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

Reduce memory usage during culling for shuffling and merge #8197

Merged
merged 4 commits into from
Sep 20, 2023
Merged

Reduce memory usage during culling for shuffling and merge #8197

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
22925f6
Reduce memory usage during culling for shuffling and merge
phofl Sep 20, 2023
3ee8fe6
Update distributed/shuffle/_merge.py
phofl Sep 20, 2023
d538c6d
Add comment
phofl Sep 20, 2023
e4cd8f9
Add comment
phofl Sep 20, 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
14 changes: 6 additions & 8 deletions distributed/shuffle/_merge.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,6 @@
# mypy: ignore-errors
from __future__ import annotations

from collections import defaultdict
from collections.abc import Iterable, Sequence
from typing import TYPE_CHECKING, Any

Expand Down Expand Up @@ -243,15 +242,14 @@ def _cull_dependencies(
all input partitions. This method does not require graph
materialization.
"""
deps = defaultdict(set)
deps = {}
parts_out = parts_out or self._keys_to_parts(keys)
keys = {(self.name_input_left, i) for i in range(self.npartitions)}
keys |= {(self.name_input_right, i) for i in range(self.npartitions)}
# Protect against mutations later on with frozenset
keys = frozenset(keys)
Copy link
Member

Choose a reason for hiding this comment

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

maybe add a comment why this is a frozenset. Using a frozenset is a bit uncommon and if we're honest this is just paranoia / testing

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

added

for part in parts_out:
deps[(self.name, part)] |= {
(self.name_input_left, i) for i in range(self.npartitions)
}
deps[(self.name, part)] |= {
(self.name_input_right, i) for i in range(self.npartitions)
}
deps[(self.name, part)] = keys
return deps

def _keys_to_parts(self, keys: Iterable[str]) -> set[str]:
Expand Down
7 changes: 5 additions & 2 deletions distributed/shuffle/_shuffle.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -227,8 +227,11 @@ def cull(
parameter.
"""
parts_out = self._keys_to_parts(keys)
input_parts = {(self.name_input, i) for i in range(self.npartitions_input)}
culled_deps = {(self.name, part): input_parts.copy() for part in parts_out}
# Protect against mutations later on with frozenset
input_parts = frozenset(
{(self.name_input, i) for i in range(self.npartitions_input)}
)
culled_deps = {(self.name, part): input_parts for part in parts_out}

if parts_out != set(self.parts_out):
culled_layer = self._cull(parts_out)
Expand Down