Skip to content

[perf] Fix MLAPO weight disposal for KV-consumer MLA in PD-mix deploy... #5192

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
zzzzwwjj merged 1 commit into from
Jan 5, 2026
Merged

[perf] Fix MLAPO weight disposal for KV-consumer MLA in PD-mix deploy... #5192

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
15 changes: 15 additions & 0 deletions vllm_ascend/attention/mla_v1.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -909,6 +909,21 @@ def _process_weights_for_fused_mlapo(self, act_dtype: torch.dtype):
self.ctkv_scale = torch.tensor([1], dtype=act_dtype, device=device)
self.q_nope_scale = torch.tensor([1], dtype=act_dtype, device=device)

# On KV consumers (decode-only) MLAPO uses the transformed weights built above;
# the original fused_qkv_a_proj/q_proj weights and quant params are no longer
# referenced, so drop them to save memory.
ascend_config = get_ascend_config()
if self.vllm_config.kv_transfer_config is not None and \
Comment thread
kiscad marked this conversation as resolved.
self.vllm_config.kv_transfer_config.is_kv_consumer and \
ascend_config.recompute_scheduler_enable:
self.fused_qkv_a_proj.weight = None
self.fused_qkv_a_proj.deq_scale = None
self.fused_qkv_a_proj.quant_bias = None
self.q_proj.weight = None
self.q_proj.deq_scale = None
self.q_proj.quant_bias = None
torch.npu.empty_cache()
Comment thread
kiscad marked this conversation as resolved.

def get_context_seq_len_npu(self, index: int,
attn_metadata: AscendMLAMetadata):
prefill_metadata = attn_metadata.prefill
Expand Down
11 changes: 8 additions & 3 deletions vllm_ascend/attention/sfa_v1.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -371,7 +371,7 @@ def __init__(
if self.enable_sfa_cp:
self.local_num_heads = self.num_heads * self.tp_size

#TODO: Temporarily adapt sfa-cp, remove after adapting near PCP. --clrs97
# TODO: Temporarily adapt sfa-cp, remove after adapting near PCP. --clrs97
self._replace_linear_class_for_sfa_cp()
from vllm_ascend.distributed.parallel_state import \
get_shared_weight_group
Expand Down Expand Up @@ -537,7 +537,7 @@ def exec_kv(
cache_mode=cache_mode,
is_output_kv=True,
)
#TODO: Temporarily adapt SFA-CP and replace it later with PCP. --clrs97
# TODO: Temporarily adapt SFA-CP and replace it later with PCP. --clrs97
k_pe = get_tp_group().all_gather(k_pe, 0)
k_nope = get_tp_group().all_gather(k_nope, 0)

Expand Down Expand Up @@ -659,8 +659,13 @@ def _process_weights_for_fused_mlapo(self, act_dtype: torch.dtype):
self.ctkv_scale = torch.tensor([1], dtype=act_dtype, device=device)
self.q_nope_scale = torch.tensor([1], dtype=act_dtype, device=device)

# On KV consumers (decode-only) MLAPO uses the transformed weights built above;
# the original fused_qkv_a_proj/q_proj weights and quant params are no longer
# referenced, so drop them to save memory.
ascend_config = get_ascend_config()
if self.vllm_config.kv_transfer_config is not None and \
self.vllm_config.kv_transfer_config.is_kv_consumer:
self.vllm_config.kv_transfer_config.is_kv_consumer and \
ascend_config.recompute_scheduler_enable:
self.fused_qkv_a_proj.weight = None
self.fused_qkv_a_proj.deq_scale = None
self.fused_qkv_a_proj.quant_bias = None
Expand Down
Loading