Affine Transformation

This is the smallest complete pipeline in the repository: one worker stage that computes

output_vector = transform_matrix @ input_vector + offset_vector

Why this example matters

Use this example when you want the shortest path to understanding:

  • stream definitions

  • one-stage kernel wiring

  • loading static parameter buffers through shared memory

  • validating outputs numerically

Files

  • examples/affine_transformation/pipeline.yaml

  • examples/affine_transformation/run_example.py

Run it

python examples/affine_transformation/run_example.py

What it demonstrates

  • a CPU shared-memory pipeline

  • one trigger input stream

  • auxiliary streams for matrix and bias terms

  • deterministic end-to-end verification against a local reference computation

Useful follow-up commands

shmpipeline validate examples/affine_transformation/pipeline.yaml
shmpipeline describe examples/affine_transformation/pipeline.yaml
shmpipeline run examples/affine_transformation/pipeline.yaml --duration 1.0

When to study this first

Start here if you are new to the package or if you are verifying a fresh install.