Skip to content

Fix num_gpus when TP > 1 #1852

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 3 commits into from
Dec 3, 2023
Merged

Fix num_gpus when TP > 1 #1852

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
62b90d1
Fix
WoosukKwon Nov 30, 2023
6a21850
Merge branch 'main' into fix-tp-mem
WoosukKwon Dec 3, 2023
6d629e2
Fix num_gpus when engine_use_ray=True
WoosukKwon Dec 3, 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
11 changes: 10 additions & 1 deletion vllm/engine/async_llm_engine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -301,7 +301,16 @@ def _init_engine(self, *args,
elif self.worker_use_ray:
engine_class = ray.remote(num_cpus=0)(self._engine_class).remote
else:
engine_class = ray.remote(num_gpus=1)(self._engine_class).remote
# FIXME(woosuk): This is a bit hacky. Be careful when changing the
# order of the arguments.
cache_config = args[1]
parallel_config = args[2]
if parallel_config.tensor_parallel_size == 1:
num_gpus = cache_config.gpu_memory_utilization
else:
num_gpus = 1
engine_class = ray.remote(num_gpus=num_gpus)(
self._engine_class).remote
return engine_class(*args, **kwargs)

async def engine_step(self) -> bool:
Expand Down
6 changes: 5 additions & 1 deletion vllm/engine/llm_engine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -158,9 +158,13 @@ def _init_workers_ray(self, placement_group: "PlacementGroup",
for bundle in placement_group.bundle_specs:
if not bundle.get("GPU", 0):
continue
if self.parallel_config.tensor_parallel_size == 1:
num_gpus = self.cache_config.gpu_memory_utilization
else:
num_gpus = 1
worker = ray.remote(
num_cpus=0,
num_gpus=self.cache_config.gpu_memory_utilization,
num_gpus=num_gpus,
scheduling_strategy=PlacementGroupSchedulingStrategy(
placement_group=placement_group,
placement_group_capture_child_tasks=True),
Expand Down