* get rid of BlockReturnType: it was not needed, and code was not always using it consistently anyway
* add topRows(), leftCols(), bottomRows(), rightCols()
* add corners unit-test covering all of that
* adapt docs, expand "porting from eigen 2 to 3"
* adapt Eigen2Support
- Updated unit tests to check above constructor.
- In the compute() method of decompositions: Made temporary matrices/vectors class members to avoid heap allocations during compute() (when dynamic matrices are used, of course).
These changes can speed up decomposition computation time when a solver instance is used to solve multiple same-sized problems. An added benefit is that the compute() method can now be invoked in contexts were heap allocations are forbidden, such as in real-time control loops.
CAVEAT: Not all of the decompositions in the Eigenvalues module have a heap-allocation-free compute() method. A future patch may address this issue, but some required API changes need to be incorporated first.
NOTE: The ComplexEigenSolver class currently _does_ allocate (line 135 of Eigenvalues/ComplexEigenSolver.h), but the reason appears to be in the implementation of matrix-matrix products, and not in the decomposition itself.
The nomalloc unit test has been extended to verify that decompositions do not allocate when max sizes are specified. There are currently two workarounds to prevent the test from failing (see comments in test/nomalloc.cpp), both of which are related to matrix products that allocate on the stack.
* renaming, e.g. LU ---> FullPivLU
* split tests framework: more robust, e.g. dont generate empty tests if a number is skipped
* make all remaining tests use that splitting, as needed.
* Fix 4x4 inversion (see stable branch)
* Transform::inverse() and geo_transform test : adapt to new inverse() API, it was also trying to instantiate inverse() for 3x4 matrices.
* CMakeLists: more robust regexp to parse the version number
* misc fixes in unit tests
- R-SVD preconditioning now done with meta selectors to avoid compiling useless code
- SVD options now honored, with options to hint "at least as many rows as cols" etc...
- fix compilation in bad cases (rectangular and fixed-size)
- the check for termination is now done on the fly, no more goto (should have done that earlier!)
- support complex numbers
- big rewrite of the 2x2 kernel, much more robust
* Jacobi:
- fix weirdness in initial design, e.g. applyJacobiOnTheRight actually did the inverse transformation
- fully support complex numbers
- fix logic to decide whether to vectorize
- remove several clumsy methods
fix for complex numbers
* normalize left Jacobi rotations to avoid having to swap rows
* set precision to 2*machine_epsilon instead of machine_epsilon, we lose 1 bit of precision
but gain between 10% and 100% speed, plus reduce the risk that some day we hit a bad matrix
where it's impossible to approach machine precision
it turns out to be better to repeat the jacobi steps on a given (p,q) pair until it
is diagonal to machine precision, before going to the next (p,q) pair. it's also
an optimization as experiments show that in a majority of cases this allows to find out
that the (p,q) pair is already diagonal to machine precision.
to guarantee the precision of the output, which is very valuable.
Here, we guarantee that the diagonal matrix returned by the SVD is
actually diagonal, to machine precision.
Performance isn't bad at all at 50% of the current householder SVD
performance for a 200x200 matrix (no vectorization) and we have
lots of room for improvement.
