Basic AO System
This example builds a compact adaptive-optics style control chain on CPU streams.
Pipeline stages
Shack-Hartmann centroid extraction
gain and offset correction
flattening into a slope vector
affine reconstruction
leaky-integrator control
Files
examples/basic_ao_system/pipeline.yamlexamples/basic_ao_system/run_example.pyexamples/basic_ao_system/run_benchmark.pyexamples/basic_ao_system/benchmark_120x120.yaml
Run the end-to-end example
python examples/basic_ao_system/run_example.py
Run the benchmark-oriented workload
python examples/basic_ao_system/run_benchmark.py
What it demonstrates
a multi-stage pipeline with meaningful intermediate products
centroid, scale-offset, flatten, affine, and control kernels working together
stage-by-stage numerical validation
a path from simple toy pipelines toward larger AO-style systems
Why this example matters
This is the best midpoint between the minimal affine example and the larger observatory-scale example. It is large enough to show graph structure clearly without the size and runtime cost of the observatory workload.