Skip to content

Toggle memory profiling start/stop explicitly #213

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 1 commit into from
Feb 15, 2024
Merged

Toggle memory profiling start/stop explicitly #213

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
20 changes: 16 additions & 4 deletions serve/mlc_serve/model/tvm_model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -196,10 +196,6 @@ def get_used_memory(self):
peak_memory = get_used_memory_func(
tvm.device("cuda", 0)
).debug_get_from_remote(0)

# TODO: temp hack to switch the VM allocator to eager recycling mode on all devices
for i in range(1, self.num_shards):
get_used_memory_func(tvm.device("cuda", i)).debug_get_from_remote(i)
else:
params = self.params

Expand Down Expand Up @@ -230,7 +226,23 @@ def profile_memory_usage(self, seq_lens):
positions = copy_to_worker_0(self.disco_session, positions)
seq_lens = copy_to_worker_0(self.disco_session, seq_lens)

start_profiling_func = self.disco_session.get_global_func(
"vm.memory_manager.start_profiling"
)
stop_profiling_func = self.disco_session.get_global_func(
"vm.memory_manager.stop_profiling"
)
else:
start_profiling_func = tvm.get_global_func(
"vm.memory_manager.start_profiling"
)
stop_profiling_func = tvm.get_global_func(
"vm.memory_manager.stop_profiling"
)

start_profiling_func()
self.mod["evaluate"](input_ids, positions, seq_lens, self.params)
stop_profiling_func()

return self.get_used_memory()

Expand Down