Difference between revisions of "ChangeLog"
(Eigen 3.0.3 changelog)
|Line 1:||Line 1:|
== Eigen 3.0.3 ==
== Eigen 3.0.3 ==
Revision as of 13:39, 6 December 2011
- 1 Eigen 2.0.17
- 2 Eigen 3.0.3
- 3 Eigen 3.0.2
- 4 Eigen 3.0.1
- 5 Eigen 2.0.16
- 6 Eigen 3.0.0
- 7 Eigen 3.0-rc1
- 8 Eigen 3.0-beta4
- 9 Eigen 3.0-beta3
- 10 Eigen 3.0-beta2
- 11 Eigen 2.0.15
- 12 Eigen 3.0-beta1
- 13 Eigen 2.0.14
- 14 Eigen 2.0.13
- 15 Eigen 2.0.12
- 16 Eigen 2.0.11
- 17 Eigen 2.0.10
- 18 Eigen 2.0.9
- 19 Eigen 2.0.8
- 20 Eigen 2.0.7
- 21 Eigen 2.0.6
- 22 Eigen 2.0.5
- 23 Eigen 2.0.4
- 24 Eigen 2.0.3
- 25 Eigen 2.0.2
- 26 Eigen 2.0.1
- 27 Eigen 2.0.0
Released December 6, 2011
Changes since 2.0.16:
- Fix a compilation bug in aligned_allocator: the allocate method should take a void pointer
- Fix a typo in ParametrizedLine documentation
Released October 6, 2011
Changes since 3.0.2:
- Fix compilation errors when Eigen2 support is enabled.
- Fix bug in evaluating expressions of the form matrix1 * matrix2 * scalar1 * scalar2.
- Fix solve using LDLT for singular matrices if solution exists.
- Fix infinite loop when computing SVD of some matrices with very small numbers.
- Allow user to specify pkgconfig destination.
- Several improvements to the documentation.
Released August 26, 2011
Changes since 3.0.1:
- Windows.h: protect min/max calls from macros having the same name (no need to #undef min/max anymore).
- MinGW: fix compilation issues and pretty gdb printer.
- Standard compliance: fix aligned_allocator and remove uses of long long.
- MPReal: updates for the new version.
- Other fixes:
- fix aligned_stack_memory_handler for null pointers.
- fix std::vector support with gcc 4.6.
- fix linking issue with OpenGL support.
- fix SelfAdjointEigenSolver for 1x1 matrices.
- fix a couple of warnings with new compilers.
- fix a few documentation issues.
Released May 30, 2011
Changes since 3.0.0:
- Fix many bugs regarding ARM and NEON (Now all tests succeed on ARM/NEON).
- Fix compilation on gcc 4.6
- Improved support for custom scalar types:
- Fix memory leak issue for scalar types throwing exceptions.
- Fix implicit scalar type conversion.
- Math functions can be defined in the scalar type's namespace.
- Fix bug in trapezoidal matrix time matrix product.
- Fix asin.
- Fix compilation with MSVC 2005 (SSE was wrongly enabled).
- Fix bug in EigenSolver: normalize the eigen vectors.
- Fix Qt support in Transform.
- Improved documentation.
Released May 28, 2011
Changes since 2.0.15:
- Fix bug in 3x3 tridiagonlisation (and consequently in 3x3 selfadjoint eigen decomposition).
- Fix compilation for new gcc 4.6.
- Fix performance regression since 2.0.12: in some matrix-vector product, complex matrix expressions were not pre-evaluated.
- Fix documentation of Least-Square.
- New feature: support for part<SelfAdjoint>.
- Fix bug in SparseLU::setOrderingMethod.
Released March 19, 2011, at the meeting.
See the Eigen 3.0 release notes.
Only change since 3.0-rc1:
- Fixed compilation of the unsupported 'openglsupport' test.
Released March 14, 2011.
Main changes since 3.0-beta4:
- Core: added new EIGEN_RUNTIME_NO_MALLOC option and new set_is_malloc_allowed() option to finely control where dynamic memory allocation is allowed. Useful for unit-testing of functions that must not cause dynamic memory allocations.
- Core: SSE performance fixes (follow-up from bug 203).
- Core: Fixed crashes when using EIGEN_DONT_ALIGN or EIGEN_DONT_ALIGN_STATICALLY (bug 213 and friends).
- Core: EIGEN_DONT_ALIGN and EIGEN_DONT_ALIGN_STATICALLY are now covered by unit tests.
- Geometry: Fixed transform * matrix products (bug 207).
- Geometry: compilation fix for mixing CompactAffine with Homogeneous objects
- Geometry: compilation fix for 1D transform
- SVD: fix non-computing constructors (correctly forward computationOptions) (bug 206)
- Sparse: fix resizing when the destination sparse matrix is row major (bug 37)
- more Eigen2Support improvements
- more unit test fixes/improvements
- more documentation improvements
- more compiler warnings fixes
- fixed GDB pretty-printer for dynamic-size matrices (bug 210)
Released February 28, 2011.
Main changes since 3.0-beta3:
- Non-vectorization bug fixes:
- fix bug 89: work around an extremely evil compiler bug on old GCC (<= 4.3) with the standard assert() macro
- fix Umfpack back-end in the complex case
- Vectorization bug fixes:
- fix a segfault in "slice vectorization" when the destination might not be aligned on a scalar (complex<double>)
- fix bug 195: fast SSE unaligned loads fail on GCC/i386 and on Clang
- fix bug 186: worked around a GCC 4.3 i386 backend issue with SSE
- fix bug 203: SSE: a workaround used in pset1() resulted in poor assembly
- worked around a GCC 4.2.4 internal compiler error with vectorization of complex numbers
- lots of AltiVec compilation fixes
- NEON compilation fixes
- API additions and error messages improvements
- Transform: prevent bad user code from compiling
- fix bug 190: directly pass Transform Options to Matrix, allowing to use RowMajor. Fix issues in Transform with non-default Options.
- factorize implementation of standard real unary math functions, and add acos, asin
- Build/tests system
- Lots of unit test improvements
- fix installation of unsupported modules
- fixed many compiler warnings, especially on the Intel compiler and on LLVM/Clang
- CTest/CMake improvements
- added option to build in 32bit mode
- BLAS/LAPACK implementation improvements
- The Blas library and tests are now automatically built as part of the tests.
- expanded LAPACK interface (including syev)
- now Sparse solver backends use our own BLAS/LAPACK implementation
- fix bug 189 (cblat1 test failure)
- improved conservativeResize methods documentation
- documented sorting of eigenvalues
- misc documentation improvements
- improve documentation of plugins
Released February 12, 2011.
The biggest news is that the API is now 100% stable.
Main changes since 3.0-beta2:
- The "too many to list them all" category:
- lots of bug fixes
- lots of performance fixes
- lots of compiler support fixes
- lots of warning fixes
- lots of unit tests improvements and fixes
- lots of documentation improvements
- lots of build system fixes
- API changes:
- replaced ei_ prefix by internal:: namespace. For example, ei_cos(x) becomes internal::cos(x).
- renamed PlanarRotation -> JacobiRotation
- renamed DenseStorageBase -> PlainObjectBase
- HouseholderSequence API cleanup
- refactored internal metaprogramming helpers to follow closely the standard library
- made UpperBidiagonalization internal
- made BandMatrix/TridiagonalMatrix internal
- Core: also see below, "const correctness".
- Sparse: EIGEN_YES_I_KNOW_SPARSE_MODULE_IS_NOT_STABLE_YET must be defined to use Eigen/Sparse
- Core: random<interger type>() now spans over range of width RAND_MAX
- New API:
- Core: added Map static methods taking strides
- SVD: added jacobiSvd() method
- Sparse: many misc improvements and new features. Improved support for Cholmod, Amd, SuperLU and other back-ends.
- Core: allow mixed real-complex dot products
- Geometry: allow mixed real-complex cross products
- Geometry: allow to pass Options parameters to Transform, Quaternion and other templates, to control memory alignment
- QR: add threshold API to FullPivHouseholderQR
- Core: added tan function
- Const correctness:
- Eigen now properly enforces const-correctness everywhere, for example with Map objects. This will break compilation of code that relied on former behavior.
- A new kind of test suite was added to check that, 'failtest'.
- Complete BLAS library built on top of Eigen. Imported BLAS test suite, which allowed to fix many issues.
- Partial LAPACK implementation. Passing part of the LAPACK test suite, which also allowed to fix some issues.
- Eigen 2 Support:
- tons of improvements in EIGEN2_SUPPORT
- new incremental migration path: see http://eigen.tuxfamily.org/dox-devel/Eigen2SupportModes.html
- imported a copy of the Eigen 2 test suite, made sure that Eigen 3 passes it. That also allowed to fix several issues.
Released October 15, 2010.
Main changes since 3.0-beta1:
- Add support for the vectorization of std::complex<> with SSE, AltiVec and NEON.
- Add support for mixed real * complex matrix products with vectorization.
- Finalize the JacobiSVD class with: compile time options, thin/full decompositions, and least-square solving.
- Several improvement of the Transform class. In particular, there is no default mode anymore.
- New methods: middleRows(), middleCols(), TriangularMatrix::conjugate()
- New unsupported modules: OpenGL, MPFR C++
- Many improvements in the support of empty objects.
- Many improvements of the vectorization logic.
- Add the possibility to extend QuaternionBase.
- Vectorize Quaternion multiplication with double.
- Significant improvements of the documentation.
- Improved compile time errors.
- Enforce static allocation of temporary buffers in gemm (when possible).
- Fix aligned_delete for null pointers and non trivial dtors.
- Fix eigen decomposition of 3x3 float matrices.
- Fix 4x4 matrix inversions (vectorization).
- Many fixes in QR: solving with m>n, use of rank, etc.
- Fixes for MSVC for windows mobile and CLang.
- Remove the Taucs backend (obsolete).
- Remove the old SVD class (was causing too much troubles, a new decompozition based on bidiagonalisation/householder should come back soon, JacobiSVD can be used meanwhile).
Released July 16, 2010
Changes since 2.0.14:
- Fix bug: certain cases of matrix-vector product (depending on storage order) were blocked by an assertion failure.
- Fix LU and QR solve when rank==0, fix LLT when the matrix is purely 0.
- Fix a couple of bugs with QR solving especially with rows>cols.
- Fix bug with custom scalar types that have non-trivial destructor.
- Fix for ICC in SSE code.
- Fix some C++ issues found by Clang (patch by Nick Lewycky).
Released July 5, 2010
See the announcement.
Released June 22, 2010
Changes since 2.0.13:
- Fix #141: crash in SSE (alignment problem) when using dynamic-size matrices with a max-size fixed at compile time that is not a multiple of 16 bytes. For example, Matrix<double,Dynamic,Dynamic,AutoAlign,5,5>.
- Fix #142: LU of fixed-size matrices was causing dynamic memory allocation (patch by Stuart Glaser).
- Fix #127: remove useless static keywords (also fixes warnings with clang++).
Released June 10, 2010
Changes since 2.0.12:
- Fix #132: crash in certain matrix-vector products. Unit test added.
- Fix #125: colwise norm() and squaredNorm() on complex types do not return real types
- Fully support the QCC/QNX compiler (thanks to Piotr Trojanek). The support in 2.0.12 was incomplete. The whole test suite is now successful.
- As part of the QCC support work, a lot of standards compliance work: put std:: in front of a lot of things such as size_t, check whether the math library needs to be linked to explicitly...
- Fix precision issues in LDLT. The isPositiveDefinite() method is now always returning true, but it was conceptually broken anyway, since a non-pivoting LDLT decomposition can't know that.
- Compilation fix in ldlt() on expressions.
- Actually install the Eigen/Eigen and Eigen/Dense public headers!
- Fix readcost for complex types.
- Fix compilation of the BTL benchmarks.
- Some dox updates.
Released February 12, 2010
Changes since 2.0.11:
- EIGEN_DEFAULT_TO_ROW_MAJOR is fully supported and tested.
- Several important fixes for row-major matrices.
- Fix support of several algorithms for mixed fixed-dynamic size matrices where the fixed dimension is greater than the dynamic dimension. For example: Matrix<float,3,Dynamic>(3,2)
- fix EIGEN_DONT_ALIGN: now it _really_ disables vectorization (was giving a #error unless you also used EIGEN_DONT_VECTORIZE).
- Fix #92: Support QNX's QCC compiler (patch by Piotr Trojanek)
- Fix #90, missing type cast in LU, allow to use LU with MPFR (patch by 'Wolf').
- Fix ICC compiler support: work around a bug present at least in ICC 11.1.
- Compilation fixes for computeInverse() on expressions.
- Fix a gap in a unit-test (thanks to Jitse Niesen)
- Backport improvements to benchmarking code.
- Documentation fixes
Released January 10, 2010
Changes since 2.0.10:
- Complete rewrite of the 4x4 matrix inversion: we now use the usual cofactors approach, so no numerical stability problems anymore (bug #70)
- Still 4x4 matrix inverse: SSE path for the float case, borrowing code by Intel, giving very high performance.
- Fix crash happening on 32-bit x86 Linux with SSE, when double's were created at non-8-byte-aligned locations (bug #79).
- Fix bug in Part making it crash in certain products (bug #80).
- Precision improvements in Quaternion SLERP (bug #71).
- Fix sparse triangular solver for lower/row-major matrices (bug #74).
- Fix MSVC 2010 compatibility.
- Some documentation improvements.
Released November 25, 2009
Changes since 2.0.9:
- Rewrite 4x4 matrix inverse to improve precision, and add a new unit test to guarantee that precision. It's less fast, but it's still faster than the cofactors method.
- Fix bug #62: crash in SSE code with MSVC 2008 (Thanks to Hauke Heibel).
- Fix bug #65: MatrixBase::nonZeros() was recursing infinitely
- Fix PowerPC platform detection on Mac OSX.
- Prevent the construction of bogus MatrixBase objects and generate good compilation errors for that. Done by making the default constructor protected, and adding some private constructors.
- Add option to initialize all matrices by zero: just #define EIGEN_INITIALIZE_MATRICES_BY_ZERO
- Improve Map documentation
- Install the pkg-config file to share/pkgconfig, instead of lib/pkgconfig (thanks to Thomas Capricelli)
- fix warnings
- fix compilation with MSVC 2010
- adjust to repository name change
Released October 24, 2009
Changes since 2.0.8:
- Really fix installation and the pkg-config file.
- Install the NewStdVector header that was introduced in 2.0.6.
Released October 23, 2009
Changes since 2.0.7:
- fix installation error introduced in 2.0.7: it was choking on the pkg-config file eigen2.pc not being found. The fix had been proposed long ago by Ingmar Vanhassel for the development branch, and when recently the pkg-config support was back-ported to the 2.0 branch, nobody thought of backporting this fix too, and apparently nobody tested "make install" !
- SVD: add default constructor. Users were relying on the compiler to generate one, and apparenty 2.0.7 triggered a little MSVC 2008 subtlety in this respect. Also added an assert.
Released October 22, 2009
Changes since 2.0.6:
- fix bug #61: crash when using Qt QVector on Windows 32-bit. By Hauke Heibel.
- fix bug #10: the reallocateSparse function was half coded
- fix bug in SparseMatrix::resize() not correctly initializing by zero
- fix another bug in SparseMatrix::resize() when outerSize==0. By Hauke Heibel.
- fully support GCC 3.3. It was working in 2.0.2, there was a compilation error in 2.0.6, now for the first time in 2.0.7 it's 100% tested (the test suite passes without any errors, warnings, or failed tests).
- SVD: add missing assert (help catch mistakes)
- fixed warnings in unit-tests (Hauke Heibel)
- finish syncing Memory.h with the devel branch. This is cleaner and fixes a warning. The choice of system aligned malloc function may be affected by this change.
- add pkg-config support by Rhys Ulerich.
- documentation fix and doc-generation-script updates by Thomas Capricelli
Released September 23, 2009
Changes since 2.0.5:
- fix bug: visitors didn't work on row-vectors.
- fix bug #50: compilation errors with swap().
- fix bug #42: Add Transform::Identity() as mentioned in the tutorial.
- allow to disable all alignment code by defining EIGEN_DONT_ALIGN (backport from devel branch).
- backport the devel branch's StdVector header as NewStdVector. You may also #define EIGEN_USE_NEW_STDVECTOR to make StdVector use it automatically. However, by default it isn't used by StdVector, to preserve compatibility.
- Vectorized quaternion product (for float) by Rohit Garg (backport from devel branch).
- allow to override EIGEN_RESTRICT and add EIGEN_DONT_USE_RESTRICT_KEYWORD
- fix a warning in ei_aligned_malloc; fixed by backporting the body from the devel branch; may result in a different choice of system aligned malloc function.
- update the documentation.
Released August 22, 2009
Changes since 2.0.4:
- fix bug: in rare situations involving mixed storage orders, a matrix product could be evaluated as its own transpose
- fix bug: LU::solve() crashed when called on the LU decomposition of a zero matrix
- fix bug: EIGEN_STACK_ALLOCATION_LIMIT was too high, resulting in stack overflow for a user. Now it is set as in the devel branch.
- fix compilation bug: our StdVector header didn't work with GCC 4.1. (Bug #41)
- fix compilation bug: missing return statement in Rotation2D::operator*= (Bug #36)
- in StdVector, a more useful #error message about the #including order
- add EIGEN_TRANSFORM_PLUGIN allowing to customize the Transform class
- fix a warning with MSVC
- fix a bug in our cmake code when building unit-tests (thanks to Marcus Hanwell)
- work around a bug in cmake that made it fail to build unit-tests when fortran wasn't installed
- in our cmake code, remove the part about retrieving the mercurial info and appending it to the version number in the dox
- dox: remove the old example list
- fix the option to build a binary library, although it's not very useful and will be removed
- add basic .hgignore file and script to build the docs (thanks to Thomas Capricelli)
Released August 1, 2009
Changes since 2.0.3:
- Several fixes in the overloaded new and delete operators. Thanks to Hauke Heibel.
- compilation fix: add the missing ei_atan2 function. Thanks to Manuel Yguel.
- Use ei_atan2 instead of using std::atan2 directly.
- several compilation fixes in the Qt interoperability code: methods toQTransform() and toQMatrix(). Thanks to Anthony Truchet.
- compilation fix and simplification in Matrix assignment
- compilation fixes in "a *= b" and "a=a*b" when a has to be resized.
- remove a "stupid" version of ei_pow. for integers for gcc >= 4.3
- bug fix in Quaternion::setFromTwoVectors()
- several ctest improvements: use our own dashboard, use a separate project for the 2.0 branch.
- documentation: improvement on the pages on unaligned arrays (the online copies have been updated immediately).
Released June 21, 2009
Changes since 2.0.2:
- precision and reliability fixes in various algorithms, especially LLT, QR, Tridiagonalization, and also a precision improvement in LU.
- fix LLT and LDLT solve() on uninitialized result (was causing an assertion).
- add Eigen/Eigen and Eigen/Dense headers for convenience
- document a newly found cause for the "unaligned array" assertion
- backport documentation improvements on transpose() and adjoint()
- updates in the Sparse module (was needed to support KDE 4.3)
Released May 22, 2009
Changes since 2.0.1:
- Fix linearRegression() compilation, actually it is reimplemented using the better fitHyperplane() which does total least-squares.
- Add missing setZero() etc... variants taking size parameters and resizing. These were mentioned in the tutorial but weren't implemented.
- Fix posix_memalign platform check. This fixes portability issues. Thanks to Ross Smith.
- Fix detection of SSE2 on the Windows 64-bit platform.
- Fix compatibility with the old GCC 3.3: it is now fully supported again.
- Fix warnings with recent GCC (4.4.0 and 4.3.3).
Released April 14, 2009
Changes since 2.0.0:
- disable alignment altogether on exotic platforms on which we don't vectorize anyway. This allows e.g. to use Eigen on ARM platforms.
- new StdVector header with a new workaround for the problems with std::vector.
- workarounds for MSVC internal compiler errors
- MSVC 9 compilation fix (patch by Hauke Heibel)
- fixes for various bugs in Maps/Blocks that could give wrong results
- fix bug in 4x4 matrix inverse that could give wrong results
- compilation fix in SliceVectorization
- fix wrong static assertion (patch by Markus Moll)
- add missing operators in aligned_allocator (thanks to Hauke Heibel)
Released February 2, 2009