biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Feature] Prefill Context Parallel (PCP) basic support (#28718)

Browse Source 
Signed-off-by: QiuChunshuo <qiuchunshuo@huawei.com>
Signed-off-by: FENP <yuanyongjie.yyj@antgroup.com>
Signed-off-by: LookAround <lixushi@huawei.com>
Signed-off-by: Jingchun Gao <gaojingchun1@huawei.com>
Signed-off-by: zhenwenqi2024 <zhenwenqi_2022@qq.com>
Co-authored-by: FENP <yuanyongjie.yyj@antgroup.com>
Co-authored-by: LookAround <lixushi@huawei.com>
Co-authored-by: Jingchun Gao <gaojingchun1@huawei.com>
Co-authored-by: zhenwenqi2024 <zhenwenqi_2022@qq.com>
Co-authored-by: Jingchun Gao <63247409+gjc0824@users.noreply.github.com>
This commit is contained in:
Qiu
2025-11-20 04:52:44 +08:00
committed by GitHub
parent 02f5903b84
commit 2fd893b4ce
27 changed files with 399 additions and 114 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
vllm/engine/arg_utils.py
View File

@@ -389,8 +389,10 @@ class EngineArgs:
nnodes: int = ParallelConfig.nnodes
node_rank: int = ParallelConfig.node_rank
tensor_parallel_size: int = ParallelConfig.tensor_parallel_size
prefill_context_parallel_size: int = ParallelConfig.prefill_context_parallel_size
decode_context_parallel_size: int = ParallelConfig.decode_context_parallel_size
dcp_kv_cache_interleave_size: int = ParallelConfig.dcp_kv_cache_interleave_size
cp_kv_cache_interleave_size: int = ParallelConfig.cp_kv_cache_interleave_size
data_parallel_size: int = ParallelConfig.data_parallel_size
data_parallel_rank: int | None = None
data_parallel_start_rank: int | None = None
@@ -770,6 +772,15 @@ class EngineArgs:
"--dcp-kv-cache-interleave-size",
**parallel_kwargs["dcp_kv_cache_interleave_size"],
)
parallel_group.add_argument(
"--cp-kv-cache-interleave-size",
**parallel_kwargs["cp_kv_cache_interleave_size"],
)
parallel_group.add_argument(
"--prefill-context-parallel-size",
"-pcp",
**parallel_kwargs["prefill_context_parallel_size"],
)
parallel_group.add_argument(
"--data-parallel-size", "-dp", **parallel_kwargs["data_parallel_size"]
)
@@ -1600,6 +1611,7 @@ class EngineArgs:
parallel_config = ParallelConfig(
pipeline_parallel_size=self.pipeline_parallel_size,
tensor_parallel_size=self.tensor_parallel_size,
prefill_context_parallel_size=self.prefill_context_parallel_size,
data_parallel_size=self.data_parallel_size,
data_parallel_rank=self.data_parallel_rank or 0,
data_parallel_external_lb=data_parallel_external_lb,
@@ -1631,6 +1643,7 @@ class EngineArgs:
worker_extension_cls=self.worker_extension_cls,
decode_context_parallel_size=self.decode_context_parallel_size,
dcp_kv_cache_interleave_size=self.dcp_kv_cache_interleave_size,
cp_kv_cache_interleave_size=self.cp_kv_cache_interleave_size,
_api_process_count=self._api_process_count,
_api_process_rank=self._api_process_rank,
)
@@ -1952,6 +1965,15 @@ class EngineArgs:
default_prefix_caching,
) = self.get_chunked_prefill_prefix_caching_defaults(model_config)
if self.prefill_context_parallel_size > 1:
default_chunked_prefill = False
default_prefix_caching = False
logger.warning(
"--prefill-context-parallel-size > 1 is not compatible with "
"chunked prefill and prefix caching now. Chunked prefill "
"and prefix caching have been disabled by default."
)
if self.enable_chunked_prefill is None:
self.enable_chunked_prefill = default_chunked_prefill