[main] remove dbo code (#3712)

### What this PR does / why we need it?
Remove codes of dbo.
Currently, vLLM has supported dbo with pr:
https://github.com/vllm-project/vllm/pull/23693.

### Does this PR introduce _any_ user-facing change?

### How was this patch tested?

- vLLM version: v0.11.0rc3
- vLLM main:
17c540a993

Signed-off-by: zzzzwwjj <1183291235@qq.com>
This commit is contained in:
zzzzwwjj
2025-10-25 15:53:01 +08:00
committed by GitHub
parent d9cdc65854
commit e5676fc36e
26 changed files with 69 additions and 1588 deletions

View File

@@ -17,11 +17,8 @@ from vllm.v1.attention.backends.utils import AttentionCGSupport
from vllm_ascend.ascend_config import get_ascend_config
from vllm_ascend.attention.attention_v1 import AscendAttentionState
from vllm_ascend.attention.mla_v1 import AscendMLAMetadata
from vllm_ascend.attention.utils import (AscendCommonAttentionMetadata,
split_decodes_and_prefills)
from vllm_ascend.multistream.base import MSAttentionMetadataSplitConfig
from vllm_ascend.multistream.ms_split import model_input_split_v1_mla_attn
from vllm_ascend.worker.npu_input_batch import InputBatch
if TYPE_CHECKING:
@@ -138,7 +135,6 @@ class AscendSFAMetadata:
decode: Optional[AscendSFADecodeMetadata] = None
prefill: Optional[AscendSFAPrefillMetadata] = None
enable_dbo_across_dp: bool = False
def __post_init__(self):
pass
@@ -149,17 +145,6 @@ class AscendSFAMetadata:
# f"Only {supported_head_sizes} are supported for head_dim,",
# f"received {self.head_dim}.")
def split_metadata_for_multistream(
self,
ms_split_config: MSAttentionMetadataSplitConfig,
) -> list["AscendSFAMetadata"]:
"""Split metadata for multi-stream with AscendSFAMetadata"""
return model_input_split_v1_mla_attn(
ms_split_config=ms_split_config,
attn_metadata=self,
_metadata_cls=AscendMLAMetadata,
)
M = TypeVar("M", bound=AscendSFAMetadata)
@@ -434,7 +419,6 @@ class AscendSFAMetadataBuilder:
query_start_loc=query_start_loc,
block_tables=block_table,
seq_lens=seq_lens,
enable_dbo_across_dp=common_attn_metadata.enable_dbo_across_dp,
)