Update Eigen to commit:dceb779ecd822f55b4ae78f760371b0e08a889f2
CHANGELOG
=========
dceb779ec - Fix test for pow with mixed integer types. We do not convert the exponent if it is an integer type.
afc014f1b - Allow mixed types for pow(), as long as the exponent is exactly representable in the base type.
b2c82a934 - Remove bad skew_symmetric_matrix3 test.
e8a2aa24a - Fix a couple of issues with unary pow():
07d075995 - [ROCm] Fix for sparse matrix related breakage on ROCm.
fb212c745 - Fix g++-6 constexpr and c++20 constexpr build errors.
ec9c7163a - Feature/skew symmetric matrix3
311ba66f7 - Fix realloc for non-trivial types.
3c37dd2a1 - Tweak bound for pow to account for floating-point types.
f9dfda28a - Add missing comparison operators for GPU packets.
242325eca - Remove unused variable.
133498c32 - Add constexpr, test for C++14 constexpr.
69f50e3a6 - Adjust overflow threshold bound for pow tests.
3e44f960e - Reduce compiler warnings for tests.
b7e21d4e3 - Call check_that_malloc_is_allowed() in aligned_realloc()
6e83e906c - fix typo in doc/TutorialSparse.dox
525f06667 - fixed msvc compilation error in GeneralizedEigenSolver.h
f241a2c18 - Add asserts for index-out-of-bounds in IndexedView.
f5364331e - Fix some cmake issues.
d816044b6 - Fix mixingtypes tests.
94cc83faa - 2 typos fix in the 3rd table.
30c42222a - Fix some test build errors in new unary pow.
bd393e15c - Vectorize acos, asin, and atan for float.
e5af9f87f - Vectorize pow for integer base / exponent types
8acbf5c11 - re-enable pow for complex types
7064ed134 - Specialize psign<Packet8i> for AVX2, don't vectorize psign<bool>.
98e51c9e2 - Avoid undefined behavior in array_cwise test due to signed integer overflow
a7c1cac18 - Fix GeneralizedEigenSolver::info() and Asserts
714678fc6 - Add missing ptr in realloc call.
b2a13c9dd - Sparse Core: Replace malloc/free with conditional_aligned
6aad0f821 - Fix psign for unsigned integer types, such as bool.
1a09defce - Protect new pblend implementation with EIGEN_VECTORIZE_AVX2
7c67dc67a - Use proper double word division algorithm for pow<double>. Gives 11-15% speedup.
7a3b667c4 - Add support for AVX512-FP16 for vectorizing half precision math
76a669fb4 - add fixed power unary operation
39fcc8979 - Removed unnecessary checks for FP16C
2f7cce2dd - [SYCL] Fix some SYCL tests
27367017b - Disable bad "deprecated warning" edge-case in BDCSVD
b8e93bf58 - Eliminate bool bitwise warnings.
66ea0c09f - Don't double-define Half functions on aarch64
97e0784dc - Vectorize the sign operator in Eigen.
be20207d1 - Fix vectorized Jacobi Rotation
7a87ed1b6 - Fix code and unit test for a few corner cases in vectorized pow()
9e0afe0f0 - Fix non-VSX PowerPC build
84a9d6fac - Fix use of Packet2d type for non-VSX.
ce60a7be8 - Partial Packet support for GEMM real-only (PowerPC). Also fix compilation warnings & errors for some conditions in new API.
5a1c7807e - Fix inner iterator for sparse block.
39d22ef46 - Fix flaky packetmath_1 test.
7896c7dc6 - Use numext::sqrt in ConjugateGradient.
e618c4a5e - Improve pblend AVX implementation
ef4654bae - Add true determinant to QR and it's variants
b7668c037 - Avoid including <sstream> with EIGEN_NO_IO
7dd3dda3d - Updated AccelerateSupport documentation after PR 966.
69714ff61 - Add Sparse Subset of Matrix Inverse
PiperOrigin-RevId: 474559050
Change-Id: Ife8edf526089b3bc65e391c975c8c852489e7225
131 files changed
