Skip to content

More aggressive memory freeing from TrainPipelineContext #1967

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

Closed
wants to merge 1 commit into from
Closed

More aggressive memory freeing from TrainPipelineContext #1967

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
2 changes: 1 addition & 1 deletion torchrec/distributed/train_pipeline/tests/pipeline_benchmarks.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -297,7 +297,7 @@ def _func_to_benchmark(
)
if rank == 0:
print(
f" {pipeline_clazz.__name__: <{35}} | Runtime (P90): {result.runtime_percentile(90)/1000:5.1f} s | Memory (P90): {result.max_mem_percentile(90)/1000:5.1f} GB"
f" {pipeline_clazz.__name__: <{35}} | Runtime (P90): {result.runtime_percentile(90)/1000:5.3f} s | Memory (P90): {result.max_mem_percentile(90)/1000:5.3f} GB"
)


Expand Down
2 changes: 2 additions & 0 deletions torchrec/distributed/train_pipeline/train_pipelines.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -378,6 +378,8 @@ def wait_sparse_data_dist(self, context: TrainPipelineContext) -> None:
for names, awaitable in context.fused_splits_awaitables:
for name, request in zip(names, awaitable.wait()):
context.input_dist_tensors_requests[name] = request
context.input_dist_splits_requests.clear()
context.fused_splits_awaitables.clear()

def _copy_batch_to_gpu(self, dataloader_iter: Iterator[In]) -> Optional[In]:
"""
Expand Down
14 changes: 7 additions & 7 deletions torchrec/distributed/train_pipeline/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -187,8 +187,10 @@ def get_context(self) -> TrainPipelineContext:
class PipelinedForward(BaseForward):
# pyre-ignore [2, 24]
def __call__(self, *input, **kwargs) -> Awaitable:
assert self._name in self._context.input_dist_tensors_requests
request = self._context.input_dist_tensors_requests[self._name]
assert (
self._name in self._context.input_dist_tensors_requests
), "Invalid PipelinedForward usage, please do not directly call model.forward()"
request = self._context.input_dist_tensors_requests.pop(self._name)
assert isinstance(request, Awaitable)
with record_function("## wait_sparse_data_dist ##"):
# Finish waiting on the dist_stream,
Expand All @@ -198,6 +200,8 @@ def __call__(self, *input, **kwargs) -> Awaitable:

# Make sure that both result of input_dist and context
# are properly transferred to the current stream.
ctx = self._context.module_contexts.pop(self._name)

if self._stream is not None:
torch.cuda.current_stream().wait_stream(self._stream)
cur_stream = torch.cuda.current_stream()
Expand All @@ -206,13 +210,9 @@ def __call__(self, *input, **kwargs) -> Awaitable:
data, (torch.Tensor, Multistreamable)
), f"{type(data)} must implement Multistreamable interface"
data.record_stream(cur_stream)

ctx = self._context.module_contexts[self._name]
ctx.record_stream(cur_stream)

return self._module.compute_and_output_dist(
self._context.module_contexts[self._name], data
)
return self._module.compute_and_output_dist(ctx, data)


class EmbeddingPipelinedForward(BaseForward):
Expand Down