CFD/eigen
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
1,384 Commits 2 Branches 88 Tags 112 MiB
044dd0c1dd
Commit Graph

994 Commits

Author SHA1 Message Date
Gael Guennebaud
d34c5ef509 fix my bad fix of Hauke's fix ;) 2009-08-06 14:54:25 +02:00
Gael Guennebaud
1d4fea48b5 fix a couple of compilations issues 2009-08-06 14:10:02 +02:00
Hauke Heibel
c2861dd41a fixed inversion for AffineCompact matrices 2009-08-06 12:25:18 +02:00
Gael Guennebaud
56d00779db more product refactoring 2009-08-06 12:20:02 +02:00
Hauke Heibel
6b2ab13ac5 fix vs.net compilation issue 2009-08-06 11:40:25 +02:00
Hauke Heibel
8163757cf0 fix vs.net compilation issue 2009-08-06 11:27:25 +02:00
Gael Guennebaud
fa55cf5ce7 fix compilation and segfault issues 2009-08-06 11:19:36 +02:00
Gael Guennebaud
b9b17ea5a5 add the missing Affine Transform * set of column vectors products... 2009-08-06 11:02:03 +02:00
Benoit Jacob
0744638b6f remove remnant of MultiplierBase 2009-08-06 10:35:13 +02:00
Gael Guennebaud
84a7659bef implement the missing outer product, 
and attempt to workaround a gcc internal error
 2009-08-05 17:39:11 +02:00
Gael Guennebaud
88147e0a91 big refactoring in Product.h: 
- all specialized products now inherits ProductBase
- the default product evaluated by Assign is still here,
  but it is currently enabled for small fixed sizes only
- => this significantly speed up compilation for large matrices
- I left the OuterProduct specialization empty as an exercise...
 2009-08-05 15:23:35 +02:00
Benoit Jacob
014c581a5b fix assertions, improve docs. 
we never assert on conditions that depend on the result of a computation!!
also the assertion that rank>0 amounts to matrix!=0 which we have to leave under the responsibility of the user.
 2009-08-05 10:15:28 +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
f3a6bc48c4 fix a couple of compilation issue due to the removal of MultiplierBase 2009-08-04 13:16:40 +02:00
Gael Guennebaud
2089a263f8 merge 2009-08-04 11:31:25 +02:00
Gael Guennebaud
4bec101470 implement two levels of blocking in PartialLU => high speedup 2009-08-04 11:28:02 +02:00
Benoit Jacob
4436a4d68c use explicit Block/VectorBlock xprs to make sure that compile-time known sizes are used 2009-08-04 00:27:58 +02:00
Benoit Jacob
523cdedf58 make the dot product linear in the second variable, not the first variable 2009-08-03 17:20:45 +02:00
Gael Guennebaud
912da9fade merge with special_matrix branch 2009-08-03 16:17:32 +02:00
Gael Guennebaud
a8f943127c merge 2009-08-03 16:11:30 +02:00
Benoit Jacob
d10c710b15 add new Householder module 2009-08-03 16:06:57 +02:00
Gael Guennebaud
3cf5bb31f6 * Bye bye MultiplierBase, extend a bit AnyMatrixBase to allow =, +=, and -= 
* This probably makes ReturnByValue needless
 2009-08-03 16:05:15 +02:00
Benoit Jacob
66ee2044ce small fixes 2009-08-03 16:05:07 +02:00
Benoit Jacob
3cde9c0e35 apply Gael's idea for auto transpose in mixed fixed/dynamic case 2009-08-03 16:04:15 +02:00
Gael Guennebaud
ce1dc1ab16 implements a blocked version of PartialLU 2009-08-03 12:11:18 +02:00
Gael Guennebaud
0103de8512 bugfix in trsm 2009-08-02 15:32:43 +02:00
Gael Guennebaud
48fc64458c add blocked LLT, and bugfix in trsm asserts 2009-08-01 23:42:51 +02:00
Gael Guennebaud
18429156a1 add selfadjointView from a trinagularView 2009-07-31 17:35:55 +02:00
Gael Guennebaud
2796bcabb1 some cleaning 2009-07-31 17:35:20 +02:00
Gael Guennebaud
a156f5a869 faster trsm kernel and fix a couple of issues 2009-07-31 13:18:19 +02:00
Gael Guennebaud
21f686846b s/std::atan2/ei_atan2 2009-07-31 10:08:23 +02:00
Manuel Yguel
ae5e26a363 add missing ei_atan2 without painfull warnings 2009-07-31 09:21:31 +02:00
Gael Guennebaud
2e9f7f80bf compilation fixes for sun CC 2009-07-31 10:04:34 +02:00
Benoit Jacob
d8bfd151d1 forward-port Anthony Truchet's changeset 8eab0bccbf 2009-07-30 16:05:38 +02:00
Gael Guennebaud
ff20a2ba94 add explicit "on the right" triangular solving, 
=> no temporary when the rhs/unknows is row major
 2009-07-30 16:03:06 +02:00
Gael Guennebaud
62d9b9b7b5 fix typo 2009-07-29 09:26:20 +02:00
Gael Guennebaud
864171df5c fix a couple of issues related to recent products 2009-07-28 18:11:30 +02:00
Gael Guennebaud
1ba35248e9 synch with main branch 2009-07-28 17:37:22 +02:00
Gael Guennebaud
54804eb626 synch with main branch 2009-07-28 17:35:07 +02:00
Gael Guennebaud
264fe82c65 add a debug mechanism to compute the number of intermediate evaluations (only for dynamic size) 2009-07-28 17:13:13 +02:00
Gael Guennebaud
7ed7ec64b5 improve the expression analyzer to bypass Transpose expression 2009-07-28 14:02:12 +02:00
Gael Guennebaud
5f3606bce9 bug fix in inverse for 1x1 matrix, 
some compilation fixes in sparse_solvers
 2009-07-27 18:09:56 +02:00
Gael Guennebaud
94cc30180e compilation fixes 2009-07-27 13:50:23 +02:00
Gael Guennebaud
0590c18555 various compilation and bug fixes in selfadjoint stuff 2009-07-27 13:17:39 +02:00
Gael Guennebaud
b5e4064289 cleaning and fix a perf issue 2009-07-27 12:13:53 +02:00
Gael Guennebaud
f95b77be62 trmm is now fully working and available via TriangularView::operator* 2009-07-27 11:42:54 +02:00
Gael Guennebaud
6aba84719d trmm is now working in all storage order configurations 2009-07-27 10:27:01 +02:00
Gael Guennebaud
1d4d9a37fd some cleaning 2009-07-26 13:53:24 +02:00
Gael Guennebaud
f3fde74695 finalize trsm: works in all situations, and it is now used by solve() and solveInPlace() 2009-07-26 13:01:37 +02:00
Gael Guennebaud
282e18da49 ok, now trsm works very well for upper triangular matrices 
TODO: link it with the meta triangular_solve_selector and handle
the case where the rhs is row major by copying it to a col-major
temporary + handle right solving: X = B * M^-1
 2009-07-26 00:49:17 +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
35927e78c2 add WIP trsm 2009-07-24 16:21:52 +02:00
Gael Guennebaud
c6d06c22ac some cleaning 2009-07-24 10:53:31 +02:00
Gael Guennebaud
6076173f0b add a simplified version of the sybb kernel built on top of gebp 2009-07-24 10:08:21 +02:00
Gael Guennebaud
82c5438c95 split and add unit tests for symm and syrk, 
the .rank*update() functions now returns a reference to *this
 2009-07-23 21:22:51 +02:00
Gael Guennebaud
b67abe22b3 oops,, update SYRK so that the rhs can be non-square² 2009-07-23 20:56:04 +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
713c92140c improve SYMV it is now faster and ready for use 2009-07-23 14:20:45 +02:00
Gael Guennebaud
eee14846e3 formating 2009-07-23 10:19:58 +02:00
Gael Guennebaud
ddb3ac98a2 addd matrix * self adjoint high level API 2009-07-23 10:05:38 +02:00
Hauke Heibel
8d2cd744b0 Added an explicit cast from int to bool to suppress MSVC warnings. 2009-07-23 00:11:25 +02:00
Gael Guennebaud
f696efc00e bugfix in SYMM 2009-07-22 23:48:42 +02:00
Gael Guennebaud
0cb4f32e12 implement high level API for SYMM and fix a couple of bugs related to complex 2009-07-22 23:12:22 +02:00
Gael Guennebaud
e7f8e939e2 * GEMM enhencement: no need to pre-transpose the rhs 
=> faster a * b.transpose() product
  => this also fix a bug in a so far untested situation
* SYMM is now ready for use => still have to write the high level
  stuff to convert natural expressions into a call to SYMM
 2009-07-22 18:04:16 +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
afa8f2ca95 * various fixes related to sub diagonals and band matrix 
* allows 0 sized objects in Block/Map
 2009-07-21 11:19:52 +02:00
Gael Guennebaud
a012aecbc4 bugfix in SVD 2009-07-20 13:44:52 +02:00
Gael Guennebaud
4375c043ac minor compilation fixes for Sun CC and ICC 2009-07-20 13:27:41 +02:00
Gael Guennebaud
4c85fa8c73 compilation fix (sun CC) 2009-07-20 10:57:31 +02:00
Gael Guennebaud
c10b919edb compilation fix 2009-07-20 10:56:03 +02:00
Gael Guennebaud
b3ad796d40 bugfix in operator*= (matrix product) 2009-07-20 10:44:07 +02:00
Gael Guennebaud
a551107cce bugfix for a = a * b; when a has to be resized 2009-07-20 10:35:47 +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
525da6a464 bugfix in blueNorm 2009-07-16 14:20:36 +02:00
Gael Guennebaud
34490f1493 * bugfixes in Product, and test/product_selfadjoint 
* speed up in the extraction of the matrix Q in Tridiagonalization
 2009-07-16 00:03:17 +02:00
Gael Guennebaud
97c9445c60 synch with main devel branch 2009-07-15 19:54:31 +02:00
Gael Guennebaud
079fa81d84 add a TridiagonalMatrix wrapper arround BandMatrix, and extend this latter 2009-07-15 19:53:08 +02:00
Gael Guennebaud
4f792583c7 add BandMatrix::col() 2009-07-15 18:00:11 +02:00
Gael Guennebaud
df6561a73f change the implementation of BandMatrix to follow the BLAS/LAPACK storage scheme 2009-07-15 17:00:49 +02:00
Gael Guennebaud
1578421ed1 fix issue #25 : the problem was that we assumed Dynamic was a multiple of a packet size 
(also disable the test of blueNorm)
 2009-07-15 14:20:45 +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
7a9519a9be fix typo in blue norm 2009-07-14 23:00:53 +02:00
Gael Guennebaud
279cedc1ce some cleaning/renaming is Triangular/SelfadjointView 2009-07-14 22:38:21 +02:00
Gael Guennebaud
f5d2317b12 add a blueNorm() function implementing the Blues's stable norm 
algorithm. it is currently provided for experimentation
purpose only.
 2009-07-13 21:14:47 +02:00
Gael Guennebaud
ddbaaebf9e one more fix of the previous commit (forgot to update ei_must_nest_by_value) 2009-07-13 15:27:01 +02:00
Gael Guennebaud
bd506d837c fix typo in previous commit 2009-07-13 15:21:32 +02:00
Gael Guennebaud
1e7b1a8a85 add a SparseNestByValue expression and fix issue in sparse adjoint evaluation 2009-07-13 14:55:03 +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
ab17f92728 more sun studio fixes 2009-07-10 16:27:01 +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
b47dea8b7a add a meta unroller for the triangular solver (only for vectors as rhs) 2009-07-10 11:30:46 +02:00
Gael Guennebaud
1a1b2e9f27 finally directly calling the low-level products is faster 2009-07-10 10:41:26 +02:00
Gael Guennebaud
1c52985aa7 merge 2009-07-10 08:21:57 +02:00
Gael Guennebaud
629e083d81 slight change in the comparison to -1 2009-07-10 08:21:20 +02:00
Gael Guennebaud
8885d56928 commit woking versions of triangular solvers naturally 
handling conjuagted expression. still have to bench whether it
is faster (runtime and compile time) to directly call the
cache friendly functions, whence all the commented piece of code...
 2009-07-09 23:59:18 +02:00
Gael Guennebaud
fa60c72398 started to simplify the triangular solvers 2009-07-09 17:11:03 +02:00
Gael Guennebaud
96e7d9f896 ok now all the complex mat-mat and mat-vec products involving conjugate, 
adjoint, -, and scalar multiple seems to be well handled. It only remains
the simpler case: C = alpha*(A*B) ... for the next commit
 2009-07-08 18:24:37 +02:00