xla::FlattenCallGraph Class Reference

#include <flatten_call_graph.h>

Inherits HloPassInterface.

Detailed Description

Google docs:

Flattening associates each call site with a unique computation (for sequential calling contexts) This simplifies buffer assignment and points-to analysis (see b/36865746 for details).

---

The documentation for this class was generated from the following files:

• tensorflow/tensorflow/compiler/xla/service/flatten_call_graph.h
• tensorflow/tensorflow/compiler/xla/service/flatten_call_graph.cc