-09-14 19:38:49 +02:00
Gael Guennebaud
26db954776 Re-enable aliasing checks when using evaluators 2014-09-14 19:06:08 +02:00
Gael Guennebaud
9b729f93a1 Resizing is done by call_assignment_noalias, so no need to perform it when dealing with aliasing. 2014-07-21 11:46:47 +02:00
Gael Guennebaud
62f332fc04 Make sure we evaluate into temporaries matching evaluator storage order requirements 2014-07-19 15:19:10 +02:00
Gael Guennebaud
b868bfb84a Make operator=(EigenBase<>) uses the new assignment mechanism and introduce a generic EigenBase to EigenBase assignment kind based on the previous evalTo mechanism. 2014-06-25 17:23:52 +02:00
Gael Guennebaud
3849cc65ee Implement binaryop and transpose evaluators for sparse matrices 2014-06-23 10:40:03 +02:00
Gael Guennebaud
5e26b7cf9d Extend evaluation traits debuging info 2014-03-12 18:13:18 +01:00
Gael Guennebaud
a6be1952f4 Fix a few regression when moving the flags 2014-03-12 16:18:34 +01:00
Gael Guennebaud
8dd3b716e3 Move evaluation related flags from traits to evaluator and fix evaluators of MapBase and Replicate 2014-03-12 13:34:11 +01:00
Gael Guennebaud
da6ec81282 Move CoeffReadCost mechanism to evaluators 2014-03-10 23:24:40 +01:00
Gael Guennebaud
af31b6c37a Generalize evaluator<Inverse<>> such that there is no need to specialize it 2014-02-21 15:22:08 +01:00
Gael Guennebaud
93125e372d Port LU module to evaluators (except image() and kernel()) 2014-02-20 15:26:15 +01:00
Gael Guennebaud
2eee6eaf3c Fix mixing scalar types with evaluators 2014-02-19 16:30:17 +01:00
Gael Guennebaud
68e8ddaf94 Fix vectorization logic wrt assignment functors 2014-02-19 13:26:07 +01:00
Gael Guennebaud
82c066b3c4 Cleaning 2014-02-18 15:44:32 +01:00
Gael Guennebaud
8cfb138e73 Finally, the simplest remains to deffer resizing at the latest 2014-02-18 13:31:44 +01:00
Gael Guennebaud
573c587e3d New design for handling automatic transposition 2014-02-18 10:53:14 +01:00
Gael Guennebaud
d595fd31f5 Deal with automatic transposition in call_assignment, fix a few shortcomings 2014-02-17 16:11:55 +01:00
Gael Guennebaud
ee1c55f923 Add missing template keyword 2014-01-26 14:55:25 +01:00
Gael Guennebaud
5fa7262e4c Refactor triangular assignment 2014-01-25 23:02:14 +01:00
Gael Guennebaud
34ca81b1bf Add direct assignment of products 2013-12-02 16:37:58 +01:00
Gael Guennebaud
c6f7337032 Get rid of call_dense_swap_loop 2013-12-02 14:44:13 +01:00
Gael Guennebaud
27ca9437a1 Fix usage of Dense versus DenseShape 2013-12-02 14:05:34 +01:00
Gael Guennebaud
c15c65990f First step toward the generalization of evaluators to triangular, sparse and other fancyness. 
Remove product_tag template parameter to Product.
 2013-11-29 17:50:59 +01:00
Gael Guennebaud
5584275325 Remove HasEvalTo and all at once eval mode 2013-11-29 13:38:59 +01:00
Gael Guennebaud
230f5c3aa9 Evaluator: introduce the main Assignment class, add call_assignment to bypass NoAlias and AssumeAliasing, and some bits of cleaning 2013-11-25 15:20:31 +01:00
Gael Guennebaud
af9851d1d7 bug #99: move the creation of the evaluator to a central place, and make generic_dense_assignment_kernel hold the destination and source evaluators 2013-11-07 12:03:12 +01:00
Gael Guennebaud
8edc964734 bug #99: refactor assignment and compound assignment mechanism through "assignment functors" and "assignement kernels". 
The former is very low level and generic. The later abstarct the former for dense expressions. This refactoring permits
to get rid of the very ugly SwapWrapper and SelfCwiseBinaryOp classes.
In the future, this will also permit to simplify all these evaluation loops and perhaps to reuse them for reduxions.
That will also permit to specialize for operations like expr1 += expr2 outside Eigen, and so for any kind
of expressions (dense, sparse, tensor, etc.)
 2013-11-06 18:17:59 +01:00
Gael Guennebaud
a37bdfc955 Fix static/inline order 2013-11-06 11:13:31 +01:00
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
0fac91ac22 Fix "storage class is not first" warnings 2013-02-28 19:27:53 +01:00
Gael Guennebaud
fa17a6da75 Fix compilation with ICC that was unable to instanciate first_aligned 2013-02-26 17:32:42 +01:00
Benoit Jacob
69124cfca2 Automatic relicensing to MPL2 using Keirs script. Manual fixup follows. 2012-07-13 14:42:47 -04:00
Jitse Niesen
746378868a Implement A.noalias() = B * C without temporaries 
* Wrap expression inside EvalToTemp in copy_using_evaluators() if we
  assume aliasing for that expression (that is, for products)
* Remove temporary kludge of evaluating expression to temporary in
  AllAtOnce traversal
* Implement EvalToTemp expression object
 2012-06-29 13:54:09 +01:00
Jitse Niesen
d0b873822f Make product eval-at-once. 
* Make product EvalAtOnce in cases OuterProduct, GemmProduct and
  GemvProduct
* Ensure that product evaluators are nested inside EvalToTemp
  evaluator
* As temporary kludge, evaluate expression to temporary in AllAtOnce
  traversal and pass expression operator to evalTo()
 2012-06-29 13:49:25 +01:00
Jitse Niesen
2393ceb380 Implement eval-at-once in evaluator. 
- Add evaluator_traits with HasEvalTo flag, which is true if evaluator
  has evalTo() function.
- Add AllAtOnce traversal, which calls evalTo() in evaluator.
- If source evaluator in copy_using_evaluator has HasEvalTo set, then
  use AllAtOnce traversal.
 2012-06-29 13:32:12 +01:00
Jitse Niesen
c1eb820e50 Implement interface for NoAlias assignments. 
* Rename the old copy_using_evaluators to noalias_copy_using_evaluators.
* Write a new copy_using_evaluators which strips NoAlias expression, if present,
  and calls noalias_copy_using_evaluators; in future, it will also take care of
  aliasing in products.
* Add expression() getter to NoAlias.
 2012-06-29 13:24:04 +01:00
Jitse Niesen
23184527fa Resize lhs automatically in copy_using_evaluator(). 2012-06-28 15:25:25 +01:00
Jitse Niesen
d0d077b212 Fix bug in evaluators with sliced vectorization. 2012-06-24 17:33:21 +01:00
Gael Guennebaud
cc6dd55028 put the resurected files into the Eigen namespace 2012-06-22 16:35:20 +02:00