devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity
13,186 Commits 15 Branches 93 Tags
gpu-dense-solvers
Commit Graph

2960 Commits

Author SHA1 Message Date
Rasmus Munk Larsen
8593c7f5a1 GPU: Add dense cuSOLVER solvers (QR, SVD, EigenSolver) 
Add QR (geqrf + ormqr + trsm), SVD (gesvd), and self-adjoint eigenvalue
decomposition (syevd) via cuSOLVER. All support host and DeviceMatrix input.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-04-09 19:11:34 -07:00
Rasmus Munk Larsen
58c44ef36d GPU: Add library dispatch module (DeviceMatrix, cuBLAS, cuSOLVER) 
Add Eigen/GPU module: A standalone GPU library dispatch layer where
DeviceMatrix<Scalar> operations map 1:1 to cuBLAS/cuSOLVER calls.
CPU and GPU solvers coexist in the same binary with compatible syntax.

Core infrastructure:
- DeviceMatrix<Scalar>: RAII dense column-major GPU memory wrapper with
  async host transfer (fromHost/toHost) and CUDA event-based cross-stream
  synchronization.
- GpuContext: Unified execution context owning a CUDA stream + cuBLAS
  handle + cuSOLVER handle. Thread-local default with explicit override
  via setThreadLocal(). Stream-borrowing constructor for integration.
- DeviceBuffer: Typed RAII device allocation with move semantics.

cuBLAS dispatch (expression syntax):
- GEMM: d_C = d_A.adjoint() * d_B (cublasXgemm)
- TRSM: d_X = d_A.triangularView<Lower>().solve(d_B) (cublasXtrsm)
- SYMM/HEMM: d_C = d_A.selfadjointView<Lower>() * d_B (cublasXsymm)
- SYRK/HERK: d_C = d_A * d_A.adjoint() (cublasXsyrk)

cuSOLVER dispatch:
- GpuLLT: Cached Cholesky factorization (cusolverDnXpotrf + Xpotrs)
- GpuLU: Cached LU factorization (cusolverDnXgetrf + Xgetrs)
- Solver chaining: auto x = d_A.llt().solve(d_B)
- Solver expressions with .device(ctx) for explicit stream control.

CI: Bump CUDA container to Ubuntu 22.04 (CMake 3.22), GCC 10->11,
Clang 12->14. Bump cmake_minimum_required to 3.17 for FindCUDAToolkit.

Tests: gpu_cublas.cpp, gpu_cusolver_llt.cpp, gpu_cusolver_lu.cpp,
gpu_device_matrix.cpp, gpu_library_example.cu
Benchmarks: bench_gpu_solvers.cpp, bench_gpu_chaining.cpp,
bench_gpu_batching.cpp
 2026-04-09 19:05:25 -07:00
Rasmus Munk Larsen
6a9405bf7a GPU: Raise CUDA/HIP minimum and remove legacy guards 
- Raise CUDA minimum from 9.0 to 11.4 (sm_70/Volta).
- Raise HIP minimum to GFX906 (Vega 20/MI50) / ROCm 5.6.
- Remove EIGEN_HAS_{CUDA,HIP,GPU}_FP16 guards — FP16 is always available
  on sm_70+ and GFX906+.
- Remove obsolete __HIP_ARCH_HAS_* preprocessor branches.
- C++14 cleanup: remove pre-C++14 workarounds in GPU code.
- Fix NVCC warnings (deprecated register keyword, unreachable code,
  tautological comparisons).
- Fix HIP test execution on gfx1151.
- Update CI configuration for new minimum versions.
 2026-04-09 15:21:39 -07:00
Rasmus Munk Larsen
b1d2ce4c85 Revert "Speed up plog_double ~1.7x with fast integer range reduction" 
This reverts merge request !2385
 2026-04-08 13:03:48 -07:00
Rasmus Munk Larsen
ab70739c9c Speed up plog_double ~1.7x with fast integer range reduction 
libeigen/eigen!2385

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-04-07 21:48:25 -07:00
Rasmus Munk Larsen
110530a4d8 Fix bugs and improve robustness of SelfAdjointEigenSolver, improve test coverage 
libeigen/eigen!2396

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-04-07 21:08:29 -07:00
Rasmus Munk Larsen
ebae0c7c10 ulp_accuracy: use dynamic work queue for thread load balancing 
libeigen/eigen!2383

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-04-02 22:40:03 -07:00
Charles Schlosser
5977635d64 fix singed integer overflow UB in integer_types and other trivial compiler warnings 
libeigen/eigen!2380
 2026-04-03 03:36:28 +00:00
Rasmus Munk Larsen
60df12437e Fix ulp_accuracy crashes in Release builds 
libeigen/eigen!2381

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-04-02 20:12:13 -07:00
Rasmus Munk Larsen
8ec68856a6 Fix basicstuff_8 casting test failure on loongarch64 
libeigen/eigen!2379

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-04-02 14:14:54 -07:00
Rasmus Munk Larsen
64885cc6a3 Fix remaining MSVC warnings in Windows CI (C4804, C4244, C4146, C4305) 
libeigen/eigen!2374

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-04-01 17:20:31 -07:00
Pavel Guzenfeld
b7f6aed1b9 Fix dangling reference in IndexedView with expression indices 
libeigen/eigen!2335

Closes #1943
 2026-03-29 09:39:13 -07:00
Rasmus Munk Larsen
624ab58e8d Add bidiagonal SVD API to BDCSVD and remove dead debug code 
libeigen/eigen!2238

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-28 20:38:31 -07:00
Charles Schlosser
ba9871e46b fix and enable realview unit tests 
libeigen/eigen!2356
 2026-03-28 20:13:54 -07:00
Rasmus Munk Larsen
5e521f3e45 Revert "add realview test" 
This reverts merge request !2352
 2026-03-28 17:27:01 -07:00
Charles Schlosser
87ae1dbe7f add realview test 
libeigen/eigen!2352
 2026-03-28 16:26:51 -07:00
Rasmus Munk Larsen
9706546a14 Add Householder blocked-right regression test 
libeigen/eigen!2348

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-27 20:49:43 -07:00
Rasmus Munk Larsen
b8baa2c49c Split eigensolver_selfadjoint test to fix NVHPC OOM 
libeigen/eigen!2347

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-27 18:09:26 -07:00
Rasmus Munk Larsen
f574cb9b18 Fix vectorization_logic test for generic clang backend 
libeigen/eigen!2333

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-22 22:12:57 -07:00
Rasmus Munk Larsen
843ffcec8b Fix warnings reported by NVHPC 26.1 
libeigen/eigen!2324

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-22 11:43:40 -07:00
Florian Maurin
71ef987edb Fixes triangular solves on indexed/sliced dense expressions 
libeigen/eigen!2340

Closes #2814
 2026-03-22 11:12:21 -07:00
Rasmus Munk Larsen
6490b17e6f Fix sanitizer regressions in sparse serializer and packet tests 
libeigen/eigen!2319

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-22 09:10:16 -07:00
Pavel Guzenfeld
a0e30732a7 Remove trailing semicolon from EIGEN_UNUSED_VARIABLE macro 
libeigen/eigen!2301

Closes #3007

Co-authored-by: Pavel Guzenfeld <67074795+PavelGuzenfeld@users.noreply.github.com>
 2026-03-21 16:54:13 -07:00
Rasmus Munk Larsen
54b04fc6b1 Fix mixed-type GEMM packing for backends without half/quarter packets 
libeigen/eigen!2297

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-21 09:46:54 -07:00
Pavel Guzenfeld
1d21d62fbc Fix computeInverseAndDetWithCheck for dynamic result matrices 
libeigen/eigen!2312

Closes #2917
 2026-03-21 08:38:27 -07:00
Rasmus Munk Larsen
cc8c7cf0e6 Fix bugs and clean up SparseCore module 
libeigen/eigen!2250

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-21 13:31:12 +00:00
Rasmus Munk Larsen
8115b45e50 Fix integer sanitizer issues in shifts and test ranges 
libeigen/eigen!2320

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-20 17:27:02 -07:00
Rasmus Munk Larsen
6540bf4787 Harden unsupported tensor tests for sanitizers 
libeigen/eigen!2321

Co-authored-by: Rasmus Munk Larsen <rlarsen@nvidia.com>
Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-20 15:12:41 -07:00
Rasmus Munk Larsen
a0b16a7e1b Fix flaky product and eigensolver_selfadjoint tests 
libeigen/eigen!2326

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-20 13:44:03 -07:00
Pavel Guzenfeld
a72172e563 Add blocking and vectorization boundary tests for LU and Cholesky 
libeigen/eigen!2317
 2026-03-20 13:27:49 -07:00
Rasmus Munk Larsen
8a47aa334b Replace empirical product test tolerances with principled Higham-Mary bounds 
libeigen/eigen!2292

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-20 11:03:58 -07:00
Pavel Guzenfeld
05295a818b Fix undefined behavior in matrix_cwise test for signed integers 
libeigen/eigen!2310

Closes #2933
 2026-03-18 11:51:01 -07:00
Rasmus Munk Larsen
ea13a98dec Fix imag_ref for real scalar types and clean up svd_fill.h 
libeigen/eigen!2303

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-15 19:56:01 -07:00
Antonio Sánchez
929785924c Fix more cache size queries. 
libeigen/eigen!2296
 2026-03-14 16:07:44 +00:00
Antonio Sánchez
b2f95d3733 Fix more cache size queries. 
libeigen/eigen!2295
 2026-03-14 15:43:24 +00:00
Rasmus Munk Larsen
c1faa74738 Add boundary test coverage: stableNorm, LinSpaced, complex GEMV, triangular solve 
libeigen/eigen!2291

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-12 18:15:30 -07:00
Rasmus Munk Larsen
6b9275d1a8 Add test coverage for transpose, reverse, bool redux, select, diagonal-of-product at boundaries 
libeigen/eigen!2290

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-12 17:02:58 -07:00
Rasmus Munk Larsen
356a9ba1da Add test coverage for matrix lpNorm, RowMajor partial reductions, selfadjoint boundaries 
libeigen/eigen!2289

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-12 14:45:51 -07:00
Rasmus Munk Larsen
15cae83485 Add test coverage for strided maps, triangular blocking, and mixed storage orders 
libeigen/eigen!2288

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-12 14:07:21 -07:00
Rasmus Munk Larsen
93aa959b8a Add vectorization boundary tests for redux and visitor 
libeigen/eigen!2287

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-12 13:47:15 -07:00
Rasmus Munk Larsen
c93116b43d Improve test coverage for inner product, fill, reductions, and IO 
libeigen/eigen!2286

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-12 12:48:45 -07:00
Rasmus Munk Larsen
5e478d3285 Improve product test coverage at critical code-path boundaries 
libeigen/eigen!2285

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-12 12:32:06 -07:00
Charles Schlosser
875fb48f0a fix various irksome compiler warnings 
libeigen/eigen!2280
 2026-03-11 21:01:20 -07:00
Charles Schlosser
c4eb3c4f4c fix custom visitors 
libeigen/eigen!2275

Closes #2920
 2026-03-11 10:52:49 +00:00
Antonio Sánchez
4387e32481 Fix row-skipping bug in general_matrix_vector_product::run_small_cols 
libeigen/eigen!2276
 2026-03-10 15:16:00 -07:00
Rasmus Munk Larsen
42b6c43cfe Revert "Remove random retry loops in tests (batch 2: indices and integer types)" 
This reverts merge request !2261
 2026-03-09 20:01:53 -07:00
Rasmus Munk Larsen
54458cb39d Remove random retry loops in tests (batch 3: geometry, sparse, umeyama) 
libeigen/eigen!2262

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-09 00:35:26 -07:00
Rasmus Munk Larsen
8eaa7552fe Fix three flaky tests: packetmath, array_cwise, polynomialsolver 
libeigen/eigen!2267

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-08 14:59:23 -07:00
Rasmus Munk Larsen
dd81698aed Fix vectorization_logic test for wide SIMD widths 
libeigen/eigen!2266

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-08 12:11:33 -07:00
Rasmus Munk Larsen
ab58784268 Remove random retry loops in tests (batch 5: geometry, mixing types, triangular) 
libeigen/eigen!2264

Co-authored-by: Rasmus Munk Larsen <rmlarsen@gmail.com>
 2026-03-08 11:51:35 -07:00