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.

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

Fix layout convert caching #5515

Merged
merged 2 commits into from
Mar 26, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
5 changes: 4 additions & 1 deletion colossalai/tensor/d_tensor/layout_converter.py
Original file line number Diff line number Diff line change
Expand Up @@ -440,7 +440,10 @@ def layout_converting(
total_steps = 0
transform_path = []
comm_action_sequence: List[CommSpec] = []
spec_pairs = (str(source_spec.sharding_sequence), str(target_spec.sharding_sequence))

src_shape = source_layout.get_sharded_shape_per_device()
dst_shape = target_layout.get_sharded_shape_per_device()
spec_pairs = ((str(source_spec.sharding_sequence), src_shape), (str(target_spec.sharding_sequence), dst_shape))

if spec_pairs in self.cached_solution:
# Solution Cache hit
Expand Down
11 changes: 9 additions & 2 deletions tests/test_tensor/test_dtensor/test_layout_converter.py
Original file line number Diff line number Diff line change
Expand Up @@ -123,8 +123,15 @@ def check_layout_converting(rank, world_size, port):
assert comm_action_sequence[2].logical_process_axis == 1

# checkout chached_spec_pairs_transform_path
assert layout_converter.cached_solution[("[R, S01, R]", "[S01, R, R]")][0] == transform_path
assert layout_converter.cached_solution[("[R, S01, R]", "[S01, R, R]")][1] == comm_action_sequence
src_shape = source_layout.get_sharded_shape_per_device()
dst_shape = target_layout.get_sharded_shape_per_device()
assert (
layout_converter.cached_solution[(("[R, S01, R]", src_shape), ("[S01, R, R]", dst_shape))][0] == transform_path
)
assert (
layout_converter.cached_solution[(("[R, S01, R]", src_shape), ("[S01, R, R]", dst_shape))][1]
== comm_action_sequence
)

comm_cost = layout_converter.get_total_comm_cost(source_layout, target_layout)

Expand Down
Loading