nystrom_preconditioner
Build and apply the inverse action of the Nyström preconditioner for regularized PSD linear systems.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- nystrom_preconditioner
- description
- Build and apply the inverse action of the Nyström preconditioner for regularized PSD linear systems.
02 资源文件 2
- 脚本scripts/preconditioner.py3.0 KB
- 脚本tests/test_preconditioner.py784 B
03 SKILL.md 正文
Nyström Preconditioner
Use this skill after obtaining U, lambda_hat from a randomized Nyström factorization. It provides the inverse preconditioner action used by Nyström PCG and dense diagnostics for small recovery experiments.
Inputs
U: orthonormal Nyström basis.lambda_hat: nonnegative approximate eigenvalues.mu > 0: regularization parameter.- Residual vector or right-hand-side matrix.
Outputs
apply_inverse_preconditioner(U, lambda_hat, mu, r)result.- Preconditioner metadata, including rank and baseline scale.
- Optional condition-number comparison for
A + mu Iversus the symmetrically preconditioned matrix.
Workflow
- Validate dimensions,
mu, and approximate orthonormality. - Sort eigenpairs by decreasing
lambda_hatif necessary. - Set
sigma = lambda_hat[-1] + mu. - Apply
P^{-1}by scaling captured components withsigma / (lambda_hat_i + mu)and leaving orthogonal components unchanged. - For dense recovery checks, construct the SPD square-root action and estimate condition numbers.
Validation
Run:
python scripts/preconditioner.py --self-test
python tests/test_preconditioner.py
The tests verify inverse-action shape, dimension checks, and condition-number reduction on a diagonal PSD example.
Limitations
Dense condition-number diagnostics are intended for small experiments. Large systems should only use the vector action.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).