overhead_experiment() — pytorch Function Reference

Architecture documentation for the overhead_experiment() function in common.py from the pytorch codebase.

Function python DynamoBenchmarks Microbenchmarking calls 1

Entity Profile

DynamoBenchmarks→ Microbenchmarking→ overhead_experiment() — pytorch Function Reference

Dependency Diagram

graph TD
  eeda7f55_d302_b33d_bc04_70a1b44397fa["overhead_experiment()"]
  04a3a4a6_8db3_854d_a893_02c9542bf9dd["speedup_experiment()"]
  eeda7f55_d302_b33d_bc04_70a1b44397fa -->|calls| 04a3a4a6_8db3_854d_a893_02c9542bf9dd
  style eeda7f55_d302_b33d_bc04_70a1b44397fa fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

benchmarks/dynamo/common.py lines 1220–1227

def overhead_experiment(*args, model_iter_fn):
    """
    Measure overheads of TorchDynamo by running with no backend (only
    eager+FX), and reporting speedup/slowdown over eager.

    Writes to ./overheads.csv
    """
    return speedup_experiment(*args, model_iter_fn)