Commit Graph

2326 Commits

Author SHA1 Message Date
Jitse Niesen
cca7b146a2 Implement evaluator for Map 2011-04-05 18:15:59 +01:00
Gael Guennebaud
a6b5314c20 Performance tunning for TRMM products 2011-04-05 11:20:50 +02:00
Jitse Niesen
ae06b8af5c Make evaluators for Matrix and Array inherit from common base class.
This gets rid of some code duplication.
2011-04-04 15:35:14 +01:00
Jitse Niesen
afdd26f229 Do some of the actual work in evaluator for Block.
Also, add simple accessor methods to Block expression class.
2011-04-04 13:44:50 +01:00
Jitse Niesen
d90a8ee8bd Evaluators: add Block evaluator as dumb wrapper, add slice vectorization. 2011-03-31 13:50:52 +01:00
Gael Guennebaud
b471161f28 fix typo and remove unused declaration. 2011-03-31 10:02:02 +02:00
Adam Szalkowski
969e92261d fix bug #239: the essential part was left uninitialized in some cases 2011-03-31 09:54:52 +02:00
Jitse Niesen
8175fe43e0 Evaluators: Make inner vectorization more similar to default traversal. 2011-03-28 21:29:47 +01:00
Gael Guennebaud
6feb1d3c0b fix trmv for Strictly* triangular matrices and trapezoidal matrices 2011-03-28 17:42:26 +02:00
Gael Guennebaud
568478ffe5 fix trmm for some unusual trapezoidal cases (a dense set of columns or rows is zero) 2011-03-28 17:41:46 +02:00
Jitse Niesen
b175bc464f Evaluators: Implement linear traversal, better testing. 2011-03-27 22:08:48 +01:00
Jitse Niesen
1b17a674dd Evaluators: Implement inner vectorization.
The implementation is minimal (I only wrote the functions called by
the unit test) and ugly (lots of copy and pasting).
2011-03-27 13:49:15 +01:00
Jitse Niesen
5c204d1ff7 Evaluators: Implement LinearVectorizedTraversal, packet ops in evaluators. 2011-03-25 16:30:41 +00:00
Gael Guennebaud
e6fa4a267a improve computation of the sub panel width 2011-03-24 23:42:25 +01:00
Gael Guennebaud
931814d7c0 improve performance of trsm 2011-03-24 23:19:53 +01:00
Jitse Niesen
c6ad2deead Bug fix in linspace_op::packetOp(row,col). Fixes bug #232.
Also, add regression test.
2011-03-24 10:42:11 +00:00
Gael Guennebaud
42bc1f77be impl basic product evaluator on top of previous one 2011-03-24 09:33:36 +01:00
Gael Guennebaud
abc8c0821c makes evaluator test use VERIFY_IS_APPROX 2011-03-23 17:23:56 +01:00
Gael Guennebaud
816541d82c add a stupid Product<A,B> expression produced by prod(a,b), and implement a first version of its evaluator 2011-03-23 16:12:21 +01:00
Gael Guennebaud
cfd5c2d74e import evaluator works 2011-03-23 11:54:00 +01:00
Gael Guennebaud
5fda8cdfb3 fix 228 (ei_aligned_stack_delete does not exist anymore) 2011-03-21 21:59:42 +01:00
Gael Guennebaud
b8ecda5c66 clean a bit the stack allocation mechanism 2011-03-19 10:27:47 +01:00
Gael Guennebaud
290205dfc0 fix memory leak when a custom scalar throw an exception 2011-03-19 01:06:50 +01:00
Benoit Jacob
5991d247f9 bump 2011-03-18 05:27:58 -04:00
Benoit Jacob
4931a719f4 bump 2011-03-14 14:10:05 -04:00
Jitse Niesen
27f34269d5 Document EIGEN_DEFAULT_DENSE_INDEX_TYPE.
Also, expand description of EIGEN_DONT_ALIGN.
2011-03-11 11:15:44 +00:00
Benoit Jacob
dc36efbb8f fix bug #219: Map Flags AlignedBit was miscomputed, didn't account for EIGEN_ALIGN 2011-03-10 10:17:17 -05:00
Oliver Ruepp
5d1263e7c5 bug #37: fix resizing when the destination sparse matrix is row major 2011-03-08 16:37:59 +01:00
Benoit Jacob
bfcad536e8 * bug #206: correctly forward computationOptions and work towards avoiding mallocs after preallocation, with unit test.
* added EIGEN_RUNTIME_NO_MALLOC and new set_is_malloc_allowed() function to implement that test
2011-03-06 20:59:25 -05:00
Benoit Jacob
b464fc19bc try to fix a ICC 11.1 compiler error (bug #217) 2011-03-06 19:27:31 -05:00
Benoit Jacob
c541d0a62e disable ICC 12 warning 279 - controlling expression is constant 2011-03-06 19:06:44 -05:00
Benoit Jacob
56818d907e Make EIGEN_ALIGN16 always align to fix crashes with EIGEN_DONT_ALIGN_STATICALLY. New macro EIGEN_USER_ALIGN16 had the old behavior i.e. honors user preference. 2011-03-04 09:57:49 -05:00
Gael Guennebaud
4f0909b5f0 fix bug #212 (installation of Eigen2Support/Geometry) 2011-03-04 14:16:58 +01:00
Gael Guennebaud
5cec29162b fix compilation in the case of 1D Transform 2011-03-02 19:29:55 +01:00
Gael Guennebaud
703c8a0cc6 fix compilation when mixing CompactAffine with Homogeneous objects 2011-03-02 19:27:13 +01:00
Gael Guennebaud
d30f0c0953 fix transform * matrix products: in particular it now truely considers the rhs as a set of (homogeneous) points and do not neglect the homogeneous coordinates in the case of affine transform 2011-03-02 19:26:38 +01:00
Gael Guennebaud
c8e1b679fa re-enable fast pset1-pstore by introducing a new higher level pstore1 function 2011-03-02 10:55:44 +01:00
Benoit Jacob
77fc6a9914 bump 2011-02-28 00:55:52 -05:00
Benoit Jacob
eef03525b8 fix bug #203: revert to using _mm_set1_p[sd] 2011-02-28 00:04:05 -05:00
Benoit Jacob
9be2712bf7 remove now-useless comments 2011-02-27 22:35:17 -05:00
Benoit Jacob
0612768c1c fix bug #201: Clang too has intrinsics bugs preventing us to use custom unaligned loads 2011-02-27 21:59:07 -05:00
Benoit Jacob
4846c76d9d shut up a stupid clang 2.8 warning 2011-02-27 20:18:03 -05:00
Benoit Jacob
afc9efca15 fix compilation with clang 2.8 2011-02-27 20:17:47 -05:00
Benoit Jacob
ea7d872181 documentation fixes 2011-02-27 17:43:10 -05:00
Benoit Jacob
b3544ce2ae bug #195 - fix this once and for all: just never use _mm_load_sd on gcc/i386, it generates redundant x87 ops 2011-02-27 17:26:59 -05:00
Jitse Niesen
a8f5ef9388 Document (non)sorting of eigenvalues.
Also, update docs for (Generalized)SelfAdjointEigenSolver to reflect that these
two classes were split apart.
2011-02-27 14:06:55 +00:00
Gael Guennebaud
ef73265987 to ease debugging let's catch invalid template options in Transform 2011-02-25 09:03:24 +01:00
Benoit Jacob
5dfae4524b fix bug #195: fast unaligned load for integer using _mm_load_sd failed when the value interpreted as a NaN 2011-02-24 10:31:57 -05:00
Hauke Heibel
2064c59878 Improved docs of PlainObjectBase::conservativeResize methods. 2011-02-24 15:48:41 +01:00
Gael Guennebaud
bb9a465c5a fix AltiVec ploaddup 2011-02-24 00:23:50 +03:00