Backend and delegation

• https://pytorch.org/executorch/stable/compiler-delegate-and-partitioner.html

• A delegate backend implementation is composed of:

  • An ahead-of-time preprocessing interface
    • program preprocessing (e.g. ahead of time compilation, transformation, optimization…).
  • A runtime initialization and execution interface
    • Program initialization (e.g. runtime compilation).
    • Program execution.
    • (optional) Program destroy (e.g. release backend owned resource)

Preprocessing

• Before any serialization, we apply passes on the subgraphs to prepare the graph.
  • These passes are essentially graph transformations that help improve the performance of the delegate.
  • Examples: Passes or transformation