Rasmus Munk Larsen 8ddbe44799 Add small fixed-size matrix benchmarks for robotics/CV workloads ...

Benchmark the operations that dominate robotics and computer vision
inner loops: fixed-size matrix multiply, matrix-vector, inverse,
determinant, LLT, LDLT, PartialPivLU, ColPivHouseholderQR, JacobiSVD,
and SelfAdjointEigenSolver for sizes 2x2 through 8x9.

Key findings from the baseline measurements:
- MatMul/MatVec: excellent (<1ns for 3x3 float)
- Inverse 3x3: excellent (3.4ns)
- LLT 3x3→4x4: 8x jump (3.9→31.7ns float) due to inlining threshold
- ColPivQR 3x3: 166ns — expensive for such a small matrix
- JacobiSVD 3x3: 498ns double — the main CV bottleneck
- SelfAdjointEig: computeDirect() is 3.2x faster than iterative for 3x3
  (71ns vs 230ns) — many users may not know this API exists

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-04-04 15:06:32 -07:00

..

bench_block_ops.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_broadcasting.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_construction.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_cwise_math.cpp

Vectorize asinh and acosh for float and double

2026-04-01 21:46:36 -07:00

bench_diagonal.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_dot.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_fixed_size.cpp

Add benchmarks for unsupported modules and extend supported benchmarks

2026-02-24 17:12:33 -08:00

bench_gemm.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_gemv.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_map.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_move_semantics.cpp

Reorganize benchmarks into subdirectories and clean up Eigen sources

2026-02-21 17:46:55 -08:00

bench_reductions.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_reverse.cpp

Reorganize benchmarks into subdirectories and clean up Eigen sources

2026-02-21 17:46:55 -08:00

bench_selfadjoint_product.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_small_matrix.cpp

Add small fixed-size matrix benchmarks for robotics/CV workloads

2026-04-04 15:06:32 -07:00

bench_symv.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_syr2.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_syr.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_triangular_product.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_trmv.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_trsm.cpp

Remove benchmark::internal::Benchmark* from all benchmarks

2026-03-20 17:42:07 -07:00

bench_vecadd.cpp

Reorganize benchmarks into subdirectories and clean up Eigen sources

2026-02-21 17:46:55 -08:00

CMakeLists.txt

Add small fixed-size matrix benchmarks for robotics/CV workloads

2026-04-04 15:06:32 -07:00