Skip to content

[P/D][bugfix]Fix the PCP port mapping error issue #5706

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
wangxiyuan merged 3 commits into from
Jan 10, 2026
+25 −18
Merged

[P/D][bugfix]Fix the PCP port mapping error issue #5706

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
c859ed6
bugfix_pcp_port_mapping
wangxiaoteng888 Jan 8, 2026
a226536
fix_lint
wangxiaoteng888 Jan 9, 2026
d291c22
fix_ut
wangxiaoteng888 Jan 9, 2026
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
19 changes: 10 additions & 9 deletions tests/ut/kv_connector/test_mooncake_connector.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1264,7 +1264,7 @@ def get_kv_split_metadata(use_mla, pcp_size, dcp_size, tp_size,
tp_rank, pcp_rank, _prefill_tp_size,
remote_pcp_size, remote_dcp_size,
remote_port, remote_block_ids,
local_block_ids):
local_block_ids, remote_engine_id):

worker = MooncakeConnectorWorker(self.vllm_config, self.engine_id)

Expand All @@ -1275,7 +1275,7 @@ def get_kv_split_metadata(use_mla, pcp_size, dcp_size, tp_size,
worker.tp_rank = tp_rank
worker.pcp_rank = pcp_rank
worker._prefill_tp_size = _prefill_tp_size
worker.local_remote_block_port_mapping = None
worker.local_remote_block_port_mapping = {}
worker.block_size = 16
worker.num_key_value_heads = 1

Expand All @@ -1289,6 +1289,7 @@ def get_kv_split_metadata(use_mla, pcp_size, dcp_size, tp_size,
meta.num_external_tokens = pcp_size * dcp_size * len(
local_block_ids) * worker.block_size
meta.num_prompt_blocks = pcp_size * dcp_size * len(local_block_ids)
meta.remote_engine_id = remote_engine_id

remote_handshake_port_list, local_block_ids_list, remote_block_ids_list = worker._get_kv_split_metadata(
'0', meta)
Expand All @@ -1297,14 +1298,14 @@ def get_kv_split_metadata(use_mla, pcp_size, dcp_size, tp_size,

self.assertEqual(
get_kv_split_metadata(True, 1, 1, 8, 1, 0, 8, 1, 8, 30000, [1],
[1]),
[1], 0),
([[30001], [30002], [30003], [30004], [30005], [30006], [30007],
[30000]], [[], [], [], [], [], [], [], [1]], [[], [], [], [], [],
[], [], [1]]))

self.assertEqual(
get_kv_split_metadata(False, 1, 1, 8, 1, 0, 8, 2, 8, 30000, [1],
[1]),
[1], 0),
([[30001], [30002], [30003], [30004], [30005], [30006], [30007],
[30008], [30009], [30010], [30011], [30012], [30013], [30014],
[30015], [30000]
Expand All @@ -1314,29 +1315,29 @@ def get_kv_split_metadata(use_mla, pcp_size, dcp_size, tp_size,

self.assertEqual(
get_kv_split_metadata(True, 1, 1, 8, 1, 0, 8, 2, 2, 30000, [1],
[1]),
[1], 0),
([[30001], [30008], [30009], [30000]], [[], [], [], [1]
], [[], [], [], [1]]))

self.assertEqual(
get_kv_split_metadata(False, 1, 1, 8, 1, 0, 8, 2, 2, 30000, [1],
[1]),
[1], 0),
([[30001], [30008], [30009], [30000]], [[], [], [], [1]
], [[], [], [], [1]]))

self.assertEqual(
get_kv_split_metadata(True, 1, 2, 8, 1, 0, 8, 2, 2, 30000, [1],
[1]),
[1], 0),
([[30000], [30008]], [[1], []], [[1], []]))

self.assertEqual(
get_kv_split_metadata(False, 1, 2, 8, 1, 0, 8, 2, 2, 30000, [1],
[1]),
[1], 0),
([[30000], [30008]], [[1], []], [[1], []]))

self.assertEqual(
get_kv_split_metadata(True, 1, 2, 8, 0, 0, 8, 2, 2, 30000,
[1, 2, 3], [1, 2, 3, 4, 5]),
[1, 2, 3], [1, 2, 3, 4, 5], 0),
([[30000], [30008]], [[1, 2, 3], [4, 5]], [[1, 2, 3], [1, 2]]))


Expand Down
24 changes: 15 additions & 9 deletions vllm_ascend/distributed/mooncake_connector.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1154,8 +1154,9 @@ def __init__(self, vllm_config: VllmConfig, engine_id: str):
num_p_block_heads = max(
1, self.num_key_value_heads // self._prefill_tp_size)
self.tp_num_need_pulls = num_d_block_heads // num_p_block_heads
self.local_remote_block_port_mapping = None
self.remote_port_send_num: dict[int, int] = {}
self.local_remote_block_port_mapping: dict[
str, Optional[List[List[int]]]] = {}
self.remote_port_send_num: dict[str, dict[int, int]] = {}

def _get_prefill_decode_size(self, vllm_config: VllmConfig):
# get prefill tp and dp size from extra config
Expand Down Expand Up @@ -1453,16 +1454,20 @@ def get_remote_port_send_num(local_remote_block_port_mappings):
remote_port_send_num[remote_port] += 1
return remote_port_send_num

if self.local_remote_block_port_mapping is None:
if meta.remote_engine_id not in self.local_remote_block_port_mapping:
self.local_remote_block_port_mapping[meta.remote_engine_id] = None
if self.local_remote_block_port_mapping[meta.remote_engine_id] is None:
local_remote_block_port_mappings = get_local_remote_block_port_mappings(
)
self.local_remote_block_port_mapping = local_remote_block_port_mappings[
self.handshake_port]
self.remote_port_send_num = get_remote_port_send_num(
local_remote_block_port_mappings)
self.local_remote_block_port_mapping[
meta.remote_engine_id] = local_remote_block_port_mappings[
self.handshake_port]
self.remote_port_send_num[
meta.remote_engine_id] = get_remote_port_send_num(
local_remote_block_port_mappings)

local_remote_block_port_mapping = copy.deepcopy(
self.local_remote_block_port_mapping)
self.local_remote_block_port_mapping[meta.remote_engine_id])

num_external_blocks = math.ceil(meta.num_external_tokens /
self.block_size)
Expand Down Expand Up @@ -1568,7 +1573,8 @@ def start_load_kv(self, metadata: MooncakeConnectorMetadata):
pcp_dcp_rank][i],
offset=i,
tp_num_need_pulls=self.tp_num_need_pulls,
remote_port_send_num=self.remote_port_send_num,
remote_port_send_num=self.remote_port_send_num[
meta.remote_engine_id],
all_task_done=(
pcp_dcp_rank
== len(remote_handshake_port_list) - 1
Expand Down