Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | Big renaming: | 2010-01-04 | |
| | | | | | | start ---> head end ---> tail Much frustration with sed syntax. Need to learn perl some day. | ||
* | make some changes to please clang, fix some warnings too. | 2010-01-04 | |
| | |||
* | Fix #69 for the second time, and add the respective regression test | 2010-01-04 | |
| | |||
* | remove debug output. sorry! | 2010-01-02 | |
| | |||
* | clarify docs as requested on forum | 2010-01-02 | |
| | |||
* | * Fix bug #79: ei_alignmentOffset was assuming that ptr is multiple of | 2010-01-02 | |
| | | | | | | | sizeof(Scalar), and that assumption breaks with double on linux x86-32. * Rename ei_alignmentOffset to ei_first_aligned * Rewrite its documentation and part of its body * The variant taking a MatrixBase doesn't need a separate size argument. | ||
* | put the assign assert and debug info before the assignment itself | 2010-01-01 | |
| | |||
* | sorry for committing this mess | 2009-12-23 | |
| | |||
* | fix #76, MayLinearVectorize depends on MaxSizeAtCompileTime and not on ↵ | 2009-12-23 | |
| | | | | MaxInnerSize | ||
* | fix the xpr analyzer for Transpose | 2009-12-23 | |
| | |||
* | clean a bit Hessenberg and make sure the subdiagonal coeff is real even for ↵ | 2009-12-23 | |
| | | | | 2x2 matrices | ||
* | fix #75, and add a basic unit test for Hessenberg | 2009-12-23 | |
| | | | | (it was indirectly tested by the eigenvalue decomposition) | ||
* | fix #69 and extend unit tests or triangular solvers | 2009-12-23 | |
| | |||
* | support gcc 4.5 | 2009-12-21 | |
| | |||
* | add missing inclusion of LU/arch, thanks to J.B. Rouault | 2009-12-19 | |
| | |||
* | oops, remove duplicated ctor | 2009-12-19 | |
| | |||
* | much better workaround for empty struct (the previous one caused GCC 4.3 to ↵ | 2009-12-18 | |
| | | | | generate wrong code leading to segfaults) | ||
* | * introduce ei_alignmentOffset(MatrixBase&,Integer) | 2009-12-16 | |
| | | | | | couldnt put it in Memory.h as it needs the definition of MatrixBase * make Redux use it | ||
* | this really fix the previous warning | 2009-12-16 | |
| | |||
* | fix warning by making ei_empty_struct::_ei_dummy_ private | 2009-12-16 | |
| | |||
* | fix compilation when mixing types | 2009-12-16 | |
| | |||
* | * fix aliasing checks when the lhs is also transposed. At the same time, | 2009-12-16 | |
| | | | | | | significantly simplify the code of these checks while extending them to catch much more expressions! * move the enabling/disabling of vectorized sin/cos to the architecture traits | ||
* | suppress unused variable warnings | 2009-12-15 | |
| | |||
* | no, this wasn't equivalent to ei_pload at all, after all! | 2009-12-15 | |
| | |||
* | Gael, who is a man of few words^Winstructions, is right, as usual. | 2009-12-15 | |
| | |||
* | Suppressed the warning for missing assignment generators (forgot that in the ↵ | 2009-12-15 | |
| | | | | | | last submission). Commented Quake3's fast inverser sqrt in SSE's MathFunction header. | ||
* | *use scalar instructions, packet not needed here | 2009-12-14 | |
| | | | | *remove unused var warning | ||
* | only include SSE path if SSE enabled | 2009-12-14 | |
| | |||
* | add SSE path for Matrix4f inverse, taken from Intel except that we do a kosher | 2009-12-14 | |
| | | | | | | division instead of RCPPS-followed-by-Newton-Raphson. The rationale for that is that elsewhere in Eigen we dont allow ourselves this approximation (which throws 2 bits of mantissa), so there's no reason we should allow it here. | ||
* | change the Arch constants: const int ---> enum, more explicit names, and use | 2009-12-14 | |
| | | | | of a namespace instead of Prefix_Name. | ||
* | Warning 4512 (assignment operators could not be generated) is now simply ↵ | 2009-12-14 | |
| | | | | | | disabled. All unimplemented assignment operators have been removed. | ||
* | Fixed a bad type conversion. | 2009-12-13 | |
| | |||
* | Correct type of ei_solve_retval<LLT,...>::operator= | 2009-12-12 | |
| | |||
* | Removed more warnings. | 2009-12-12 | |
| | |||
* | Added to possibility to compile unit tests at maximum warning level. | 2009-12-12 | |
| | | | | Silenced (amongst others) many conversion related warnings. | ||
* | * 4x4 inverse: revert to cofactors method | 2009-12-09 | |
| | | | | | | | * inverse tests: use createRandomMatrixOfRank, use more strict precision * tests: createRandomMatrixOfRank: support 1x1 matrices * determinant: nest the xpr * Minor: add comment | ||
* | Documentation clean up. | 2009-12-08 | |
| | | | | | | | | * remove non-existant reference to CwiseAll * define \householder_module (used in HouseholderSequence.h) * update I01_TopicLazyEvaluation.dox - Product is now called GeneralProduct * remove reference to list of examples which was deleted ages ago * rename PartialLU_solve.cpp snippet to PartialPivLU_solve.cpp | ||
* | Fix poor Quaternion::slerp snapping | 2009-12-04 | |
| | |||
* | merge | 2009-12-04 | |
|\ | |||
* | | fix compilation and clean a bit Map<Quaternion> | 2009-12-04 | |
|/ | |||
* | fix MSVC10 compilation | 2009-12-02 | |
| | |||
* | * HouseholderSequence: | 2009-12-02 | |
| | | | | | | | * be aware of number of actual householder vectors (optimization in non-full-rank case, no behavior change) * fix applyThisOnTheRight, it was using k instead of actual_k * QR: rename matrixQ() to householderQ() where applicable | ||
* | * remove empty destructors in Matrix.h and MatrixStorage.h | 2009-12-02 | |
| | |||
* | merge | 2009-12-02 | |
|\ | |||
| * | Added NestByValue and .nestByValue() again for the sake of backwards ↵ | 2009-12-02 | |
| | | | | | | | | compatibility. | ||
| * | Removed unnecessary code. | 2009-12-02 | |
| | | |||
* | | merge with tip | 2009-12-01 | |
|\ \ | |||
* | | | added comments to help vim understand the header files are c++. | 2009-12-01 | |
| | | | |||
| * | | ColPivQR: now the unit tests even succeeds: | 2009-12-01 | |
| | | | | | | | | | | | | | | | | | | * with random matrices multiplied by 1e+8 (i.e. fixed wrong absolute fuzzy compare) * with 10,000 repetitions (i.e. the fuzzy compare is really clever) and when it occasionnally fails, less than once in 10,000 repeats, it is only on the exact rank computation. | ||
| * | | Big reworking of ColPivQR and its unit test, which now passes even with ↵ | 2009-12-01 | |
| | | | | | | | | | | | | thousands of repetitions and correctly tests matrices of all sizes. Several surprises along the way: for example, a major cause of trouble was the optimized "table of column squared norms" where the accumulation of imprecision was a serious issue; another surprise is that tests like "x!=0" before dividing by x really benefit from being replaced by fuzzy tests, as i hit real cases where i got wrong results in 1/epsilon. |