1043 Commits

Author SHA1 Message Date
Benoit Jacob
d209120180 * Introduce EIGEN_DEFAULT_TO_ROW_MAJOR tests option
---> Now only product_large fails with EIGEN_DEFAULT_TO_ROW_MAJOR.
* Fix EIGEN_NO_ASSERTION_CHECKING tests option
* Fix a crash in Tridiagonalization on row-major matrices + SSE
* Fix inverse test (numeric stability noise)
* Extend map test (see previous fixes in MapBase)
* Fix vectorization_logic test for row-major
* Disable sparse tests with EIGEN_DEFAULT_TO_ROW_MAJOR
2010-01-25 14:00:02 -05:00
Thomas Capricelli
55c0707b1d fix the script again (definitely?) + cleaning 2010-01-22 19:28:33 +01:00
Benoit Jacob
72044ca925 fix a super nasty bug: on row-major expressions that are NOT vectors but that
do have LinearAccess, the MapBase::coeff(int) and MapBase::coeffRef(int)
methods were broken.
2010-01-21 23:33:20 -05:00
Benoit Jacob
c2b8ca7493 if EIGEN_DONT_ALIGN then don't try to vectorize (was giving a #error later on). 2010-01-21 22:32:16 -05:00
Gael Guennebaud
018cb8975a fix plugin doc 2010-01-17 19:55:08 +01:00
Benoit Jacob
3ab280ce4e add missing semicolon in the example 2010-01-17 12:40:19 -05:00
Benoit Jacob
b40030753b Added tag 2.0.11 for changeset 5f73a8df20235c6ff5a14c1f354c17b26fae022b 2010-01-10 11:30:40 -05:00
Benoit Jacob
5f73a8df20 bump 2.0.11 2010-01-10 11:30:10 -05:00
Thomas Capricelli
8a6d5f10dc backport from tip : actually stop on compile failure 2010-01-06 17:17:40 +01:00
Benoit Jacob
ba6ed5fa5f Fix CoeffReadCost in Part: it must account for the cost of the
conditional jump. This makes Part considered an "expensive" xpr
 that must be evaluated in operations such as Product.
This fixes bug #80.
2010-01-02 13:04:04 -05:00
Benoit Jacob
e4c88c14ec clarify docs as requested on the forum 2010-01-02 12:54:55 -05:00
Benoit Jacob
74207a31fa backport the fix to bug #79, and the unit test 2010-01-02 12:45:49 -05:00
Benoit Jacob
6fd9248c09 add Intel copyright info 2009-12-15 08:43:31 -05:00
Benoit Jacob
4262117f84 backport 4x4 inverse changes:
- use cofactors
 - use Intel's SSE code in the float case
2009-12-15 08:16:48 -05:00
Gael Guennebaud
b581cb870c fix #74: sparse triangular solver for lower/row-major matrices 2009-12-14 10:20:35 +01:00
Gael Guennebaud
72fc81dd9d backport quaternion slerp precision fix 2009-12-05 18:28:17 +01:00
Gael Guennebaud
f36650b00a fix MSVC10 compilation issues 2009-12-02 19:34:37 +01:00
Benoit Jacob
8d31f58ea1 fix bug #70
Was trying to apply stupid invertibility check to top-left 2x2 corner.
2009-11-26 15:33:07 -05:00
Benoit Jacob
a161a70696 Added tag 2.0.10 for changeset 8f1ce52e76887ecc026e8d5a68fb6ebd301113b6 2009-11-25 08:54:17 -05:00
Benoit Jacob
8f1ce52e76 bump 2.0.10 2009-11-25 08:46:42 -05:00
Benoit Jacob
268df314f1 make the inverse_4x4 test pass more consistently 2009-11-25 08:43:20 -05:00
Benoit Jacob
522022ebfc wow, restore Gael's changeset 5455d6fbe8d889f100b1526d046da16b375fc451
that I had accidentally undone in my changeset c64ca6870ea3f76b5ea5182bba1e62fd5a632f89
.
2009-11-25 08:31:25 -05:00
Thomas Capricelli
d048d7e712 fix bitbucket url after recent change 2009-11-24 23:08:16 +01:00
Benoit Jacob
cd3c8a9404 typo 2009-11-23 11:23:30 -05:00
Benoit Jacob
ec8f37ac75 improve precision test 2009-11-23 11:20:48 -05:00
Benoit Jacob
fc7f39980c backport improvement in 4x4 inverse precisio, and rigorous precision test 2009-11-23 10:27:10 -05:00
Gael Guennebaud
70af59c455 an attempt to fix compilation with recent MSVC 2009-11-23 10:29:40 +01:00
Benoit Jacob
f4dd399499 fix warnings 2009-11-16 14:15:47 -05:00
Benoit Jacob
153557527e backport: init-by-zero option: resize with same size must be a NOP 2009-11-16 13:47:02 -05:00
Benoit Jacob
6aad7f80ff avoid infinite loop, optimization not important, so a plain for loop is the safe way 2009-11-12 14:09:53 -05:00
Benoit Jacob
e3f6c3115a backport the initialize-by-0 option 2009-11-12 12:53:24 -05:00
Benoit Jacob
a2c838ff8f fix PowerPC platform detection 2009-11-11 10:52:00 -05:00
Thomas Capricelli
1e2f56c35a backport of b53c2fcc996d9944aba152dfa5917bbd8b9615eb
: fix install dir for *.pc

Ingmar Vanhassel <ingmar@exherbo.org>:
Packages that don't install architecture-specific files should install
their pkg-config file to datadir, not libdir.
2009-11-11 15:35:12 +01:00
Hauke Heibel
808c4e9581 Fixed the packport of 62 - Packet4f/d/i does not exist in 2.0. 2009-11-05 10:49:49 +01:00
Hauke Heibel
65331c3884 backporting3979f6d8aad001174160774b49b747430a7686b5
: fixed bug #62
2009-11-04 17:49:34 +01:00
Benoit Jacob
e158cdd61d fix Matrix::Map/MapAligned documentation, and rephrase the tutorial on Map 2009-10-31 14:45:50 -04:00
Benoit Jacob
c64ca6870e this explicit keyword can't hurt 2009-10-31 11:49:20 -04:00
Benoit Jacob
6a90f6c5f0 * default MatrixBase ctor: make it protected, make it a static assert, only do the check when debugging eigen to avoid slowing down compilation for everybody (this check is paranoiac, it's very seldom useful)
* add private MatrixBase ctors to catch cases when the user tries to construct MatrixBase objects directly
2009-10-31 11:48:33 -04:00
Gael Guennebaud
22dd13fdb9 backporting fix of #65 2009-10-29 14:26:38 +01:00
Gael Guennebaud
5455d6fbe8 backporting fix of #65 2009-10-29 14:26:00 +01:00
Benoit Jacob
de693cf34a remove extra ; 2009-10-28 10:04:13 -04:00
Benoit Jacob
21c4e0802d fix potential warning 2009-10-28 09:45:09 -04:00
Benoit Jacob
241b9d34a7 Hey, I was insomniac too ;)
This restores much of the performance benefit of Euler's method, without compromising accuracy (tested on 1e+7 matrices). Namely, my benchmark now runs in 1.5 s instead of 2.2 s. The same in the default branch runs in 1.08 s instead of 1.9 s, so the default branch benefits even more!
2009-10-28 03:50:29 -04:00
Benoit Jacob
9e15a6da2e Fix 4x4 matrix inversion. Applying Euler's trick is more tricky than what "high performance matrix inversion" websites would have you believe. Our 4x4 matrix inversion wasn't numerically stable, because in applying the Euler trick we didn't take the 2x2 block of biggest determinant. As a result, with float, we got relative errors above 1% every 1000 random matrices, and we got completely wrong results every 10000 matrices.
Note that this decreases the performance, but we're still significantly faster than the brutal cofactors approach.
2009-10-27 07:35:25 -04:00
Benoit Jacob
3d365a75cd Added tag 2.0.9 for changeset 38bc82a6f7e854e91bcb23162bc423ce40f141f4 2009-10-24 19:38:58 -04:00
Benoit Jacob
38bc82a6f7 bump 2.0.9 2009-10-24 16:35:46 -04:00
Benoit Jacob
6173eb67ff really fix pkgconfig support and make install.
* mistake: was using the install dir instead of binary dir
* was also using INCLUDE_INSTALL_DIR before it was set, so on a first cmake run, the pkgconfig file was bad
2009-10-24 16:16:48 -04:00
Benoit Jacob
9f89431cea install NewStdVector 2009-10-23 19:58:37 -04:00
Benoit Jacob
79e392472a Added tag 2.0.8 for changeset e1c96f3fe0aa36b19a4a8024aa3e2f701c674943 2009-10-23 18:47:33 -04:00
Benoit Jacob
e1c96f3fe0 bump 2.0.8 2009-10-23 18:37:05 -04:00