Benchmarks¶

These plots track the runtime of the standalone nanobench benchmarks in benchmarks/. The Benchmarks workflow builds them in Release on every push to main, runs them, and publishes the per-benchmark JSON reports onto the orphan benchmarks branch. The documentation build fetches those reports and renders, for each benchmark series, the median runtime over the report date so that performance regressions become visible.

The table shows the most recent measurement for every series; each plot below it shows that series’ history. Error bars correspond to nanobench’s median-absolute-percent-error; hover over a point to see the commit it was measured at.

Benchmark

Median runtime

Error

Date

assembly_kernels__continuous_lagrange / 2d__laplace__p1

90.134 ms

0.0 %

2026-06-28

assembly_kernels__continuous_lagrange / 2d__laplace__p2

1.372 s

0.0 %

2026-06-28

assembly_kernels__continuous_lagrange / 2d__laplace__p3

7.267 s

0.0 %

2026-06-28

assembly_kernels__continuous_lagrange / 2d__product__p1

86.955 ms

0.0 %

2026-06-28

assembly_kernels__continuous_lagrange / 2d__product__p2

1.365 s

0.0 %

2026-06-28

assembly_kernels__continuous_lagrange / 2d__product__p3

7.580 s

0.0 %

2026-06-28

assembly_kernels__continuous_lagrange / 3d__laplace__p1

39.489 ms

0.0 %

2026-06-28

assembly_kernels__continuous_lagrange / 3d__laplace__p2

1.697 s

0.0 %

2026-06-28

assembly_kernels__continuous_lagrange / 3d__product__p1

35.752 ms

0.0 %

2026-06-28

assembly_kernels__continuous_lagrange / 3d__product__p2

1.617 s

0.0 %

2026-06-28

assembly_kernels__sipdg / sipdg_p1__assemble_only

22.117 ms

0.0 %

2026-06-28

burgers__1d__explicit__fv / explicit_euler__upwind

3.957 s

0.0 %

2026-06-28

stationary_heat_equation__ESV2007 / sipdg_p1__assemble_and_solve

17.683 ms

0.0 %

2026-06-28

stokes__taylorhood / taylor_hood_order2__assemble_and_solve

101.280 ms

0.0 %

2026-06-28