Skip to content

[Engine] reject duplicate GPUs in automatic replica groups #3387

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

Open
j-luwierski wants to merge 1 commit into
base: main
Choose a base branch
from
Open

[Engine] reject duplicate GPUs in automatic replica groups #3387

Show file tree
Hide file tree
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
11 changes: 6 additions & 5 deletions serve/replicas.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,11 +56,12 @@ def gpu_groups(cfg: dict) -> list[list[int]]:
raise ReplicaConfigError(f'"replicas": {spec} does not divide the {len(cards)} cards {cards} evenly: name the '
f'groups ("replicas": [{{"gpus": [0, 1]}}, ...]) or change the count')
n = len(cards) // spec
return [cards[i * n:(i + 1) * n] for i in range(spec)]
groups = []
for e in spec:
g = e.get("gpus") if isinstance(e, dict) else e
groups.append(_cards(g))
groups = [cards[i * n:(i + 1) * n] for i in range(spec)]
else:
groups = []
for e in spec:
g = e.get("gpus") if isinstance(e, dict) else e
groups.append(_cards(g))
if any(not g for g in groups):
raise ReplicaConfigError('every replica needs its cards: {"gpus": [0, 1]}')
flat = [c for g in groups for c in g]
Expand Down
19 changes: 19 additions & 0 deletions serve/test_replicas.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,25 @@ def test_named_groups(self):
self.assertEqual(gpu_groups({"replicas": [{"gpus": [0, 1]}, {"gpus": [2]}]}), [[0, 1], [2]])
self.assertEqual(gpu_groups({"replicas": [{"gpus": [0, 1]}]}), [])

def test_automatic_groups_reject_duplicate_cards(self):
for cards in ([0, 0], [0, 1, 1, 2], [0, 0, 1, 2], "0,0", "0,1,1,2", "0,00"):
with self.subTest(cards=cards):
with self.assertRaisesRegex(ReplicaConfigError, "a card is in two replicas"):
gpu_groups({"gpu": cards, "replicas": 2})

def test_named_groups_still_reject_duplicate_cards(self):
for groups in ([[0], [0]], [[0, 1], [1, 2]], [[0, 0], [1, 2]], [[0, 0]]):
with self.subTest(groups=groups):
with self.assertRaisesRegex(ReplicaConfigError, "a card is in two replicas"):
gpu_groups({"replicas": [{"gpus": g} for g in groups]})

def test_valid_card_order_and_disabled_replicas_are_unchanged(self):
for cards in ([3, 1, 2, 0], "3,1,2,0"):
self.assertEqual(gpu_groups({"gpu": cards, "replicas": 2}), [[3, 1], [2, 0]])
for count in (None, False, 0, 1):
with self.subTest(count=count):
self.assertEqual(gpu_groups({"gpu": [0, 0], "replicas": count}), [])

def test_mistakes_are_said(self):
for cfg in ({"gpu": [0, 1, 2], "replicas": 2}, {"replicas": 2}, {"replicas": "2"},
{"replicas": [{"gpus": [0, 1]}, {"gpus": [1, 2]}]}, {"replicas": [{"gpus": []}, {"gpus": [1]}]}):
Expand Down
Loading