4 Commits

Author SHA1 Message Date
f345b0e520 Performance optimization for the TwoPunctures module
* Re-enabled OpenMP.

1. Batch spectral derivatives (Chebyshev & Fourier) via precomputed matrices:
Chebyshev/Fourier transforms and derivatives are precomputed as explicit physical-space operator matrices.
Batch DGEMM now applies to entire tensor fields, mathematically identical to original per-line transforms but vastly faster.

2. Gauss-Seidel relaxation & tridiagonal solver workspace reuse:
Per-thread reusable workspaces replace per-call heap new/delete in all tridiagonal and relaxation routines.

3. Efficient OpenMP multithreading throughout relaxation/deriv:
relax_omp and friends parallelize over grouped lines/planes, maximizing threading efficiency and memory independence.

Co-authored-by: copilot-swe-agent[bot] <198982749+copilot@users.noreply.github.com>
2026-02-07 14:48:47 +08:00
f5ed23d687 Revert "Eliminate hot-path heap allocations in TwoPunctures spectral solver"
This reverts commit 09ffdb553d.
2026-02-07 14:45:25 +08:00
09ffdb553d Eliminate hot-path heap allocations in TwoPunctures spectral solver
Pre-allocate workspace buffers as class members to remove ~8M malloc/free
pairs per Newton iteration from LineRelax, ThomasAlgorithm, JFD_times_dv,
J_times_dv, chebft_Zeros, fourft, Derivatives_AB3, and F_of_v.
Rewrite ThomasAlgorithm to operate in-place on input arrays.
Results are bit-identical; no algorithmic changes.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-06 21:20:35 +08:00
f2fc9af70e asc26 amss-ncku initialized 2026-01-13 15:01:15 +08:00