process_caching_precompile() — pytorch Function Reference

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

Function python DynamoBenchmarks Configuration called by 1

Entity Profile

DynamoBenchmarks→ Configuration→ process_caching_precompile() — pytorch Function Reference

Dependency Diagram

graph TD
  87423322_176a_b228_ba7e_98f801a037cf["process_caching_precompile()"]
  27245bba_7bce_3446_86ed_29630ca8fe8a["process_entry()"]
  27245bba_7bce_3446_86ed_29630ca8fe8a -->|calls| 87423322_176a_b228_ba7e_98f801a037cf
  style 87423322_176a_b228_ba7e_98f801a037cf fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

benchmarks/dynamo/common.py lines 3712–3725

def process_caching_precompile():
    """
    After every process_entry, save precompile artifacts to DynamoCache
    """
    if not torch._dynamo.config.caching_precompile:
        raise AssertionError(
            "Caching precompile should be enabled with --caching-precompile"
        )
    from torch._dynamo.precompile_context import PrecompileContext

    debug_info = PrecompileContext.save_to_dynamo_cache()
    print(
        f"Saved {len(debug_info['dynamo'])} precompile artifacts with {len(debug_info['backends'])} backends"
    )