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

[zero++] Synchronize at the end of secondary partitioning and simplify the logic #5216

Merged
merged 5 commits into from
Mar 1, 2024
Merged

[zero++] Synchronize at the end of secondary partitioning and simplify the logic #5216

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
644a822
Synchronize at the end of secondary partitioning
ByronHsu Mar 1, 2024
18967fc
trim
ByronHsu Mar 1, 2024
d36ec0c
trim
ByronHsu Mar 1, 2024
9837fcc
simplify parition logic
ByronHsu Mar 1, 2024
00e76e9
only sync current stream
ByronHsu Mar 1, 2024
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
21 changes: 9 additions & 12 deletions deepspeed/runtime/zero/partition_parameters.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1635,19 +1635,16 @@ def _partition_param_sec(self, param, buffer=None, has_been_updated=False):
secondary_end = secondary_start + secondary_partition_size

one_dim_param = param.contiguous().view(-1)
start = partition_size * self.rank
end = start + partition_size
if start < param.ds_numel and end <= param.ds_numel:
if secondary_start < param.ds_numel and secondary_end <= param.ds_numel:
sec_src_tensor = one_dim_param.narrow(0, secondary_start, secondary_partition_size)
param.ds_secondary_tensor.copy_(sec_src_tensor)

else:
if start < param.ds_numel:
elements_to_copy = param.ds_numel - start
elements_to_copy_sec = elements_to_copy * param.ds_secondary_tensor_num_of_groups
param.ds_secondary_tensor.narrow(0, 0, elements_to_copy_sec).copy_(
one_dim_param.narrow(0, secondary_start, elements_to_copy_sec))
# ds_numel is unpadded, so the last chunk of the secondary tensor might not be secondary_partition_size
sec_numel = param.ds_numel - secondary_start if secondary_end > param.ds_numel else secondary_partition_size

# copy from full tensor to secondary tensor
param.ds_secondary_tensor.narrow(0, 0,
sec_numel).copy_(one_dim_param.narrow(0, secondary_start, sec_numel))

# TODO: This is a temporary fix to avoid the issue that 2nd tensor all-gather happens before 2nd tensor partition is done
get_accelerator().current_stream().synchronize()

print_rank_0(f"{param.ds_id} partitioned type {param.dtype} dev {param.device} shape {param.shape}",
force=False)
Expand Down
Loading