My initial fix was incorrect, the libraries must be quoted when being
passed to the add test macro, but must be unquoted when passed to the
target_link_libraries function.
- rename EvalBeforeAssignBit to MayAliasBit
- make .lazy() remove the MayAliasBit only, and mark it as deprecated
- add a NoAlias pseudo expression, and MatrixBase::noalias() function
Todo:
- we have to decide whether += and -= assume no aliasing by default ?
- once we agree on the API: update the Sparse module and the unit tests respectively.
- all specialized products now inherits ProductBase
- the default product evaluated by Assign is still here,
but it is currently enabled for small fixed sizes only
- => this significantly speed up compilation for large matrices
- I left the OuterProduct specialization empty as an exercise...
* this allows to optimize xpr like C -= lazy_product, still have to catch "scalar_product_of_lazy_product"
* started to support conjugate in cache friendly products (very useful to evaluate A * B.adjoint() without
evaluating B.adjoint() into a temporary
* compilation fix
* move solveTriangular*() to TriangularView::solve*()
* move .llt() to SelfAdjointView
* add a high level wrapper to the efficient selfadjoint * vector product
* improve LLT so that we can specify which triangular part is meaningless
=> there are still many things to do (doc, cleaning, improve the matrix products, etc.)
* rename qr() to householderQr(), for same reason.
* clarify that it's non-pivoting, non-rank-revealing, so remove all the rank API, make solve() be void instead of bool, update the docs/test, etc.
* fix warning in SVD
this is still not Eigen style code but at least it works for
n>m and it is more accurate than the JAMA based version. (I needed
it now, this is why I did that)
* Fixed a bug in umeyama for fixed size matrices.
* Fixed the umeyama unit test for fixed size matrices.
* Added XprHelper::ei_plain_matrix_type_row_major.
* switched lu/qr tests to be using createRandomMatrixOfRank
* removed unused methods doSomeRankPreservingOperations
* removed NOTE about doSomeRankPreservingOperations
introduce ei_is_diagonal to check for it
DiagonalCoeffs ---> Diagonal and allow Index to by Dynamic
-> add MatrixBase::diagonal(int) with unittest and doc
deprecated). Basically there are now only 2 functions to set a
coefficient:
1) mat.coeffRef(row,col) = value;
2) mat.insert(row,col) = value;
coeffRef has no limitation, insert assumes the coeff has not already
been set, and raises an assert otherwise.
In addition I added a much lower level, but more efficient filling
mechanism for
internal use only.
Until now, the user had to edit the source code to do that.
Internally, add EIGEN_ALIGN that takes into account both EIGEN_DONT_ALIGN.and
EIGEN_ARCH_WANTS_ALIGNMENT. From now on, only EIGEN_ALIGN should be used to
test whether we want to align.
* Cholesky decs are NOT rank revealing so remove all the rank/isPositiveDefinite etc stuff.
* fix bug in LLT: s/return/continue/
* introduce machine_epsilon constants, they are actually needed for Higman's formula determining
the cutoff in Cholesky. Btw fix the page reference to his book (chat with Keir).
* solve methods always return true, since this isn't a rank revealing dec. Actually... they already did always return true!! Now it's explicit.
* updated dox and unit-test
* in LDLT, support the negative semidefinite case
* fix bad floating-point comparisons, improves greatly the accuracy of methods like
isPositiveDefinite() and rank()
* simplifications
* identify (but not resolve) bug: claim that only triangular part is used, is inaccurate
* expanded unit-tests
Pommier. They are for float only, and they return exactly the same
result as the standard versions in about 90% of the cases. Otherwise the max error
is below 1e-7. However, for very large values (>1e3) the accuracy of sin and cos
slighlty decrease. They are about 3 or 4 times faster than 4 calls to their respective
standard versions. So, is it ok to enable them by default in their respective functors ?
* add Homogeneous expression for vector and set of vectors (aka matrix)
=> the next step will be to overload operator*
* add homogeneous normalization (again for vector and set of vectors)
* add a Replicate expression (with uni-directional replication
facilities)
=> for all of them I'll add examples once we agree on the API
* fix gcc-4.4 warnings
* rename reverse.cpp array_reverse.cpp
* add an efficient selfadjoint * vector implementation (= blas symv)
perf are inbetween MKL and GOTO
=> the interface is still missing (have to be rethougth)
* extend PartialRedux::cross() to any matrix sizes with automatic
vectorization when possible
* unit tests: add "geo_" prefix to all unit tests related to the
geometry module and start splitting the big "geometry.cpp" tests to
multiple smaller ones (also include new tests)
DiagonalCoeffs. The current API is simply:
m.diagonal<1>() => 1st super diagonal
m.diagonal<-2>() => the 2nd sub diagonal
I'll add a code snippet once we agree on this API.
This should remove most portability issues to other platforms where data alignment issues (including
overloading operator new and new[]) can be tricky, and where data alignment is not needed in the first place.
The unsupported module documentation is automatically generated in:
build/doc/unsupported/
with bidirectional cross references.
I leave a class Foo in AdolcForward module to illustrate the
cross-reference behavior. I will remove it in the next commit.
* previous DiagonalMatrix expression is now DiagonalMatrixWrapper
* DiagonalMatrix class is now for storage
* add the DiagonalMatrixBase class to factorize code of the
two previous classes
* remove Scaling class (it is now a global function)
* add UniformScaling helper class
(don't use it directly, use the Scaling function)
* add the Scaling global function to simplify the creation
of scaling objects
There is still a lot to do, in particular about DiagonalProduct for which
the goal is to get rid of the "if()" in the coeff() function. At least
it is not worse than before ! Also need to uptade the tutorial and add more doc.
EIGEN_WORK_AROUND_QT_BUG_CALLING_WRONG_OPERATOR_NEW_FIXED_IN_QT_4_5
is the right way to go for allowing placement new on a class having
overloaded operator new. Qt 4.5 won't add the :: prefix. (I still do not
understand how you can distinghish a placement new from an overloaded
operator new taking an allocator as argument...)
Question 1: why are *=scalar and /=scalar working right away ?
Same weirdness in DynamicSparseMatrix where operators += and -= work wihout
having to redefine them ???
* try to be clever in matrix ctors and operator=: be lazy when we can, always allow
to copy rowvector into columnvector, check the template parameters,
try to factor the code better
* add missing copy ctor in UnalignedType
* fix bug in the traits of DiagonalProduct
* renaming: EIGEN_TUNE_FOR_CPU_CACHE_SIZE
* update the dox a little
That means a lot of features which were available for sparse matrices
via the dense (and super slow) implemention are no longer available.
All features which make sense for sparse matrices (aka can be implemented efficiently) will be
implemented soon, but don't expect to see an API as rich as for the dense path.
Other changes:
* no block(), row(), col() anymore.
* instead use .innerVector() to get a col or row vector of a matrix.
* .segment(), start(), end() will be back soon, not sure for block()
* faster cwise product
*Add Eigen/StdVector header.
Including it #includes<vector> and "Core" and generates a partial
specialization of std::vector<T> for T=Eigen::Matrix<...>
that will work even with vectorizable fixed-size Eigen types
(working around a design issue in the c++ STL)
*Add unit-test
CCMAIL: alex.stapleton@gmail.com
only used as fallback for now, needs benchmarking.
also notice that some malloc() impls do waste memory to keep track of alignment
and other stuff (check msdn's page on malloc).
* expand test_dynalloc to cover low level aligned alloc funcs. Remove the old
#ifdef EIGEN_VECTORIZE...
* rewrite the logic choosing an aligned alloc, some new stuff:
* malloc() already aligned on freebsd and windows x64 (plus apple already)
* _mm_malloc() used only if EIGEN_VECTORIZE
* posix_memalign: correct detection according to man page (not necessarily
linux specific), don't attempt to declare it if the platform didn't declare it
(there had to be a reason why it didn't declare it, right?)
ei_aligned_malloc now really behaves like a malloc
(untyped, doesn't call ctor)
ei_aligned_new is the typed variant calling ctor
EIGEN_MAKE_ALIGNED_OPERATOR_NEW now takes the class name as parameter
* the dashboard is there: http://my.cdash.org/index.php?project=Eigen
* now you can run the tests from the top build dir
and submit report like that (from the top build dir):
ctest -D Experimental
* todo:
- add some nighlty builds (I'll add a few on my computer)
- add valgrind memory checks, performances tests, compilation time tests, etc.
