Update on "[compiled autograd] Always proxy autograd.Function nodes; handle AOT backwards"

zou3519 · zou3519 · commit db79cf9775cd · 2025-01-23T20:01:14.000-08:00
We will always proxy autograd.Function nodes in compiled autograd's
initial graph capture (previously there was an
option to proxy vs trace into the autograd.Function)

We have some requirements for the AOTBackward. Compiled Autograd runs
accumulate grad reordering passes on the AOTBackward graph directly
after the initial graph capture, so we can't just proxy a single node for it.

Instead, we:
- proxy the AOTBackward prologue function into the CA graph
- copy-paste the AOTBackward graph into the CA graph
- trace directly through the epilogue (the traced nodes go into the CA
  graph).

Tracing through the epilogue is safe (assuming no Tensor subclasses)
because the only thing the epilogue does is drop some outputs. The
Tensor subclass situation was already broken so this doesn't regress
anything but this PR sets it up to be fixed (in a followup, where we
will proxy "make_subclass" calls into the graph from the epilogue).

Test Plan:
- existing tests

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy yf225 chenyang78 kadeng muchulee8 ColinPeppler amjames desertfire chauhang aakhundov xmfan

[ghstack-poisoned]
diff --git a/torch/csrc/autograd/functions/tensor.cpp b/torch/csrc/autograd/functions/tensor.cpp
@@ -74,8 +74,8 @@ variable_list CopyBackwards::apply_with_saved(
     std::vector<at::TypePtr> schema = {
         IValuePacker<std::array<bool, 2>>::packed_type(),
         IValuePacker<c10::TensorOptions>::packed_type()};
-    const auto& interface = torch::dynamo::autograd::getPyCompilerInterface();
-    interface->bind_function(
+    const auto& pyinterface = torch::dynamo::autograd::getPyCompilerInterface();
+    pyinterface->bind_function(
         saved.get_py_compiler(),
         name(),
         CopyBackwards_apply_functional_ivalue,
@@ -91,8 +91,8 @@ variable_list CopyBackwards::apply_with_saved(
       IValuePacker<std::vector<std::optional<InputMetadata>>>::pack(
           torch::dynamo::autograd::get_input_metadata(next_edges()));
 
-  const auto& interface = torch::dynamo::autograd::getPyCompilerInterface();
-  auto result = interface->call_function(
+  const auto& pyinterface = torch::dynamo::autograd::getPyCompilerInterface();
+  auto result = pyinterface->call_function(
       saved.get_py_compiler(),
       "apply_functional",
       name(),
