tree 9439dcad874e2d9fe5ed12cbb38a4c883a9bc2b8
parent 34de9d370db7d3d3e25d8f02330eaecf3875f54e
author C. Antonio Sanchez <cantonios@google.com> 1776975467 -0700
committer Copybara-Service <copybara-worker@google.com> 1776975489 -0700

Update Eigen to commit:ea13a98decd497a8c5588fb5de71b57bcf10d864

CHANGELOG
=========
ea13a98de - Fix imag_ref for real scalar types and clean up svd_fill.h
929785924 - Fix more cache size queries.
b2f95d373 - Fix more cache size queries.
9ae0e0f19 - Remove include from within Eigen namespace.
c1faa7473 - Add boundary test coverage: stableNorm, LinSpaced, complex GEMV, triangular solve
6b9275d1a - Add test coverage for transpose, reverse, bool redux, select, diagonal-of-product at boundaries
356a9ba1d - Add test coverage for matrix lpNorm, RowMajor partial reductions, selfadjoint boundaries
15cae8348 - Add test coverage for strided maps, triangular blocking, and mixed storage orders
93aa959b8 - Add vectorization boundary tests for redux and visitor
c93116b43 - Improve test coverage for inner product, fill, reductions, and IO
5e478d328 - Improve product test coverage at critical code-path boundaries
3a2ba7c43 - Optimize `predux_any<Packet4f>`
8190c82cb - Add missing SIMD math function benchmarks
8368a12f0 - Add runtime cache size detection for ARM and improve GEMM blocking
42c1dbd2c - Add aarch64 smoke test pipeline for MRs
875fb48f0 - fix various irksome compiler warnings
2a2456c87 - restore Eigen/src/Core/arch/Altivec/MatrixProduct.h to b1e74b1cc
c4eb3c4f4 - fix custom visitors
4387e3248 - Fix row-skipping bug in general_matrix_vector_product::run_small_cols
81550faea - Use Web Archive for dead link for the PDF referenced in Geometry/EulerAngles.h
42b6c43cf - Revert "Remove random retry loops in tests (batch 2: indices and integer types)"
54458cb39 - Remove random retry loops in tests (batch 3: geometry, sparse, umeyama)
a3cb1c659 - cxx11_tensor_random: use retry loop for low-precision RNG collisions
f80d7b825 - Fix three more flaky tests: igamma, tensor_random, matrix_power
8eaa7552f - Fix three flaky tests: packetmath, array_cwise, polynomialsolver
dd81698ae - Fix vectorization_logic test for wide SIMD widths
ab5878426 - Remove random retry loops in tests (batch 5: geometry, mixing types, triangular)
411422f2d - Remove random retry loop in SVD min-norm test
7c3a34476 - Remove random retry loops in tests (batch 2: indices and integer types)
be7538ed6 - Remove random retry loops in tests (batch 1: simple scalar cases)
5790d716c - Simplify and optimize pow/cbrt special case handling
3041ab44a - Fix GEBP asm register constraints for custom scalar types
20fce70e5 - Fix another complex div edge case.
5bacb5be9 - Fix null pointer dereference in Sparse-Dense products for Sparse vectors.
d8c8ee6fb - Fix crash on construction of SparseMatrix with zero-length diagonal
265496e86 - Fix heap overflow in BM_BatchContraction benchmark
eea4d31f5 - Simplify and modernize XprHelper.h
dd826edb4 - Replace typedef with using in tensor contraction files
abc3d6014 - Fix CUDA+Clang build warnings.
0269c017a - Revise Tensor module README.md: fix bugs, add missing docs
ca94be70d - fix uninitialized variable in constexpr function
b0ebf966a - Fix default rank-detection threshold in QR and LU decompositions
d36a7db7b - Fix Eigen::array constructors.
661cdb227 - Fix relative paths after move.
57b1de233 - Fix row-major GEMV dropping rows when n8 heuristic disables main loop
662d5c21f - Optimize SYMV, SYR, SYR2, and TRMV product kernels
c66fc5286 - Add ULP accuracy measurement tool and documentation for vectorized math functions
c20b6f5c4 - Restore EIGEN_EMPTY_STRUCT_CTOR as a no-op macro for backward compatibility
77d917359 - Add ca-certificates to clang-tidy CI job
444ae9761 - Clamp igamma/igammac output to [0,1] for numerical stability
eddb470a0 - Fix flaky array_cwise and sparse_basic tests
25dba492e - Use stack-constructed variable for SVD block sweep.
f64d1e0ac - Improve ConditionEstimator docs and tighten test bounds
8525491eb - Add dedicated unit tests and benchmark for ConditionEstimator
e730b1fe3 - Fix mixed products GEMM.
3adfa9bd3 - Add const to non-mutating member functions across remaining modules
13b61529f - Add const to non-mutating member functions in products/ and Serializer
aaca9e585 - Add missing const qualifiers in Eigen/src/Core/
1b1b7e347 - Fix EIGEN_NO_AUTOMATIC_RESIZING not resizing empty destinations
064d686c5 - Remove CXX11/ directory nesting for Tensor modules
11eb66e1b - Remove pre-C++14 workarounds from unsupported/ tensor code
a95440de1 - Remove obsolete bench/ and btl/ directories
6e2aff6b5 - Fix ambiguous static_cast in JacobiSVD blocking threshold
d8ed4f688 - Fix GEBP half/quarter-packet loops for nr>=8 RHS packing on ARM64
6b6d0d8c8 - Revert "Fix ambiguous static_cast in JacobiSVD blocking threshold computation"
ba2fc4e77 - Revert "Fix GEBP half/quarter-packet loops for nr>=8 RHS packing on ARM64"
888d708dc - Fix GEBP half/quarter-packet loops for nr>=8 RHS packing on ARM64
e567151ce - Fix ambiguous static_cast in JacobiSVD blocking threshold computation
a31de4778 - Blocked Jacobi SVD sweep with L2-cache-adaptive threshold
647e0009b - Refactor BDCSVD D&C code to reduce compilation time and memory footprint
4fab38d79 - Make clang generic vector backend support 16, 32, and 64-byte vectors
ea25ea52b - Revert accidental changes from !2212 squash merge
38f0f4275 - Update rmlarsen email address from @google.com to @gmail.com
d0d70a952 - Consolidate complex math function boilerplate with shared macros
c4c704e5d - Install libclang-rt-19-dev for asan-ubsan CI job
61895c597 - Selectively add constexpr to Core expression template scaffolding
34092d278 - Fix flaky tests: add iteration guards, yield in busy-waits, cap thread count
28d090a49 - Refactor GenericPacketMathFunctions.h into smaller focused headers
16da0279f - Add benchmarks for unsupported modules and extend supported benchmarks
fa567f6bc - Add CUDA CI jobs with NVHPC (nvc++) as host and device compiler
2cd9bb738 - Fix sparse product with entities that do not have direct access.
00cc497d3 - Add clang-tidy, codespell, and sanitizer checks to CI pipeline
241af1c0b - Add NVHPC (nvc++) compiler support and CI build/test jobs
f3f2c676b - Fix direct access for sparse blocks.
d537b51ed - Fix ComplexEigenSolver NaN with flush-to-zero arithmetic
667cabe3a - Clean up comments in unsupported module
78b76986b - Comment cleanup v3: trailing ??, informal language, FIXME/TODO colons
112c2324b - Consolidate BF16/F16 wrapper macros and simplify arch math functions
d5e67adbe - Clean up informal language, vague TODOs, and dead code in comments
7d727d26b - Refactor GenericPacketMathFunctions.h into smaller focused headers
9810969c0 - Suppress false-positive GCC and clang warnings in test builds
ad7f1fe70 - Improve clang vector extension backend
1f49bf96c - Add new benchmarks for Core, LU, and QR operations
8c35441f1 - Fix typos: misspellings, French variable names, and hyphenation
44c613216 - Fix ~40 typos found by codespell across the codebase
f52ad04bb - Fix ASAN-detected bugs in Diagonal::data() and array_cwise test
d4077a6e9 - Reorganize benchmarks into subdirectories and clean up Eigen sources
832b94097 - Update COPYING.README to clarify third-party license status
e6accc73f - Fix comment typos, doubled words, grammar errors, and copy-paste mistakes
0e424f405 - Remove dead code, commented-out blocks, and outdated comments
18791a81b - Fix MSVC build: disable [[msvc::forceinline]] on generic lambdas
95e8bc326 - Add EIGEN_LAMBDA_ALWAYS_INLINE macro for MSVC lambda inlining
a87ecfb17 - Use m_ prefix consistently for private/protected member variables
270ea539f - Remove redundant EIGEN_STRONG_INLINE from trivial constexpr and = default functions
e0a8d6c9d - Fix compile warnings
1dcea43c4 - Fix RowMajor performance for triangular/dense assignment
374fe225b - Reduce GEMV and TRSM benchmark sizes for faster routine runs
2c898e8b9 - Remove unused LhsPacketType typedef in gebp_peeled_loop
4fdc82d69 - Fix mixed-type compilation error in row-major GEMV small-cols path
4141d1fd2 - Fix -Wtautological-overlap-compare warning in row-major GEMV dispatch
53e3408cb - Optimize GEMV kernels: row-major small-cols and template deduplication
9c63d26de - Remove reference to nonexistent spmv.cpp in benchmarks
5f09b3b63 - Fix missing template argument list in trsmKernelR for Clang 20/21
c9eab4087 - Fix unused variable warning for phys_l1 on non-AVX512 builds
3c86a013b - Vectorize generic trsmKernelR for non-AVX512 targets
43a01f06a - update AVX and AVX512 to support gcc < 10.1 and clang < 10
552ca8f15 - Simplify GEBP micro-kernel and improve blocking heuristics
e953f1e50 - Modernize C++14 usage and minor optimizations in Core
f69745b67 - Fix real x complex GEMM for backends where half == full packet size
073190be0 - Fix outdated documentation across multiple .dox files
bdec88009 - Remove const from return-by-value types (issue #1087)
3108f6360 - Migrate Eigen benchmarks to the Google benchmark framework
740cac97b - Fix AVX double-precision trig and complex exp without AVX2
50d6d92a7 - Optimize sparse-dense product by bypassing InnerIterator for compressed storage
b6b2f31ba - Fix compiler warnings from GCC 13 and Clang 18
113207a9d - Optimize JacobiSVD 2x2 kernel and hoist sweep threshold
e6e5b5c4c - Fix pexp_complex for `complex<double>` (issue #3022)
2b561f928 - Revert "Specialized enable_borrowed_ranges for VectorwiseOp class range iteration"
d0654a201 - Specialized enable_borrowed_ranges for VectorwiseOp class range iteration
1a2b80727 - Fix pdiv for complex packets involving infinites.
9b709e826 - Diagonalview example typo
23fcc1c6c - MatrixBase::diagonalView issue 604
004d81a85 - Fix cblat3/zblat3 test program with gfortran.
0ac2a2df9 - Prevent predux_half for DoublePacket from accidentally catching complex Packets of size >= 16
4d05fcf8d - Fix packetmath tests on M* macs.
752911927 - betainc edge case checks at start of calculation
afb438053 - Fix RunQueue race condition on weak memory architectures (ARM64)
c64829636 - Update HVX floating-point reduction to support V79 architecture.
ddfc68d39 - Fix clang vector backend type compatibility issues.
93ff38884 - Fix relative tolerance scaling by multiplying with RHS norm in BiCGSTAB
26c242ab5 - make EIGEN_BLAS macro names consistent and undef at end of file
3d6f5fe8f - Tests: skip denorms in ARM ieee tests
fdfdd4c96 - test suite: emit the function name when an ieee test fails
e246f9cb6 - Use memset if !NumTraits<Scalar>::RequireInitialization
f46a2c561 - Fix bad static access for TensorDeviceGpu.
c09151a2c - Wrote resizing documentation page
f7772e394 - Gcc warnings
918a5f1a6 - Fix warnings related to variable_if_dynamic.
9a37aca9f - Fix assignment size assertion for EIGEN_NO_AUTOMATIC_RESIZING.
251bff288 - CUDA 13 compatibility update for unit test gpu_basic
0315fb319 - Change inline hint for `general_matrix_vector_product<>::run()` to gain performance
7aea350ba - Fix more packetmath issues for RVV
5d9beb81a - Initial version of reactivating RVV features like GeneralBlockPanelKernel
d90a0534b - fix polynomialsolver test failures
711118b74 - docs does not exists
c30af8f3d - fix UB in random implementation and tests
c5aa40675 - Fix `check_that_free_is_allowed` so that it properly checks `is_free_allowed` and not `is_malloc_allowed`
5793499a5 - Fix AVX512FP16 build.
2ac496ff8 - Revert !1953 and !1954
9164d3f16 - Fix undefined behavior in packetmath.
748e0a651 - Add missing semicolon
fe973ab0c - Force early evaluation of boost expressions.
976f15ebc - fix doc generation with doxygen 1.14 & 1.15
4f14da11d - Add basic support for packetmath for BF16 RVV
21e4582d1 - Merge remote-tracking branch 'refs/remotes/origin/master'
a7209fad7 - GemmKernel: Define static constexpr member variables out-of-class for C++14 compatibility
cdc62b84c - Merge remote-tracking branch 'origin2/master'
26fe567dd - Fix FP16 for RVV so that it will compile for gcc
afbf8173d - Merge remote-tracking branch 'origin2/master'
9b00db8cb - Simplify thread-safe initialization of GpuDeviceProperties.
8cdc0fa67 - Fix naming of predux_half for RVV when LMUL > 1
f610edadc - Merge remote-tracking branch 'origin2/master'
5aefbab77 - Merge remote-tracking branch 'origin2/master'
3ff3d0378 - Merge remote-tracking branch 'origin2/master'
196eed3d6 - Merge branch 'master' of https://gitlab.com/libeigen/eigen

PiperOrigin-RevId: 904604242
Change-Id: I3b70435e8fee585af45ad5de8384531d716a57e8