Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[wip][compiled autograd] runtime wrapper for verbose debugging #125417

Closed
wants to merge 2 commits into from
Closed

[wip][compiled autograd] runtime wrapper for verbose debugging #125417

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d83cdbd
[wip][compiled autograd] runtime wrapper for verbose debugging
xmfan May 2, 2024
c39d929
Update on "[wip][compiled autograd] runtime wrapper for verbose debug…
xmfan May 2, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
8 changes: 7 additions & 1 deletion torch/_dynamo/compiled_autograd.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -220,7 +220,13 @@ def end_capture(self, outputs):
"compiled_autograd_graph",
payload_fn=lambda: graph.print_readable(print_output=False),
)
return self.compiler_fn(graph)

def runtime_wrapper(compiled_fn, inputs, sizes, hooks):
# insert debug code here
verbose_log.debug(f"lifted hooks={hooks}")
return compiled_fn(inputs, sizes, hooks)

return runtime_wrapper, self.compiler_fn(graph)

def reorder_accumulate_grad_nodes(self):
"""
Expand Down
7 changes: 5 additions & 2 deletions torch/csrc/dynamo/python_compiled_autograd.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -228,6 +228,7 @@ struct CacheNode {
std::vector<CacheKeyBuffer> key_storage;
std::vector<SizeInput> expected_sizes;

THPObjectPtr runtime_wrapper;
THPObjectPtr compiled_fn;
};

Expand Down Expand Up @@ -504,7 +505,9 @@ CacheNode* _compiled_autograd_impl(
}
}

cache->compiled_fn = check(call_end_capture(py_compiler, state.outputs));
PyObject* res = check(call_end_capture(py_compiler, state.outputs));
cache->runtime_wrapper = PyTuple_GetItem(res, 0);
cache->compiled_fn = PyTuple_GetItem(res, 1);
state.debug_asserts();
} // End cache miss region

Expand Down Expand Up @@ -552,7 +555,7 @@ variable_list compiled_autograd(
&hooks);

THPObjectPtr pyresult(check(PyObject_CallFunctionObjArgs(
cache->compiled_fn.get(), inputs.get(), sizes.get(), hooks.get(), NULL)));
cache->runtime_wrapper.get(), cache->compiled_fn.get(), inputs.get(), sizes.get(), hooks.get(), NULL)));
variable_list outputs = THPVariable_UnpackList(pyresult);
TORCH_INTERNAL_ASSERT(outputs.size() == output_edges.size());
return outputs;
Expand Down