Commit Graph

  • 71b64d3498 Updated tests for enhanced AlignedBox with volume, diagonal and better handling of integral types. Manuel Yguel 2010-01-20 14:55:10 +01:00
  • 8918d18e21 Improved patch from Manuel Yguel: Enhance AlignedBox to accept integral types and add some usefull methods: diagonal, volume, sample. Gael Guennebaud 2010-02-10 11:40:55 +01:00
  • bb290977b8 add highest and lowest functions to NumTraits Gael Guennebaud 2010-02-10 11:11:21 +01:00
  • fe0827495a * move dummy_precision and epsilon to NumTraits * make NumTraits inherits std::numeric_limits Gael Guennebaud 2010-02-10 10:52:28 +01:00
  • c11df02f0d Deactivated test which requires variadic macros. Hauke Heibel 2010-02-09 20:52:15 +01:00
  • 5da3049b80 Regression tests for number of nested temporaries. Moved EIGEN_DEBUG_MATRIX_CTOR to ei_matrix_storage to capture resize related allocations. Hauke Heibel 2010-02-09 20:32:48 +01:00
  • f96e6a6944 merge Gael Guennebaud 2010-02-09 16:41:48 +01:00
  • 71e580c4aa fix nesting in Arraywrapper and nesting_ops Gael Guennebaud 2010-02-09 16:38:36 +01:00
  • 04333c6ace Now variadic macro related warnings should be supressed as well under Linux. Hauke Heibel 2010-02-09 16:37:24 +01:00
  • 905050b239 extend sparse product benchmark with ublas Gael Guennebaud 2010-02-09 15:55:36 +01:00
  • 285bc336d5 document lazyProduct Gael Guennebaud 2010-02-09 14:45:17 +01:00
  • 840977529f * as promised, remove the "optimization" for Product::diagonal() * add MatrixBase::lazyProduct Gael Guennebaud 2010-02-09 14:34:31 +01:00
  • 9ce1212d7c For the record, here is a solution for (a*b).diagonal, at the cost of extra copies if a and/or b as to be evaluated. So in the next commit I'll remove it. A nice solution would be to evaluate the lhs/rhs into member of the initial product, but that would be overkill. Gael Guennebaud 2010-02-09 14:28:22 +01:00
  • 0a680a9857 Added a non-diagonal product nesting test. Hauke Heibel 2010-02-09 14:24:17 +01:00
  • dd283b3f82 Fixed paste&copy error. Hauke Heibel 2010-02-09 14:12:41 +01:00
  • 928ae382b4 Added debug only unit test for nesting ops - just run ./check nesting. Hauke Heibel 2010-02-09 14:07:37 +01:00
  • 8185a3c6cf fix one useless temp & copy Gael Guennebaud 2010-02-09 13:16:29 +01:00
  • 1cb59e4781 fix nesting lazy prod by ref Gael Guennebaud 2010-02-09 11:27:30 +01:00
  • d104d2cd29 add accessors to coeff based product Gael Guennebaud 2010-02-09 11:13:22 +01:00
  • 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) Gael Guennebaud 2010-02-09 11:05:39 +01:00
  • 0398e21198 s/UnrolledProduct/CoeffBasedProduct Gael Guennebaud 2010-02-09 10:02:26 +01:00
  • c076fec734 fix the multiple temporary issue for nested products Gael Guennebaud 2010-02-09 09:58:34 +01:00
  • 8b016e717f get rid of NestParentByRefBit Gael Guennebaud 2010-02-08 16:51:41 +01:00
  • 871698d3aa Introduced NestParentByRefBit and NestByRefBit - this should fix temporaries related to nested products. Fixed a few typos and a few warnings. Hauke Heibel 2010-02-06 17:43:32 +01:00
  • 6f3f857897 make noalias works for coefficient based products Gael Guennebaud 2010-02-05 23:44:24 +01:00
  • 52167be4c8 make sure the correct diagoanl() function is called in trace() Gael Guennebaud 2010-02-04 18:51:29 +01:00
  • 73eb0e633c * resurected Flagged from Eigen2Support * reimplement .diagonal() for ProductBase to make (A*B).diagonal() more efficient! Gael Guennebaud 2010-02-04 18:28:09 +01:00
  • b44240180f optiization: make hybrid small/large outer products use the unrolled path Gael Guennebaud 2010-02-04 17:17:57 +01:00
  • d0b4ef81f0 Prevent temporaries for reductions. Hauke Heibel 2010-02-04 14:26:03 +01:00
  • 1a77334d54 Silenced type conversion warnings. Hauke Heibel 2010-02-03 19:20:25 +01:00
  • 05837be8fb Fixed a warning. Transform::Identity() is now returning a Transform. Hauke Heibel 2010-02-03 09:46:27 +01:00
  • 8861dce7ee Fixed 32bit builds. Hauke Heibel 2010-02-03 09:07:17 +01:00
  • 7b2dd988fa Fixes #89. Added regression test. Hauke Heibel 2010-02-02 09:27:41 +01:00
  • 7c41fb66f8 fix compilation on 32bits systems Gael Guennebaud 2010-02-01 11:45:08 +01:00
  • 08f154b93a remove some trailing nestbyvalue Gael Guennebaud 2010-02-01 11:44:44 +01:00
  • dd817361f5 use unrolled product path for small outer product Gael Guennebaud 2010-01-31 12:28:16 +01:00
  • 15c0a78c6e One warning less... Hauke Heibel 2010-01-30 17:24:18 +01:00
  • 43f0c0cbb3 fix triangular view assignment Gael Guennebaud 2010-01-30 09:02:18 +01:00
  • b6521b799f add specialization of ei_ref_selector for Array (fix a big perf issue \!) Gael Guennebaud 2010-01-29 21:28:23 +01:00
  • 6dee5440e4 Adapted mean to work with complex numbers. Added regression test. Hauke Heibel 2010-01-29 12:12:02 +01:00
  • 42b88983ff Fixed mean reduction leading to unresolved symbol. Hauke Heibel 2010-01-29 11:48:16 +01:00
  • ae06365bbd Disable variadic macro warning when compiling at full warning level. I was not able to get a macro version running and thus I opted for a cmake patch. Hauke Heibel 2010-01-29 09:53:19 +01:00
  • 2b2fcc9460 erm.... using nxn is the actual purpose of this variant, fix this. Thomas Capricelli 2010-01-29 08:59:25 +01:00
  • 375b5faa8a Fix copy-paste error in first_aligned test. Jitse Niesen 2010-01-28 17:06:20 +00:00
  • 13b078efc9 remove reference to dead option Benoit Jacob 2010-01-28 08:46:01 -05:00
  • 33abe75afa Fixed Quaternion operator*= added regression test. Hauke Heibel 2010-01-28 10:32:44 +01:00
  • 98a584ceb8 Put the Status outside of the class, it really does not depend on the FunctorType or Scalar template parameters. Thomas Capricelli 2010-01-28 10:29:26 +01:00
  • 27cf1b3a97 eigenization of ei_r1updt() Thomas Capricelli 2010-01-28 04:28:52 +01:00
  • 40eac2d8a0 misc cleaning / eigenization Thomas Capricelli 2010-01-28 04:19:39 +01:00
  • fcd074c928 silent warning of icc Thomas Capricelli 2010-01-27 23:43:32 +01:00
  • bd732986bc fix compilation Thomas Capricelli 2010-01-27 23:43:15 +01:00
  • 0ce5bc0d14 add support for global math function for array Gael Guennebaud 2010-01-27 23:23:59 +01:00
  • 7d92b7ad5f Modified license header. Hauke Heibel 2010-01-27 20:47:37 +01:00
  • 3bfba8c9a9 Added the missing unit test file. Hauke Heibel 2010-01-27 20:34:42 +01:00
  • 5b9cc65418 Added EIGEN_DEFINE_STL_VECTOR_SPECIALIZATION macro including unit tests and documentation. Hauke Heibel 2010-01-27 20:34:05 +01:00
  • 828d058b4b EIGEN_ENUM_MIN ---> EIGEN_SIZE_MIN Benoit Jacob 2010-01-27 11:42:04 -05:00
  • dcbf104bcc add EIGEN_DEFAULT_TO_ROW_MAJOR cmake option for the tests. Benoit Jacob 2010-01-27 07:48:48 -05:00
  • 35bacf7cb8 *forward port fix in MapBase::coeff(int) and coeffRef(int) *forward port expanded map.cpp unit test *fix unused variable warnings Benoit Jacob 2010-01-27 07:11:49 -05:00
  • 7ba9dc07ed port ei_rwupdt to c++, and misc cleaning Thomas Capricelli 2010-01-27 09:01:13 +01:00
  • e97529c2e3 doc : update code, mention examples Thomas Capricelli 2010-01-27 08:14:50 +01:00
  • 4365a48748 Added an ei_linspaced_op to create linearly spaced vectors. Added setLinSpaced/LinSpaced functionality to DenseBase. Improved vectorized assignment - overcomes MSVC optimization issues. CwiseNullaryOp is now requiring functors to offer 1D and 2D operators. Adapted existing functors to the new CwiseNullaryOp requirements. Added ei_plset to create packages as [a, a+1, ..., a+size]. Added more nullaray unit tests. Hauke Heibel 2010-01-26 19:42:17 +01:00
  • afb9bf6281 use PlanarRotation<> instead of handmade givens rotation in cminpack code + cleaning. This results in some more memory being used, but not much. Thomas Capricelli 2010-01-26 17:40:55 +01:00
  • c04a93df31 clean r1mpyq: remove fortran leftovers Thomas Capricelli 2010-01-26 14:08:52 +01:00
  • 55f328b264 misc cleaning Thomas Capricelli 2010-01-26 13:20:24 +01:00
  • 71f513e250 forgot to commit this: qform.h is not used anymore Thomas Capricelli 2010-01-26 13:08:29 +01:00
  • 69f11c08a1 more eigenization, dropped 'ipvt' in lm Thomas Capricelli 2010-01-26 12:09:52 +01:00
  • 8a690299c6 fix possible segfault Thomas Capricelli 2010-01-26 11:48:32 +01:00
  • d791b51112 remove spaces Thomas Capricelli 2010-01-26 10:50:43 +01:00
  • 113995355b get rid of ei_qform + lot of other cleaning, now that we do not depend on minpack qr factorization. Thomas Capricelli 2010-01-26 08:42:48 +01:00
  • ba2a9cce03 some more eigenization Thomas Capricelli 2010-01-26 07:36:15 +01:00
  • a3034ee079 cleaning Thomas Capricelli 2010-01-26 06:05:01 +01:00
  • 91561cded4 use a plain matrix to store the upper triangular matrix 'R', instead of the "compact inside a vector" scheme used by fortran/minpack. The most difficult part is to fix all related code. Tests pass. Thomas Capricelli 2010-01-26 05:59:58 +01:00
  • 4b859c8554 cleaning Thomas Capricelli 2010-01-26 01:59:32 +01:00
  • c759814f11 some more (thoroughly checked) eigenization Thomas Capricelli 2010-01-26 01:43:58 +01:00
  • bdb0e9fcd0 Clean up one compilation error and two warnings. Jitse Niesen 2010-01-26 16:02:19 +00:00
  • 1403cea087 fix a bug introduced between the cminpack version of Manolis Lourakis and the one from Frédéric Devernay. Here, we want to compute the max over the column, the -1 is not needed in fortran because indices start at 1, contrary to c/c++. Thomas Capricelli 2010-01-26 04:55:36 +01:00
  • 9651e0c503 Use eigen methods for solving triangular systems. We loose again very slightly on both speed and precision on some tests. Thomas Capricelli 2010-01-25 11:34:52 +01:00
  • 92be7f461b define ei_lmpar2() that takes a ColPivHouseholderQR as argument. We still need to keep the old one around, though. Thomas Capricelli 2010-01-25 07:23:38 +01:00
  • ee0e39284c Replace the qr factorization from (c)minpack (qrfac) by Eigen's own stuff. Results as checked by unit tests are very slightly worse, but not much. Thomas Capricelli 2010-01-25 07:22:28 +01:00
  • 1cb0be05b0 useful cleaning Thomas Capricelli 2010-01-25 07:08:08 +01:00
  • cbf6022e5a useless cleaning Thomas Capricelli 2010-01-25 07:07:31 +01:00
  • 6ae7d842a3 generate a compilation error when using ReturnByValue::coeff() or coeffRef(), instead of doing an infinite recursion Benoit Jacob 2010-01-24 21:44:18 -05:00
  • 858539a6af Use matrices with clustered eigenvalues in matrix function test. This is in order to get better code coverage. Test matrix_function_3 now fails regularly because ComplexSchur reaches the max number of iterations; further study needed. Jitse Niesen 2010-01-24 22:56:28 +00:00
  • d08035f3e1 fix the script again (definitely?) + cleaning Thomas Capricelli 2010-01-22 19:26:29 +01:00
  • be11a254ac rm ExpressionMaker stuff (weird as I was pretty sure that I had already removed them) Gael Guennebaud 2010-01-22 10:17:43 +01:00
  • d40c110053 lot of cleaning: - clean the *_PUBLIC_INTERFACE_* - update Diagonal, ReturnByValue, ForceAlignedAccess, UnaryView, etc. to support array - many other small stuff Gael Guennebaud 2010-01-22 10:15:41 +01:00
  • f1a025185a Added array() to ArrayBase and matrix() to MatrixBase(). Hauke Heibel 2010-01-21 17:55:09 +01:00
  • dbf3af866e Remove some Array #includes. Jitse Niesen 2010-01-21 12:31:03 +00:00
  • 7bf5930496 Adapted Geometry includes. Adapted the decomposition documentation regarding the solve signature. Hauke Heibel 2010-01-21 09:43:30 +01:00
  • ecc71abdda Added the Array include's warning for GCC. Hauke Heibel 2010-01-20 21:36:50 +01:00
  • 26df0609e2 Corrected the Array include's deprecation warning for MSVC. Hauke Heibel 2010-01-20 20:56:52 +01:00
  • 85d80d0fcd merge Hauke Heibel 2010-01-20 20:51:42 +01:00
  • 5d48cc1f5b Moved the Array module to Core. Hauke Heibel 2010-01-20 20:51:01 +01:00
  • e78e3cd41b Fix bug in MatrixBase::setIdentity(int, int). Jitse Niesen 2010-01-20 12:07:46 +00:00
  • e0c2c6385f Add small test for Matrix::setIdentity() This is to exhibit the bug that makes the jacobisvd_7 test fail. Jitse Niesen 2010-01-20 10:51:59 +00:00
  • 89ee9f092f Fixed compilation of MatrixFunctions module. Hauke Heibel 2010-01-20 11:32:13 +01:00
  • d5d5417062 add SSE code (from Intel) for the fast inversion of 4x4 matrices of double Gael Guennebaud 2010-01-19 16:04:04 +01:00
  • 60b0ddc3e1 update the fast 4x4 SSE inversion code from more recent Intel's code Gael Guennebaud 2010-01-19 15:33:45 +01:00
  • a13ffbd836 Get rid off GCC warning on comparing enums from different types. Jitse Niesen 2010-01-19 11:05:52 +00:00
  • dacfa97e82 merge Mark Borgerding 2010-01-18 19:53:44 -05:00