DeepSeek-V3.2

DeepSeek-V3.2#

简介#

DeepSeek-V3.2 是一个稀疏注意力模型。其主要架构与 DeepSeek-V3.1 相似，但采用了稀疏注意力机制，旨在探索和验证长上下文场景下训练和推理效率的优化方案。

本文档将展示模型的主要验证步骤，包括支持的特性、特性配置、环境准备、单节点和多节点部署、精度评估和性能评估。

支持的特性#

请参阅支持的特性矩阵获取模型支持的特性列表。

请参阅特性指南获取特性的配置方法。

环境准备#

模型权重#

DeepSeek-V3.2-Exp-W8A8（量化版本）：需要 1 个 Atlas 800 A3（64G × 16）节点或 2 个 Atlas 800 A2（64G × 8）节点。下载模型权重
DeepSeek-V3.2-w8a8（量化版本）：需要 1 个 Atlas 800 A3（64G × 16）节点或 2 个 Atlas 800 A2（64G × 8）节点。下载模型权重

建议将模型权重下载到多节点共享目录，例如 /root/.cache/。

验证多节点通信（可选）#

如果要部署多节点环境，需要按照验证多节点通信环境验证多节点通信。

安装#

您可以使用官方 Docker 镜像直接运行 DeepSeek-V3.2。

A3 系列

在每个节点上启动 Docker 镜像。

export IMAGE=quay.io/ascend/vllm-ascend:v0.20.2rc1-a3
docker run --rm \
    --name vllm-ascend \
    --shm-size=1g \
    --net=host \
    --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 /root/.cache:/root/.cache \
    -it $IMAGE bash

A2 系列

在每个节点上启动 Docker 镜像。

export IMAGE=quay.io/ascend/vllm-ascend:v0.20.2rc1
docker run --rm \
    --name vllm-ascend \
    --shm-size=1g \
    --net=host \
    --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/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 /root/.cache:/root/.cache \
    -it $IMAGE bash

此外，如果您不想使用上述 Docker 镜像，也可以从源码构建：

从源码安装 vllm-ascend，请参考安装指南。

如果要部署多节点环境，需要在每个节点上设置环境。

部署#

备注

在本教程中，假设您已将模型权重下载到 /root/.cache/。请根据实际情况修改为您自己的路径。

单节点部署#

量化模型 DeepSeek-V3.2-w8a8 可以部署在 1 台 Atlas 800 A3（64G × 16）上。

运行以下脚本执行在线推理。

export HCCL_OP_EXPANSION_MODE="AIV"
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export VLLM_USE_V1=1
export HCCL_BUFFSIZE=200
export VLLM_ASCEND_ENABLE_MLAPO=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/DeepSeek-V3.2-W8A8 \
--host 0.0.0.0 \
--port 8000 \
--data-parallel-size 2 \
--tensor-parallel-size 8 \
--quantization ascend \
--seed 1024 \
--served-model-name deepseek_v3_2 \
--enable-expert-parallel \
--max-num-seqs 16 \
--max-model-len 8192 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--no-enable-prefix-caching \
--gpu-memory-utilization 0.92 \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

多节点部署#

DeepSeek-V3.2-w8a8：至少需要 2 台 Atlas 800 A2（64G × 8）。

分别在两个节点上运行以下脚本。

A3 系列

节点 0

# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxx"
local_ip="xxx"

# The value of node0_ip must be consistent with the value of local_ip set in node0 (master node)
node0_ip="xxxx"

export HCCL_OP_EXPANSION_MODE="AIV"

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 VLLM_USE_V1=1
export HCCL_BUFFSIZE=200
export VLLM_ASCEND_ENABLE_MLAPO=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/DeepSeek-V3.2-W8A8 \
--host 0.0.0.0 \
--port 8077 \
--data-parallel-size 2 \
--data-parallel-size-local 1 \
--data-parallel-address $node0_ip \
--data-parallel-rpc-port 12890 \
--tensor-parallel-size 16 \
--quantization ascend \
--seed 1024 \
--served-model-name deepseek_v3_2 \
--enable-expert-parallel \
--max-num-seqs 16 \
--max-model-len 8192 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--no-enable-prefix-caching \
--gpu-memory-utilization 0.92 \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

节点 1

# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxx"
local_ip="xxx"

# The value of node0_ip must be consistent with the value of local_ip set in node0 (master node)
node0_ip="xxxx"

export HCCL_OP_EXPANSION_MODE="AIV"

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 VLLM_USE_V1=1
export HCCL_BUFFSIZE=200
export VLLM_ASCEND_ENABLE_MLAPO=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/DeepSeek-V3.2-W8A8 \
--host 0.0.0.0 \
--port 8077 \
--headless \
--data-parallel-size 2 \
--data-parallel-size-local 1 \
--data-parallel-start-rank 1 \
--data-parallel-address $node0_ip \
--data-parallel-rpc-port 12890 \
--tensor-parallel-size 16 \
--quantization ascend \
--seed 1024 \
--served-model-name deepseek_v3_2 \
--enable-expert-parallel \
--max-num-seqs 16 \
--max-model-len 8192 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--no-enable-prefix-caching \
--gpu-memory-utilization 0.92 \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

A2 系列

节点 0

# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxx"
local_ip="xxx"

# The value of node0_ip must be consistent with the value of local_ip set in node0 (master node)
node0_ip="xxxx"

export HCCL_OP_EXPANSION_MODE="AIV"

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=100
export VLLM_USE_V1=1
export HCCL_BUFFSIZE=200
export VLLM_ASCEND_ENABLE_MLAPO=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1
export HCCL_CONNECT_TIMEOUT=120
export HCCL_INTRA_PCIE_ENABLE=1
export HCCL_INTRA_ROCE_ENABLE=0

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/DeepSeek-V3.2-W8A8 \
--host 0.0.0.0 \
--port 8077 \
--data-parallel-size 2 \
--data-parallel-size-local 1 \
--data-parallel-address $node0_ip \
--data-parallel-rpc-port 13389 \
--tensor-parallel-size 8 \
--quantization ascend \
--seed 1024 \
--served-model-name deepseek_v3_2 \
--enable-expert-parallel \
--max-num-seqs 16 \
--max-model-len 8192 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--no-enable-prefix-caching \
--gpu-memory-utilization 0.92 \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY", "cudagraph_capture_sizes":[8, 16, 24, 32, 40, 48]}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

节点 1

# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxx"
local_ip="xxx"

# The value of node0_ip must be consistent with the value of local_ip set in node0 (master node)
node0_ip="xxxx"

export HCCL_OP_EXPANSION_MODE="AIV"

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=100
export VLLM_USE_V1=1
export HCCL_BUFFSIZE=200
export VLLM_ASCEND_ENABLE_MLAPO=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1
export HCCL_CONNECT_TIMEOUT=120
export HCCL_INTRA_PCIE_ENABLE=1
export HCCL_INTRA_ROCE_ENABLE=0

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/DeepSeek-V3.2-W8A8 \
--host 0.0.0.0 \
--port 8077 \
--headless \
--data-parallel-size 2 \
--data-parallel-size-local 1 \
--data-parallel-start-rank 1 \
--data-parallel-address $node0_ip \
--data-parallel-rpc-port 13389 \
--tensor-parallel-size 8 \
--quantization ascend \
--seed 1024 \
--served-model-name deepseek_v3_2 \
--enable-expert-parallel \
--max-num-seqs 16 \
--max-model-len 8192 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--no-enable-prefix-caching \
--gpu-memory-utilization 0.92 \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY", "cudagraph_capture_sizes":[8, 16, 24, 32, 40, 48]}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

预填充-解码分离部署#

我们将在多节点环境下以 1P1D 方式展示 DeepSeek-V3.2 的部署指南，以获得更好的性能。

开始之前，请

在每个节点上准备脚本 launch_online_dp.py：

import argparse
import multiprocessing
import os
import subprocess
import sys

def parse_args():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--dp-size",
        type=int,
        required=True,
        help="Data parallel size."
    )
    parser.add_argument(
        "--tp-size",
        type=int,
        default=1,
        help="Tensor parallel size."
    )
    parser.add_argument(
        "--dp-size-local",
        type=int,
        default=-1,
        help="Local data parallel size."
    )
    parser.add_argument(
        "--dp-rank-start",
        type=int,
        default=0,
        help="Starting rank for data parallel."
    )
    parser.add_argument(
        "--dp-address",
        type=str,
        required=True,
        help="IP address for data parallel master node."
    )
    parser.add_argument(
        "--dp-rpc-port",
        type=str,
        default=12345,
        help="Port for data parallel master node."
    )
    parser.add_argument(
        "--vllm-start-port",
        type=int,
        default=9000,
        help="Starting port for the engine."
    )
    return parser.parse_args()

args = parse_args()
dp_size = args.dp_size
tp_size = args.tp_size
dp_size_local = args.dp_size_local
if dp_size_local == -1:
    dp_size_local = dp_size
dp_rank_start = args.dp_rank_start
dp_address = args.dp_address
dp_rpc_port = args.dp_rpc_port
vllm_start_port = args.vllm_start_port

def run_command(visible_devices, dp_rank, vllm_engine_port):
    command = [
        "bash",
        "./run_dp_template.sh",
        visible_devices,
        str(vllm_engine_port),
        str(dp_size),
        str(dp_rank),
        dp_address,
        dp_rpc_port,
        str(tp_size),
    ]
    subprocess.run(command, check=True)

if __name__ == "__main__":
    template_path = "./run_dp_template.sh"
    if not os.path.exists(template_path):
        print(f"Template file {template_path} does not exist.")
        sys.exit(1)

    processes = []
    num_cards = dp_size_local * tp_size
    for i in range(dp_size_local):
        dp_rank = dp_rank_start + i
        vllm_engine_port = vllm_start_port + i
        visible_devices = ",".join(str(x) for x in range(i * tp_size, (i + 1) * tp_size))
        process = multiprocessing.Process(target=run_command,
                                        args=(visible_devices, dp_rank,
                                                vllm_engine_port))
        processes.append(process)
        process.start()

    for process in processes:
        process.join()

在每个节点上准备脚本 run_dp_template.sh。

预填充节点 0

nic_name="enp48s3u1u1" # change to your own nic name
local_ip=141.61.39.105 # change to your own ip

export HCCL_OP_EXPANSION_MODE="AIV"

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_USE_V1=1
export HCCL_BUFFSIZE=256

export ASCEND_AGGREGATE_ENABLE=1
export ASCEND_TRANSPORT_PRINT=1
export ACL_OP_INIT_MODE=1
export ASCEND_A3_ENABLE=1
# Timeout (in seconds) for automatically releasing the prefiller’s KV cache for a particular request.
export VLLM_MOONCAKE_ABORT_REQUEST_TIMEOUT=480

export ASCEND_RT_VISIBLE_DEVICES=$1

export VLLM_ASCEND_ENABLE_FLASHCOMM1=1


vllm serve /root/.cache/Eco-Tech/DeepSeek-V3.2-w8a8-mtp-QuaRot \
    --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 \
    --speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name deepseek_v3.2 \
    --max-model-len 68000 \
    --max-num-batched-tokens 32560 \
    --trust-remote-code \
    --max-num-seqs 64 \
    --gpu-memory-utilization 0.82 \
    --quantization ascend \
    --enforce-eager \
    --no-enable-prefix-caching \
    --additional-config '{"layer_sharding": ["q_b_proj", "o_proj"], "enable_dsa_cp": true}' \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeLayerwiseConnector",
    "kv_role": "kv_producer",
    "kv_port": "30000",
    "engine_id": "0",
    "kv_connector_extra_config": {
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 8,
                        "tp_size": 4
                }
        }
    }'

预填充节点 1

nic_name="enp48s3u1u1" # change to your own nic name
local_ip=141.61.39.113 # change to your own ip

export HCCL_OP_EXPANSION_MODE="AIV"

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_USE_V1=1
export HCCL_BUFFSIZE=256

export ASCEND_AGGREGATE_ENABLE=1
export ASCEND_TRANSPORT_PRINT=1
export ACL_OP_INIT_MODE=1
export ASCEND_A3_ENABLE=1
# Timeout (in seconds) for automatically releasing the prefiller’s KV cache for a particular request.
export VLLM_MOONCAKE_ABORT_REQUEST_TIMEOUT=480

export ASCEND_RT_VISIBLE_DEVICES=$1

export VLLM_ASCEND_ENABLE_FLASHCOMM1=1


vllm serve /root/.cache/Eco-Tech/DeepSeek-V3.2-w8a8-mtp-QuaRot \
    --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 \
    --speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name deepseek_v3.2 \
    --max-model-len 68000 \
    --max-num-batched-tokens 32560 \
    --trust-remote-code \
    --max-num-seqs 64 \
    --gpu-memory-utilization 0.82 \
    --quantization ascend \
    --enforce-eager \
    --no-enable-prefix-caching \
    --additional-config '{"layer_sharding": ["q_b_proj", "o_proj"], "enable_dsa_cp": true}' \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeLayerwiseConnector",
    "kv_role": "kv_producer",
    "kv_port": "30000",
    "engine_id": "0",
    "kv_connector_extra_config": {
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 8,
                        "tp_size": 4
                }
        }
    }'

解码节点 0

nic_name="enp48s3u1u1" # change to your own nic name
local_ip=141.61.39.117 # change to your own ip

export HCCL_OP_EXPANSION_MODE="AIV"

export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name

#Mooncake
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10

export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_USE_V1=1
export HCCL_BUFFSIZE=256


export ASCEND_AGGREGATE_ENABLE=1
export ASCEND_TRANSPORT_PRINT=1
export ACL_OP_INIT_MODE=1
export ASCEND_A3_ENABLE=1
# Timeout (in seconds) for automatically releasing the prefiller’s KV cache for a particular request.
export VLLM_MOONCAKE_ABORT_REQUEST_TIMEOUT=480

export TASK_QUEUE_ENABLE=1

export ASCEND_RT_VISIBLE_DEVICES=$1


vllm serve /root/.cache/Eco-Tech/DeepSeek-V3.2-w8a8-mtp-QuaRot \
    --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 \
    --speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name deepseek_v3.2 \
    --max-model-len 68000 \
    --max-num-batched-tokens 12 \
    --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY", "cudagraph_capture_sizes":[3, 6, 9, 12]}' \
    --trust-remote-code \
    --max-num-seqs 4 \
    --gpu-memory-utilization 0.95 \
    --no-enable-prefix-caching \
    --quantization ascend \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeLayerwiseConnector",
    "kv_role": "kv_consumer",
    "kv_port": "30100",
    "engine_id": "1",
    "kv_connector_extra_config": {
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 8,
                        "tp_size": 4
                }
        }
    }' \
    --additional-config '{"recompute_scheduler_enable" : true}'

解码节点 1

nic_name="enp48s3u1u1" # change to your own nic name
local_ip=141.61.39.181 # change to your own ip

export HCCL_OP_EXPANSION_MODE="AIV"

export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name

#Mooncake
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10

export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_USE_V1=1
export HCCL_BUFFSIZE=256

export ASCEND_AGGREGATE_ENABLE=1
export ASCEND_TRANSPORT_PRINT=1
export ACL_OP_INIT_MODE=1
export ASCEND_A3_ENABLE=1
# Timeout (in seconds) for automatically releasing the prefiller’s KV cache for a particular request.
export VLLM_MOONCAKE_ABORT_REQUEST_TIMEOUT=480

export TASK_QUEUE_ENABLE=1

export ASCEND_RT_VISIBLE_DEVICES=$1


vllm serve /root/.cache/Eco-Tech/DeepSeek-V3.2-w8a8-mtp-QuaRot \
    --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 \
    --speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name deepseek_v3.2 \
    --max-model-len 68000 \
    --max-num-batched-tokens 12 \
    --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY",  "cudagraph_capture_sizes":[3, 6, 9, 12]}' \
    --trust-remote-code \
    --max-num-seqs 4 \
    --gpu-memory-utilization 0.95 \
    --no-enable-prefix-caching \
    --quantization ascend \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeLayerwiseConnector",
    "kv_role": "kv_consumer",
    "kv_port": "30100",
    "engine_id": "1",
    "kv_connector_extra_config": {
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 8,
                        "tp_size": 4
                }
        }
    }' \
    --additional-config '{"recompute_scheduler_enable" : true}'

准备工作完成后，您可以在每个节点上使用以下命令启动服务器：详细启动方法请参考大规模专家并行分布式 DP 服务器。

预填充节点 0

# change ip to your own
python launch_online_dp.py --dp-size 2 --tp-size 16 --dp-size-local 1 --dp-rank-start 0 --dp-address 141.61.39.105 --dp-rpc-port 12890 --vllm-start-port 9100

预填充节点 1

# change ip to your own
python launch_online_dp.py --dp-size 2 --tp-size 16 --dp-size-local 1 --dp-rank-start 1 --dp-address 141.61.39.105 --dp-rpc-port 12890 --vllm-start-port 9100

解码节点 0

# change ip to your own
python launch_online_dp.py --dp-size 8 --tp-size 4 --dp-size-local 4 --dp-rank-start 0 --dp-address 141.61.39.117 --dp-rpc-port 12777 --vllm-start-port 9100

解码节点 1

# change ip to your own
python launch_online_dp.py --dp-size 8 --tp-size 4 --dp-size-local 4 --dp-rank-start 4 --dp-address 141.61.39.117 --dp-rpc-port 12777 --vllm-start-port 9100

请求转发#

要设置请求转发，请在任意机器上运行以下脚本。您可以在仓库的示例中获取代理程序：load_balance_proxy_layerwise_server_example.py

unset http_proxy
unset https_proxy

python load_balance_proxy_layerwise_server_example.py \
    --port 8000 \
    --host 141.61.39.105 \
    --prefiller-hosts \
       141.61.39.105 \
       141.61.39.113 \
    --prefiller-ports \
       9100 \
       9100 \
    --decoder-hosts \
      141.61.39.117 \
      141.61.39.117 \
      141.61.39.117 \
      141.61.39.117 \
      141.61.39.181 \
      141.61.39.181 \
      141.61.39.181 \
      141.61.39.181 \
    --decoder-ports \
      9100 9101 9102 9103 \
      9100 9101 9102 9103 \

功能验证#

服务器启动后，您可以使用输入提示词查询模型：

Note:

<node0_ip>: The IP address of the node where the server is running (e.g., localhost).
<port>: The port number specified in the server startup command (e.g., 8000).

curl http://<node0_ip>:<port>/v1/completions \
    -H "Content-Type: application/json" \
    -d '{
        "model": "deepseek_v3.2",
        "prompt": "The future of AI is",
        "max_completion_tokens": 50,
        "temperature": 0
    }'

精度评估#

以下是两种精度评估方法。

使用 AISBench#

详细信息请参考使用 AISBench 进行精度评估。
执行后即可获得结果。

使用 Language Model Evaluation Harness#

以 gsm8k 数据集为例，以在线模式运行 DeepSeek-V3.2-W8A8 的精度评估。

请参考使用 lm_eval 进行精度评估了解 lm_eval 的安装方法。

运行 lm_eval 执行精度评估。

lm_eval \
--model local-completions \
--model_args model=/root/.cache/Eco-Tech/DeepSeek-V3.2-w8a8-mtp-QuaRot,base_url=http://127.0.0.1:8000/v1/completions,tokenized_requests=False,trust_remote_code=True \
--tasks gsm8k \
--output_path ./

执行后即可获得结果。

性能#

使用 AISBench#

详细信息请参考使用 AISBench 进行性能评估。

性能结果如下：

硬件：A3-752T，4 节点

部署：1P1D，预填充节点：DP2+TP16，解码节点：DP8+TP4

输入/输出：64k/3k

性能：533 tps，TPOT 32 毫秒

使用 vLLM 基准测试#

以 DeepSeek-V3.2-W8A8 为例运行性能评估。

更多详细信息请参考 vLLM 基准测试。

有三个 vllm bench 子命令：

latency：基准测试单批次请求的延迟。
serve：基准测试在线服务的吞吐量。
throughput：基准测试离线推理的吞吐量。

以 serve 为例，运行如下代码。

export VLLM_USE_MODELSCOPE=True
vllm bench serve --model /root/.cache/Eco-Tech/DeepSeek-V3.2-w8a8-mtp-QuaRot  --dataset-name random --random-input 200 --num-prompts 200 --request-rate 1 --save-result --result-dir ./

函数调用#

函数调用功能从 v0.13.0rc1 版本开始支持。请使用最新版本。

详细信息请参考 DeepSeek-V3.2 使用指南。