Skip to content

quick refactor on _get_group_assignment #5318

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

Merged
merged 1 commit into from
Jul 18, 2023
Merged

quick refactor on _get_group_assignment #5318

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
7 changes: 3 additions & 4 deletions torch_xla/experimental/xla_sharding.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -335,12 +335,11 @@ def _get_tile_assignment(mesh: Mesh,
return mesh_list_tensor.permute(partition_spec_list).tolist()


def _get_group_assignment(sharding_type: ShardingType, mesh: Mesh,
def _get_group_assignment(sharding_type: ShardingType,
partition_spec: Tuple[Union[int, None]],
tile_assignment: List) -> Tuple[List, List]:
group_assignment = list()
replication_groups = list()
# TODO(JackCaoG): 3d mesh on 2d tensor
mesh_shape_list = list(torch.tensor(tile_assignment).size())
if sharding_type is ShardingType.PARTIAL:
# Shard across groups and replicate within subgroups; replicated dims
Expand Down Expand Up @@ -426,7 +425,7 @@ def mark_sharding(t: Union[torch.Tensor, XLAShardedTensor], mesh: Mesh,
tile_assignment = _get_tile_assignment(mesh, partition_spec)
sharding_type = _get_sharding_type(partition_spec, num_devices)
group_assignment, replication_groups = _get_group_assignment(
sharding_type, mesh, partition_spec, tile_assignment)
sharding_type, partition_spec, tile_assignment)

def tensor_squeeze(t, tensor_expand):
if tensor_expand:
Expand Down Expand Up @@ -482,7 +481,7 @@ def __post_init__(self):
self._sharding_type = _get_sharding_type(partition_spec,
xr.global_device_count())
self._group_assignment, self._replication_groups = _get_group_assignment(
self._sharding_type, mesh, partition_spec, self._tile_assignment)
self._sharding_type, partition_spec, self._tile_assignment)

def xla_spec(self, t: torch.Tensor) -> Union['XlaShardingSpec', None]:
"""
Expand Down