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

Sync xla_args before computation. #5823

Merged
merged 2 commits into from
Nov 21, 2023
Merged

Sync xla_args before computation. #5823

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
6b62e8b
Sync `xla_args` before actually computing them.
ysiraichi Nov 18, 2023
99e6966
Only sync `FunctionalTensorWrapper` tensors.
ysiraichi Nov 18, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 11 additions & 0 deletions test/dynamo/test_bridge.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -240,6 +240,17 @@ def foo(x):
x = torch.randint(0, 10, (10,), device=device)
foo(x)

def test_inputs_not_computed(self):

@torch.compile(backend="openxla")
def foo(x):
return x * 2

device = xm.xla_device()
x = torch.rand(5, device=device)
x = x.unsqueeze(dim=-1)
foo(x)


if __name__ == "__main__":
from torch._dynamo.test_case import run_tests
Expand Down
6 changes: 6 additions & 0 deletions torch_xla/core/dynamo_bridge.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -455,6 +455,12 @@ def call_module(self, target, args, kwargs):


def extract_compiled_graph(xla_model: torch.fx.GraphModule, xla_args):
# Synchronize xla_args, so that each FunctionalTensorWrapper argument updates its
# value reference before actually computing it.
for a in xla_args:
if isinstance(a, torch.Tensor) and torch._is_functional_tensor(a):
torch._functionalize_sync(a)
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

first time learnt this api.. interesting.


# This call is critical to make sure xla_args' tensor id show up in graph_input_tensor_ids
xm.mark_step()

Expand Down