Update Eigen to commit:2620cb930b7ad87ed1d77a26319739f4b1c86d33
CHANGELOG
=========
2620cb930 - Update file Geometry_SIMD.h
b2c9ba2be - Fix preprocessor condition on when to use fast float logistic implementation.
283d69294 - Guard AVX2 implementation of psignbit in PacketMath.h
be54cc8de - Fix preverse for PowerPC.
c5b234196 - Fix unused variable warning in TensorIO.h
86aee3d9c - Fix long double random
776d86d8d - AVX: guard Packet4l definition
e63d9f6cc - Fix random again
f75e2297d - AVX2 - double->int64_t casting
13092b5d0 - Fix usages of Eigen::array to be compatible with std::array.
77833f932 - Allow symbols to be used in compile-time expressions.
d26e19714 - Add missing cwiseSquare, tests for cwise matrix ops.
35bf6c8ed - Add SimplicialNonHermitianLLT and SimplicialNonHermitianLDLT
4dccaa587 - Use truncation rather than rounding when casting Packet2d to Packet2l.
7b5d32b7c - Sparse move
c8d368bda - More fixes for 32-bit.
de304ab96 - Fix using ScalarPrinter redefinition for gcc.
c54303848 - Undef macro in TensorContractionGpu.h that causes buildbreakages.
d8aa4d6ba - Fix another instance of Packet2l on win32.
9f77ce4f1 - Add custom formatting of complex numbers for Numpy/Native.
5570a2786 - cross3_product vectorization
0b3df4a6e - Remove "extern C" in CholmodSupport.
a39ade4cc - Protect use of alloca.
b86641a4c - Add support for casting between double and int64_t for SSE and AVX2.
d88393258 - Fix Packet*l for 32-bit builds.
d792f13a6 - Make more Matrix functions constexpr
d3cd31265 - Remove slow index check in Tensor::resize from release mode.
386e2079e - Fix Jacobi module doc.
8b101ade2 - Fix CwiseUnaryView for MSVC.
0951ad2a8 - Don'\''t hide rbegin/rend for GPU.
24f8fdeb4 - Fix CwiseUnaryView const access (Attempt 2).
285da30ec - Fix const input and c++20 compatibility in unary view.
126ba1a16 - Add Packet2l for SSE.
1d4369c2f - Fix CwiseUnaryView.
352ede96e - Fix incomplete cholesky.
f1adb0ccc - Split up cxx11_tensor_gpu to reduce timeouts.
17f3bf898 - Fix pexp test for ARM.
6da34d9d9 - Allow aligned assignment in TRMV.
3e8e63eb4 - Fix packetmath plog test on Windows.
5ffb307af - Fix deprecated anonymous enum-enum conversion warnings
55dd48747 - Revert "fix unaligned access in trmv"
38fcedaf8 - Fix pexp complex test edge-cases.
251ec4208 - Return 0 volume for empty AlignedBox
64edfbed0 - Fix static_assert for c++14.
3f3144f53 - fix unaligned access in trmv
23f6c2685 - Rip out make_coherent, add CoherentPadOp.
edaf9e16b - Fix triangular matrix-vector multiply uninitialized warning.
98620b58c - Eliminate FindCUDA cmake warning.
cc941d69a - Update error about c++14 requirement.
6893287c9 - Add degenerate checks before calling BLAS routines.
fa201f1bb - Fix QR colpivoting warnings and test failure.
b33491070 - delete shadowed typedefs
a962a2759 - Fix MSVC GPU build.
a2f8eba02 - Speed up sparse x dense dot product.
7a88cdd6a - Fix signed integer UB in random.
a6dc930d1 - Speed up SparseQR.
feaafda30 - Change array_size result from enum to constexpr.
8a73c6490 - Remove "using namespace Eigen" from blas/common.h.
6ed4d80cc - Fix crash in IncompleteCholesky when the input has zeros on the diagonal.
3859e8d5b - Add method signDeterminant() to QR and related decompositions.
db6b9db33 - Make header guards in GeneralMatrixMatrix.h and Parallelizer.h consistent:...
b56e30841 - Enable direct access for IndexedView.
90087b990 - Fix use of uninitialized memory in kronecker_product test.
6b365e74d - Fix GPU build for ptanh_float.
b14c5d0fa - Fix real schur and polynomial solver.
8a4118746 - fix exp complex test: use int instead of index
960892ca1 - JacobiSVD: get rid of m_scaledMatrix, m_adjoint, hopefully fix some compiler warnings
18a161bf1 - fix pexp_complex_test
be06c9ad5 - Implement float pexp_complex
4d419e220 - Rename generic_fast_tanh_float to ptanh_float and move it to...
PiperOrigin-RevId: 622268063
Change-Id: Ia444528ed7190f328160d8784de910a3722b5aed
139 files changed
