Skip to content

enable VLLM_LIMIT_HPU_GRAPH to disable graph warmup for prompt #66

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

enable VLLM_LIMIT_HPU_GRAPH to disable graph warmup for prompt #66

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
8 changes: 7 additions & 1 deletion vllm/worker/habana_model_runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -308,6 +308,10 @@ def load_model(self) -> None:
self.model = self.lora_manager.create_lora_manager(self.model)

def _use_graphs(self, batch_size, seq_len, is_prompt):
if is_prompt:
if os.environ.get('VLLM_LIMIT_HPU_GRAPH', 'false').lower() == 'true':
return False

if self.enforce_eager:
return False
return (batch_size, seq_len, is_prompt) in self.graphed_buckets
Expand Down Expand Up @@ -999,7 +1003,9 @@ def warmup_model(self, kv_caches: List[torch.Tensor]) -> None:
decode_available_memory = free_mem - prompt_available_memory
prompt_strategy = 'min_tokens'
decode_strategy = os.environ.get('VLLM_GRAPH_DECODE_STRATEGY', 'max_bs')
self.warmup_graphs(prompt_strategy, self.prompt_buckets, True, kv_caches, prompt_available_memory)

if os.environ.get('VLLM_LIMIT_HPU_GRAPH', 'false').lower() != 'true':
self.warmup_graphs(prompt_strategy, self.prompt_buckets, True, kv_caches, prompt_available_memory)
self.warmup_graphs(decode_strategy, self.decode_buckets, False, kv_caches, decode_available_memory)

end_time = time.perf_counter()
Expand Down