Skip to content

[Feature] Optimize forward metadata collection across dp ranks #1857

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
jianzs wants to merge 1 commit into from
Closed

[Feature] Optimize forward metadata collection across dp ranks #1857

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
22 changes: 9 additions & 13 deletions vllm_ascend/worker/model_runner_v1.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,6 @@
import numpy.typing as npt
import torch
import torch._dynamo.cache_size
import torch.distributed as dist
import torch.nn as nn
from tqdm import tqdm # type: ignore
from vllm.attention import AttentionType, get_attn_backend
Expand Down Expand Up @@ -596,18 +595,15 @@ def _update_states(self, scheduler_output: "SchedulerOutput") -> None:
def _get_forward_metadata_across_dp(
self, num_tokens: int, with_prefill: bool,
enable_dbo: bool) -> tuple[torch.Tensor, bool, bool]:

# Compose: all_reduce metadata (num_tokens of each rank, with_prefill, enable_dbo)
num_tokens_across_dp = torch.zeros(self.dp_size + 2,
dtype=torch.int32,
device="cpu")
num_tokens_across_dp[self.dp_rank] = num_tokens
num_tokens_across_dp[-2] = int(with_prefill)
num_tokens_across_dp[-1] = int(not enable_dbo)
dist.all_reduce(num_tokens_across_dp, group=get_dp_group().cpu_group)
with_prefill = bool(num_tokens_across_dp[-2])
enable_dbo = not bool(num_tokens_across_dp[-1])
num_tokens_across_dp = num_tokens_across_dp[:-2]
local_forward_metadata = torch.tensor(
Copy link
Copy Markdown
Collaborator

@momo609 momo609 Sep 14, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please provide the data for the memory increase.

[[num_tokens, with_prefill, enable_dbo]],
device="npu",
dtype=torch.int32)
global_forward_metadata = get_dp_group().all_gather(
local_forward_metadata, dim=0)
num_tokens_across_dp = global_forward_metadata[:, 0].cpu()
with_prefill = bool(global_forward_metadata[:, 1].any())
enable_dbo = bool(global_forward_metadata[:, 2].any())
return num_tokens_across_dp, with_prefill, enable_dbo

def _get_forward_metadata_across_dp_and_pad(
Expand Down
Loading