Prefill-Decode Disaggregation (Deepseek)

Prefill-Decode Disaggregation (Deepseek)#

开始使用#

vLLM-Ascend 现在支持带有 EP（专家并行）选项的预填充-解码（PD）分离。本指南将逐步介绍如何使用有限资源验证这些功能。

以 Deepseek-r1-w8a8 模型为例，使用 4 台 Atlas 800T A3 服务器部署 "2P1D" 架构。假设预填充服务器的 IP 为 192.0.0.1（预填充 1）和 192.0.0.2（预填充 2），解码器服务器为 192.0.0.3（解码器 1）和 192.0.0.4（解码器 2）。每台服务器使用 8 个 NPU 16 个芯片部署一个服务实例。

验证多节点通信环境#

物理层要求#

  • 物理机必须位于同一 WLAN 中，并具有网络连接。

  • 所有 NPU 必须互连。节点内连接通过 HCCS，节点间连接通过 RDMA。

验证流程#

在每个节点上按顺序执行以下命令。结果必须全部为 success 且状态必须为 UP：

  1. 单节点验证：

在每个节点上按顺序执行以下命令。结果必须全部为 success 且状态必须为 UP：

# Check the remote switch ports
for i in {0..15}; do hccn_tool -i $i -lldp -g | grep Ifname; done 
# Get the link status of the Ethernet ports (UP or DOWN)
for i in {0..15}; do hccn_tool -i $i -link -g ; done
# Check the network health status
for i in {0..15}; do hccn_tool -i $i -net_health -g ; done
# View the network detected IP configuration
for i in {0..15}; do hccn_tool -i $i -netdetect -g ; done
# View gateway configuration
for i in {0..15}; do hccn_tool -i $i -gateway -g ; done

  1. 检查 NPU HCCN 配置：

确保环境中存在 hccn.conf 文件。如果使用 Docker，请将其挂载到容器中。

cat /etc/hccn.conf

  1. 获取 NPU IP 地址

# Get virtual npu ip
for i in {0..15}; do hccn_tool -i $i -vnic -g;done

  1. 获取 superpodid 和 SDID

for i in {0..15}; do npu-smi info -t spod-info -i $i -c 0;npu-smi info -t spod-info -i $i -c 1;done

  1. 跨节点 PING 测试

# Execute on the target node (replace 'x.x.x.x' with virtual npu ip address)
for i in {0..15}; do hccn_tool -i $i -hccs_ping -g address x.x.x.x;done

  1. 检查 NPU TLS 配置

# The tls settings should be consistent across all nodes
for i in {0..15}; do hccn_tool -i $i -tls -g ; done | grep switch

  1. 单节点验证：

在每个节点上按顺序执行以下命令。结果必须全部为 success 且状态必须为 UP：

# Check the remote switch ports
for i in {0..7}; do hccn_tool -i $i -lldp -g | grep Ifname; done
# Get the link status of the Ethernet ports (UP or DOWN)
for i in {0..7}; do hccn_tool -i $i -link -g ; done
# Check the network health status
for i in {0..7}; do hccn_tool -i $i -net_health -g ; done
# View the network detected IP configuration
for i in {0..7}; do hccn_tool -i $i -netdetect -g ; done
# View gateway configuration
for i in {0..7}; do hccn_tool -i $i -gateway -g ; done

  1. 检查 NPU HCCN 配置：

确保环境中存在 hccn.conf 文件。如果使用 Docker，请将其挂载到容器中。

cat /etc/hccn.conf

  1. 获取 NPU IP 地址

for i in {0..7}; do hccn_tool -i $i -ip -g;done

  1. 跨节点 PING 测试

# Execute on the target node (replace 'x.x.x.x' with actual npu ip address)
for i in {0..7}; do hccn_tool -i $i -ping -g address x.x.x.x;done

  1. 检查 NPU TLS 配置

# The tls settings should be consistent across all nodes
for i in {0..7}; do hccn_tool -i $i -tls -g ; done | grep switch

使用 Docker 运行#

在每个节点上启动一个 Docker 容器。

# Update the vllm-ascend image
export IMAGE=m.daocloud.io/quay.io/ascend/vllm-ascend:v0.13.0
export NAME=vllm-ascend

# Run the container using the defined variables
# Note: If you are running bridge network with docker, please expose available ports for multiple nodes communication in advance
docker run --rm \
--name $NAME \
--net=host \
--shm-size=1g \
--device /dev/davinci0 \
--device /dev/davinci1 \
--device /dev/davinci2 \
--device /dev/davinci3 \
--device /dev/davinci4 \
--device /dev/davinci5 \
--device /dev/davinci6 \
--device /dev/davinci7 \
--device /dev/davinci8 \
--device /dev/davinci9 \
--device /dev/davinci10 \
--device /dev/davinci11 \
--device /dev/davinci12 \
--device /dev/davinci13 \
--device /dev/davinci14 \
--device /dev/davinci15 \
--device /dev/davinci_manager \
--device /dev/devmm_svm \
--device /dev/hisi_hdc \
-v /usr/local/dcmi:/usr/local/dcmi \
-v /usr/local/Ascend/driver/tools/hccn_tool:/usr/local/Ascend/driver/tools/hccn_tool \
-v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
-v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
-v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
-v /etc/ascend_install.info:/etc/ascend_install.info \
-v /etc/hccn.conf:/etc/hccn.conf \
-v /mnt/sfs_turbo/.cache:/root/.cache \
-it $IMAGE bash

安装 Mooncake#

Mooncake 是 Kimi（由 Moonshot AI 提供的领先 LLM 服务）的服务平台。安装和编译指南：https://github.com/kvcache-ai/Mooncake?tab=readme-ov-file#build-and-use-binaries。首先，我们需要获取 Mooncake 项目。参考以下命令：

git clone -b v0.3.7.post2 --depth 1 https://github.com/kvcache-ai/Mooncake.git

（可选）如果网络状况不佳，请替换 go install 网址

cd Mooncake
sed -i 's|https://go.dev/dl/|https://golang.google.cn/dl/|g' dependencies.sh

安装 mpi

apt-get install mpich libmpich-dev -y

安装相关依赖项。不需要安装 Go。

bash dependencies.sh -y

编译和安装

mkdir build
cd build
cmake .. -DUSE_ASCEND_DIRECT=ON
make -j
make install

设置环境变量

注意：

  • 根据您的具体 Python 安装调整 Python 路径

  • 确保 /usr/local/lib 和 /usr/local/lib64 在您的 LD_LIBRARY_PATH 中

export LD_LIBRARY_PATH=/usr/local/lib64/python3.11/site-packages/mooncake:$LD_LIBRARY_PATH

预填充器/解码器部署#

我们可以分别运行以下脚本在预填充器/解码器节点上启动服务器。请注意，每个 P/D 节点将占用从 kv_port 到 kv_port + num_chips 的端口来初始化套接字监听器。为避免任何问题，应防止端口冲突。此外，确保每个节点的 engine_id 被唯一分配以避免冲突。

launch_online_dp.py#

使用 launch_online_dp.py 启动外部 dp vllm 服务器。launch_online_dp.py

run_dp_template.sh#

在每个节点上修改 run_dp_template.sh。run_dp_template.sh

层分离#

nic_name="eth0"  # network card name
local_ip="192.0.0.1"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export HCCL_BUFFSIZE=256
export TASK_QUEUE_ENABLE=1
export HCCL_OP_EXPANSION_MODE="AIV"
export VLLM_USE_V1=1
export ASCEND_RT_VISIBLE_DEVICES=$1
vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
  --host 0.0.0.0 \
  --port $2 \
  --data-parallel-size $3 \
  --data-parallel-rank $4 \
  --data-parallel-address $5 \
  --data-parallel-rpc-port $6 \
  --tensor-parallel-size $7 \
  --enable-expert-parallel \
  --seed 1024 \
  --served-model-name ds_r1 \
  --max-model-len 40000 \
  --max-num-batched-tokens 16384 \
  --max-num-seqs 8 \
  --enforce-eager \
  --trust-remote-code \
  --gpu-memory-utilization 0.9  \
  --quantization ascend \
  --no-enable-prefix-caching \
  --speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
  --additional-config '{"recompute_scheduler_enable":true,"enable_shared_expert_dp": true}' \
  --kv-transfer-config \
  '{"kv_connector": "MooncakeLayerwiseConnector",
  "kv_role": "kv_producer",
  "kv_port": "30000",
  "engine_id": "0",
  "kv_connector_module_path": "vllm_ascend.distributed.mooncake_layerwise_connector",
  "kv_connector_extra_config": {
            "prefill": {
                    "dp_size": 2,
                    "tp_size": 8
             },
             "decode": {
                    "dp_size": 32,
                    "tp_size": 1
             }
      }
  }'

nic_name="eth0"  # network card name
local_ip="192.0.0.2"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export HCCL_BUFFSIZE=256
export TASK_QUEUE_ENABLE=1
export HCCL_OP_EXPANSION_MODE="AIV"
export VLLM_USE_V1=1
export ASCEND_RT_VISIBLE_DEVICES=$1
vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
  --host 0.0.0.0 \
  --port $2 \
  --data-parallel-size $3 \
  --data-parallel-rank $4 \
  --data-parallel-address $5 \
  --data-parallel-rpc-port $6 \
  --tensor-parallel-size $7 \
  --enable-expert-parallel \
  --seed 1024 \
  --served-model-name ds_r1 \
  --max-model-len 40000 \
  --max-num-batched-tokens 16384 \
  --max-num-seqs 8 \
  --enforce-eager \
  --trust-remote-code \
  --gpu-memory-utilization 0.9  \
  --quantization ascend \
  --no-enable-prefix-caching \
  --speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
  --additional-config '{"recompute_scheduler_enable":true,"enable_shared_expert_dp": true}' \
  --kv-transfer-config \
  '{"kv_connector": "MooncakeLayerwiseConnector",
  "kv_role": "kv_producer",
  "kv_port": "30100",
  "engine_id": "1",
  "kv_connector_module_path": "vllm_ascend.distributed.mooncake_layerwise_connector",
  "kv_connector_extra_config": {
            "prefill": {
                    "dp_size": 2,
                    "tp_size": 8
             },
             "decode": {
                    "dp_size": 32,
                    "tp_size": 1
             }
      }
  }'

nic_name="eth0"  # network card name
local_ip="192.0.0.3"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_MLAPO=1
export HCCL_BUFFSIZE=600
export TASK_QUEUE_ENABLE=1
export HCCL_OP_EXPANSION_MODE="AIV"
export VLLM_USE_V1=1
export ASCEND_RT_VISIBLE_DEVICES=$1
vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
  --host 0.0.0.0 \
  --port $2 \
  --data-parallel-size $3 \
  --data-parallel-rank $4 \
  --data-parallel-address $5 \
  --data-parallel-rpc-port $6 \
  --tensor-parallel-size $7 \
  --enable-expert-parallel \
  --seed 1024 \
  --served-model-name ds_r1 \
  --max-model-len 40000 \
  --max-num-batched-tokens 256 \
  --max-num-seqs 40 \
  --trust-remote-code \
  --gpu-memory-utilization 0.94  \
  --quantization ascend \
  --no-enable-prefix-caching \
  --speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
  --additional-config '{"recompute_scheduler_enable":true,"multistream_overlap_shared_expert": true,"finegrained_tp_config": {"lmhead_tensor_parallel_size":16}}' \
  --compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
  --kv-transfer-config \
  '{"kv_connector": "MooncakeLayerwiseConnector",
  "kv_role": "kv_consumer",
  "kv_port": "30200",
  "engine_id": "2",
  "kv_connector_module_path": "vllm_ascend.distributed.mooncake_layerwise_connector",
  "kv_connector_extra_config": {
            "prefill": {
                    "dp_size": 2,
                    "tp_size": 8
             },
             "decode": {
                    "dp_size": 32,
                    "tp_size": 1
             }
      }
  }'

nic_name="eth0"  # network card name
local_ip="192.0.0.4"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_MLAPO=1
export HCCL_BUFFSIZE=600
export TASK_QUEUE_ENABLE=1
export HCCL_OP_EXPANSION_MODE="AIV"
export VLLM_USE_V1=1
export ASCEND_RT_VISIBLE_DEVICES=$1
vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
  --host 0.0.0.0 \
  --port $2 \
  --data-parallel-size $3 \
  --data-parallel-rank $4 \
  --data-parallel-address $5 \
  --data-parallel-rpc-port $6 \
  --tensor-parallel-size $7 \
  --enable-expert-parallel \
  --seed 1024 \
  --served-model-name ds_r1 \
  --max-model-len 40000 \
  --max-num-batched-tokens 256 \
  --max-num-seqs 40 \
  --trust-remote-code \
  --gpu-memory-utilization 0.94  \
  --quantization ascend \
  --no-enable-prefix-caching \
  --speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
  --additional-config '{"recompute_scheduler_enable":true,"multistream_overlap_shared_expert": true,"finegrained_tp_config": {"lmhead_tensor_parallel_size":16}}' \
  --compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
  --kv-transfer-config \
  '{"kv_connector": "MooncakeLayerwiseConnector",
  "kv_role": "kv_consumer",
  "kv_port": "30200",
  "engine_id": "2",
  "kv_connector_module_path": "vllm_ascend.distributed.mooncake_layerwise_connector",
  "kv_connector_extra_config": {
            
            "prefill": {
                    "dp_size": 2,
                    "tp_size": 8
             },
             "decode": {
                    "dp_size": 32,
                    "tp_size": 1
             }
      }
  }'

非层分离#

nic_name="eth0"  # network card name
local_ip="192.0.0.1"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export HCCL_BUFFSIZE=256
export TASK_QUEUE_ENABLE=1
export HCCL_OP_EXPANSION_MODE="AIV"
export VLLM_USE_V1=1
export ASCEND_RT_VISIBLE_DEVICES=$1
vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
  --host 0.0.0.0 \
  --port $2 \
  --data-parallel-size $3 \
  --data-parallel-rank $4 \
  --data-parallel-address $5 \
  --data-parallel-rpc-port $6 \
  --tensor-parallel-size $7 \
  --enable-expert-parallel \
  --seed 1024 \
  --served-model-name ds_r1 \
  --max-model-len 40000 \
  --max-num-batched-tokens 16384 \
  --max-num-seqs 8 \
  --enforce-eager \
  --trust-remote-code \
  --gpu-memory-utilization 0.9  \
  --quantization ascend \
  --no-enable-prefix-caching \
  --speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
  --additional-config '{"recompute_scheduler_enable":true,"enable_shared_expert_dp": true}' \
  --kv-transfer-config \
  '{"kv_connector": "MooncakeConnectorV1",
  "kv_role": "kv_producer",
  "kv_port": "30000",
  "engine_id": "0",
  "kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector",
  "kv_connector_extra_config": {
            "prefill": {
                    "dp_size": 2,
                    "tp_size": 8
             },
             "decode": {
                    "dp_size": 32,
                    "tp_size": 1
             }
      }
  }'

nic_name="eth0"  # network card name
local_ip="192.0.0.2"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export HCCL_BUFFSIZE=256
export TASK_QUEUE_ENABLE=1
export HCCL_OP_EXPANSION_MODE="AIV"
export VLLM_USE_V1=1
export ASCEND_RT_VISIBLE_DEVICES=$1
vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
  --host 0.0.0.0 \
  --port $2 \
  --data-parallel-size $3 \
  --data-parallel-rank $4 \
  --data-parallel-address $5 \
  --data-parallel-rpc-port $6 \
  --tensor-parallel-size $7 \
  --enable-expert-parallel \
  --seed 1024 \
  --served-model-name ds_r1 \
  --max-model-len 40000 \
  --max-num-batched-tokens 16384 \
  --max-num-seqs 8 \
  --enforce-eager \
  --trust-remote-code \
  --gpu-memory-utilization 0.9  \
  --quantization ascend \
  --no-enable-prefix-caching \
  --speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
  --additional-config '{"recompute_scheduler_enable":true,"enable_shared_expert_dp": true}' \
  --kv-transfer-config \
  '{"kv_connector": "MooncakeConnectorV1",
  "kv_role": "kv_producer",
  "kv_port": "30100",
  "engine_id": "1",
  "kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector",
  "kv_connector_extra_config": {
            "prefill": {
                    "dp_size": 2,
                    "tp_size": 8
             },
             "decode": {
                    "dp_size": 32,
                    "tp_size": 1
             }
      }
  }'

nic_name="eth0"  # network card name
local_ip="192.0.0.3"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_MLAPO=1
export HCCL_BUFFSIZE=600
export TASK_QUEUE_ENABLE=1
export HCCL_OP_EXPANSION_MODE="AIV"
export VLLM_USE_V1=1
export ASCEND_RT_VISIBLE_DEVICES=$1
vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
  --host 0.0.0.0 \
  --port $2 \
  --data-parallel-size $3 \
  --data-parallel-rank $4 \
  --data-parallel-address $5 \
  --data-parallel-rpc-port $6 \
  --tensor-parallel-size $7 \
  --enable-expert-parallel \
  --seed 1024 \
  --served-model-name ds_r1 \
  --max-model-len 40000 \
  --max-num-batched-tokens 256 \
  --max-num-seqs 40 \
  --trust-remote-code \
  --gpu-memory-utilization 0.94  \
  --quantization ascend \
  --no-enable-prefix-caching \
  --speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
  --additional-config '{"recompute_scheduler_enable":true,"multistream_overlap_shared_expert": true,"finegrained_tp_config": {"lmhead_tensor_parallel_size":16}}' \
  --compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
  --kv-transfer-config \
  '{"kv_connector": "MooncakeConnectorV1",
  "kv_role": "kv_consumer",
  "kv_port": "30200",
  "engine_id": "2",
  "kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector",
  "kv_connector_extra_config": {
            "prefill": {
                    "dp_size": 2,
                    "tp_size": 8
             },
             "decode": {
                    "dp_size": 32,
                    "tp_size": 1
             }
      }
  }'

nic_name="eth0"  # network card name
local_ip="192.0.0.4"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_MLAPO=1
export HCCL_BUFFSIZE=600
export TASK_QUEUE_ENABLE=1
export HCCL_OP_EXPANSION_MODE="AIV"
export VLLM_USE_V1=1
export ASCEND_RT_VISIBLE_DEVICES=$1
vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
  --host 0.0.0.0 \
  --port $2 \
  --data-parallel-size $3 \
  --data-parallel-rank $4 \
  --data-parallel-address $5 \
  --data-parallel-rpc-port $6 \
  --tensor-parallel-size $7 \
  --enable-expert-parallel \
  --seed 1024 \
  --served-model-name ds_r1 \
  --max-model-len 40000 \
  --max-num-batched-tokens 256 \
  --max-num-seqs 40 \
  --trust-remote-code \
  --gpu-memory-utilization 0.94  \
  --quantization ascend \
  --no-enable-prefix-caching \
  --speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
  --additional-config '{"recompute_scheduler_enable":true,"multistream_overlap_shared_expert": true,"finegrained_tp_config": {"lmhead_tensor_parallel_size":16}}' \
  --compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
  --kv-transfer-config \
  '{"kv_connector": "MooncakeConnectorV1",
  "kv_role": "kv_consumer",
  "kv_port": "30200",
  "engine_id": "2",
  "kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector",
  "kv_connector_extra_config": {
            "prefill": {
                    "dp_size": 2,
                    "tp_size": 8
             },
             "decode": {
                    "dp_size": 32,
                    "tp_size": 1
             }
      }
  }'

启动服务#

# on 190.0.0.1
python launch_online_dp.py --dp-size 2 --tp-size 8 --dp-size-local 2 --dp-rank-start 0 --dp-address 190.0.0.1 --dp-rpc-port 12321 --vllm-start-port 7100
# on 190.0.0.2
python launch_online_dp.py --dp-size 2 --tp-size 8 --dp-size-local 2 --dp-rank-start 0 --dp-address 190.0.0.2 --dp-rpc-port 12321 --vllm-start-port 7100
# on 190.0.0.3
python launch_online_dp.py --dp-size 32 --tp-size 1 --dp-size-local 16 --dp-rank-start 0 --dp-address 190.0.0.3 --dp-rpc-port 12321 --vllm-start-port 7100
# on 190.0.0.4
python launch_online_dp.py --dp-size 32 --tp-size 1 --dp-size-local 16 --dp-rank-start 16 --dp-address 190.0.0.3 --dp-rpc-port 12321 --vllm-start-port 7100

部署示例代理#

在部署预填充服务实例的同一节点上运行代理服务器。您可以在仓库的示例目录中找到代理实现。

我们提供两种具有不同请求路由行为的代理实现：

  • load_balance_proxy_layerwise_server_example.py：请求首先路由到 D 节点，然后根据需要转发到 P 节点。此代理设计用于与 MooncakeLayerwiseConnector 配合使用。load_balance_proxy_layerwise_server_example.py

  • load_balance_proxy_server_example.py：请求首先路由到 P 节点，然后转发到 D 节点进行后续处理。此代理设计用于与 MooncakeConnector 配合使用。load_balance_proxy_server_example.py

python load_balance_proxy_layerwise_server_example.py \
  --port 1999 \
  --host 192.0.0.1 \
  --prefiller-hosts \
    192.0.0.1 \
    192.0.0.1 \
    192.0.0.2 \
    192.0.0.2 \
  --prefiller-ports  \
    7100 7101 7100 7101 \
  --decoder-hosts \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
  --decoder-ports  \
    7100 7101 7102 7103 7104 7105 7106 7107 7108 7109 7110 7111 7112 7113 7114 7115\
    7100 7101 7102 7103 7104 7105 7106 7107 7108 7109 7110 7111 7112 7113 7114 7115\

python load_balance_proxy_server_example.py \
  --port 1999 \
  --host 192.0.0.1 \
  --prefiller-hosts \
    192.0.0.1 \
    192.0.0.1 \
    192.0.0.2 \
    192.0.0.2 \
  --prefiller-ports  \
    7100 7101 7100 7101 \
  --decoder-hosts \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.3  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
    192.0.0.4  \
  --decoder-ports  \
    7100 7101 7102 7103 7104 7105 7106 7107 7108 7109 7110 7111 7112 7113 7114 7115\
    7100 7101 7102 7103 7104 7105 7106 7107 7108 7109 7110 7111 7112 7113 7114 7115\

参数

含义

--port

代理服务端口

--host

代理服务主机 IP

--prefiller-hosts

预填充器节点主机

--prefiller-ports

预填充器节点端口

--decoder-hosts

解码器节点主机

--decoder-ports

解码器节点端口

您可以在仓库的示例中获取代理程序，load_balance_proxy_server_example.py

基准测试#

我们推荐使用 aisbench 工具评估性能。aisbench 执行以下命令安装 aisbench

git clone https://gitee.com/aisbench/benchmark.git
cd benchmark/
pip3 install -e ./

在评估性能之前，您需要取消 http 代理，如下所示

# unset proxy
unset http_proxy
unset https_proxy

  • 您可以将数据集放置在目录：benchmark/ais_bench/datasets

  • 您可以在目录 benchmark/ais_bench/benchmark/configs/models/vllm_api 中更改配置。以 vllm_api_stream_chat.py 为例

models = [
    dict(
        attr="service",
        type=VLLMCustomAPIChatStream,
        abbr='vllm-api-stream-chat',
        path="/root/.cache/ds_r1",
        model="dsr1",
        request_rate = 14,
        retry = 2,
        host_ip = "192.0.0.1", # Proxy service host IP
        host_port = 8000,  # Proxy service Port
        max_out_len = 10,
        batch_size=768,
        trust_remote_code=True,
        generation_kwargs = dict(
            temperature = 0,
            seed = 1024,
            ignore_eos=False,
        )
    )
]

  • 以 gsm8k 数据集为例，执行以下命令评估性能。

ais_bench --models vllm_api_stream_chat --datasets gsm8k_gen_0_shot_cot_str_perf  --debug  --mode perf

  • 有关 aisbench 命令和参数的更多详细信息，请参考 aisbench

常见问题#

1.预填充器节点需要预热#

由于一些 NPU 算子的计算需要多轮预热才能达到最佳性能，我们建议在进行性能测试之前先用一些请求预热服务，以实现最佳端到端吞吐量。

验证#

使用代理服务器端点检查服务健康状况。

curl http://192.0.0.1:8080/v1/completions \
    -H "Content-Type: application/json" \
    -d '{
        "model": "qwen3-moe",
        "prompt": "Who are you?",
        "max_tokens": 100,
        "temperature": 0
    }'