跳转至

GLM-5/GLM-5.1

1.简介

本文档同时适用于 GLM-5 和 GLM-5.1。除非另有说明，本文档中所有针对 GLM-5 的描述、配置和部署流程同样适用于 GLM-5.1。为简洁起见，下文统一使用 GLM-5 指代 GLM-5 和 GLM-5.1。

GLM-5 采用混合专家（MoE）架构，面向复杂的系统工程和长周期智能体任务。

GLM-5 模型首次在 vllm-ascend:v0.17.0rc1 中得到支持，所有 v0.17.0rc1 及更高版本 均可稳定运行。要使用最新特性（如 PD 分离、MTP），建议使用最新候选版本或正式版本。transformers 版本需要升级到 5.2.0 或更高版本。

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。

2.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。

3.前提条件

3.1.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/

3.2.验证多节点通信（可选）

如果需要多节点部署，请按照验证多节点通信环境指南进行通信验证。

4.安装

4.1.Docker 镜像安装

您可以使用官方 Docker 镜像直接运行 GLM-5/5.1。

在每个节点上启动 Docker 镜像。

export IMAGE=quay.io/ascend/vllm-ascend:v0.22.1rc1-a3
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 /root/.cache:/root/.cache \
-it $IMAGE bash

Start the docker image on your each node.

export IMAGE=quay.io/ascend/vllm-ascend:v0.22.1rc1
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

If you want to deploy multi-node environment, you need to set up environment on each node.

To verify the successful installation of the environment, please refer to installation.

4.2 Source Code Installation

In addition, if you don't want to use the docker image as above, you can also build all from source:

If you want to deploy multi-node environment, you need to set up environment on each node.

5 Online Service Deployment

5.1 Single-Node Online Deployment

  • Quantized model glm-5-w4a8 and glm-5.1-w4a8 can be deployed on 1 Atlas 800 A3 (64G × 16) .

Run the following script to execute online inference.

Common Issues Tip: If you encounter issues, Refer to FAQs.

# The version of transformers needs to be upgraded to 5.2.0.
# pip install transformers==5.2.0 --upgrade

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

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w4a8 \
--host 0.0.0.0 \
--port 8077 \
--data-parallel-size 1 \
--tensor-parallel-size 16 \
--enable-expert-parallel \
--seed 1024 \
--served-model-name glm-5 \
--max-num-seqs 8 \
--max-model-len 200000 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.95 \
--quantization ascend \
--enable-chunked-prefill \
--enable-prefix-caching \
--additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true}' \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'
  • Quantized model glm-5-w8a8 and glm-5.1-w8a8 can be deployed on 1 Atlas 800 A3 (64G × 16) .

Run the following script to execute online inference.

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

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w8a8 \
--host 0.0.0.0 \
--port 8077 \
--data-parallel-size 1 \
--tensor-parallel-size 16 \
--enable-expert-parallel \
--seed 1024 \
--served-model-name glm-5 \
--max-num-seqs 8 \
--max-model-len 40960 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.95 \
--quantization ascend \
--enable-chunked-prefill \
--enable-prefix-caching \
--additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true}' \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'
  • Quantized model glm-5-w4a8 can be deployed on 1 Atlas 800 A2 (64G × 8) .

Run the following script to execute online inference.

Common Issues Tip: If you encounter issues, Refer to FAQs.

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

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w4a8 \
--host 0.0.0.0 \
--port 8077 \
--data-parallel-size 1 \
--tensor-parallel-size 8 \
--enable-expert-parallel \
--seed 1024 \
--served-model-name glm-5 \
--max-num-seqs 2 \
--max-model-len 32768 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.95 \
--quantization ascend \
--enable-chunked-prefill \
--enable-prefix-caching \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

Notice: The parameters are explained as follows:

  • For single-node deployment, we recommend using dp1tp16 and turn off expert parallel in low-latency scenarios.

5.2 Multi-node Deployment

If you want to deploy multi-node environment, you need to verify multi-node communication according to verify multi-node communication environment.

Common Issues Tip: If you encounter issues, Refer to FAQs.

  • glm-5-bf16 and glm-5.1-bf16: require at least 2 Atlas 800 A3 (64G × 16).

Run the following scripts on two nodes respectively.

node 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=1
export HCCL_BUFFSIZE=200
export VLLM_ASCEND_BALANCE_SCHEDULING=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-bf16 \
--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 \
--seed 1024 \
--served-model-name glm-5 \
--enable-expert-parallel \
--max-num-seqs 16 \
--max-model-len 8192 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.95 \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

node 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=1
export HCCL_BUFFSIZE=200
export VLLM_ASCEND_BALANCE_SCHEDULING=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-bf16 \
--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 \
--seed 1024 \
--served-model-name glm-5 \
--enable-expert-parallel \
--max-num-seqs 16 \
--max-model-len 8192 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.95 \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

Run the following scripts on two nodes respectively.

node 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="xxx"

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=1
export HCCL_BUFFSIZE=200
export VLLM_ASCEND_BALANCE_SCHEDULING=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w4a8 \
--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 glm-5 \
--enable-expert-parallel \
--max-num-seqs 2 \
--max-model-len 131072 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.95 \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

node 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="xxx"

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=1
export HCCL_BUFFSIZE=200
export VLLM_ASCEND_BALANCE_SCHEDULING=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w4a8 \
--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 glm-5 \
--enable-expert-parallel \
--max-num-seqs 2 \
--max-model-len 131072 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.95 \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

python adjust_weight.py "path_of_bf16_weight"

# adjust_weight.py
from safetensors.torch import safe_open, save_file
import torch
import json
import os
import sys

target_keys = ["model.embed_tokens.weight", "lm_head.weight"]

def get_tensor_info(file_path):
   with safe_open(file_path, framework="pt", device="cpu") as f:
         tensor_names = f.keys()
         tensor_dict = {}
         for name in tensor_names:
            tensor = f.get_tensor(name)
            tensor_dict[name] = tensor
         return tensor_dict

if __name__ == "__main__":
   directory_path = sys.argv[1]
   json_name = "model.safetensors.index.json"
   json_path = os.path.join(directory_path, json_name)
   with open(json_path, 'r', encoding='utf-8') as f:
         json_data = json.load(f)
   weight_map = json_data.get('weight_map', {})
   file_list = []
   for key in target_keys:
         safetensor_file = weight_map.get(key)
         file_list.append(directory_path + safetensor_file)

   new_dict = {}
   for file_path in file_list:
         tensor_dict = get_tensor_info(file_path)
         for key in target_keys:
            if key in tensor_dict:
               if key == "model.embed_tokens.weight":
                     new_key = "model.layers.78.embed_tokens.weight"
               elif key == "lm_head.weight":
                     new_key = "model.layers.78.shared_head.head.weight"
               new_dict[new_key] = tensor_dict[key]

   new_file_name = os.path.join(directory_path, "mtp-others.safetensors")
   new_keys = ["model.layers.78.embed_tokens.weight", "model.layers.78.shared_head.head.weight"]
   save_file(tensors=new_dict, filename=new_file_name)
   for key in new_keys:
         json_data["weight_map"][key] = "mtp-others.safetensors"
   with open(json_path, 'w', encoding='utf-8') as f:
         json.dump(json_data, f, indent=2)
  • glm-5-w8a8：需要 2 台 Atlas 800 A3（64G × 16）。

分别在两个节点上运行以下脚本。

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

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-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 \
--seed 1024 \
--served-model-name glm-5 \
--enable-expert-parallel \
--max-num-seqs 16 \
--max-model-len 200000 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.95 \
--quantization ascend \
--enable-chunked-prefill \
--enable-prefix-caching \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

node 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=1
export HCCL_BUFFSIZE=200
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_BALANCE_SCHEDULING=1
export VLLM_ASCEND_ENABLE_MLAPO=1
export VLLM_ASCEND_ENABLE_FLASHCOMM1=1

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-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 \
--seed 1024 \
--served-model-name glm-5 \
--enable-expert-parallel \
--max-num-seqs 16 \
--max-model-len 200000 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.95 \
--quantization ascend \
--enable-chunked-prefill \
--enable-prefix-caching \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "deepseek_mtp"}'

5.3 Prefill-Decode Disaggregation

We'd like to show the deployment guide of GLM-5 on multi-node environment with 1P1D for better performance. Prefill-Decode Disaggregation refers to the separation of the prefill stage and the decode stage across different nodes to improve throughput and latency.

Before you start, please

  1. prepare the script launch_online_dp.py on each node:

    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()

  2. prepare the script run_dp_template.sh on each node.

    To support a 200k context window on the stage of prefill, the parameter --additional_config needs to be added to "layer_sharding": ["q_b_proj", "o_proj"] on each prefill node. 1. Prefill node 0

    ```shell
nic_name="xxxx" # change to your own nic name
local_ip="xxxx" # 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=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
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

export VLLM_ASCEND_ENABLE_FUSED_MC2=1
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w8a8 \
    --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": 3, "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name glm-5 \
    --max-model-len 131072 \
    --additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true, "enable_dsa_cp": true}' \
    --max-num-batched-tokens 4096 \
    --trust-remote-code \
    --max-num-seqs 64 \
    --enable-chunked-prefill \
    --quantization ascend \
    --gpu-memory-utilization 0.95 \
    --enforce-eager \
    --enable-auto-tool-choice \
    --tool-call-parser glm47 \
    --reasoning-parser glm45 \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeConnectorV1",
    "kv_role": "kv_producer",
    "kv_port": "30000",
    "kv_connector_extra_config": {
                "use_ascend_direct": true,
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 16,
                        "tp_size": 4
                }
        }
    }'

```

    1. Prefill node 1

      nic_name="xxxx" # change to your own nic name
local_ip="xxxx" # 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=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
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 PYTORCH_NPU_ALLOC_CONF=expandable_segments:True

export VLLM_ASCEND_ENABLE_FLASHCOMM1=1

export VLLM_ASCEND_ENABLE_FUSED_MC2=1
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w8a8 \
    --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": 3, "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name glm-5 \
    --max-model-len 131072 \
    --additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true, "enable_dsa_cp": true}' \
    --max-num-batched-tokens 4096 \
    --trust-remote-code \
    --max-num-seqs 64 \
    --enable-chunked-prefill \
    --gpu-memory-utilization 0.95 \
    --quantization ascend \
    --enforce-eager \
    --enable-auto-tool-choice \
    --tool-call-parser glm47 \
    --reasoning-parser glm45 \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeConnectorV1",
    "kv_role": "kv_producer",
    "kv_port": "30000",
    "kv_connector_extra_config": {
                "use_ascend_direct": true,
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 16,
                        "tp_size": 4
                }
        }
    }'

    2. Decode node 0

      nic_name="xxxx" # change to your own nic name
local_ip="xxxx" # 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=1

export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
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

export VLLM_ASCEND_ENABLE_FUSED_MC2=1
export VLLM_ASCEND_ENABLE_MLAPO=1
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w8a8 \
    --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": 3,  "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name glm-5 \
    --max-model-len 200000 \
    --max-num-batched-tokens 32 \
    --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY"}' \
    --additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true, "recompute_scheduler_enable": true}' \
    --trust-remote-code \
    --max-num-seqs 8 \
    --gpu-memory-utilization 0.92 \
    --quantization ascend \
    --enable-auto-tool-choice \
    --tool-call-parser glm47 \
    --reasoning-parser glm45 \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeConnectorV1",
    "kv_role": "kv_consumer",
    "kv_port": "30100",
    "kv_connector_extra_config": {
                "use_ascend_direct": true,
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 16,
                        "tp_size": 4
                }
        }
    }'

    3. Decode node 1

      nic_name="xxxx" # change to your own nic name
local_ip="xxxx" # 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=1

export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
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

export VLLM_ASCEND_ENABLE_FUSED_MC2=1
export VLLM_ASCEND_ENABLE_MLAPO=1
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w8a8 \
    --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": 3,  "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name glm-5 \
    --max-model-len 200000 \
    --max-num-batched-tokens 32 \
    --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY"}' \
    --additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true, "recompute_scheduler_enable": true}' \
    --trust-remote-code \
    --max-num-seqs 8 \
    --gpu-memory-utilization 0.92 \
    --quantization ascend \
    --enable-auto-tool-choice \
    --tool-call-parser glm47 \
    --reasoning-parser glm45 \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeConnectorV1",
    "kv_role": "kv_consumer",
    "kv_port": "30100",
    "kv_connector_extra_config": {
                "use_ascend_direct": true,
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 16,
                        "tp_size": 4
                }
        }
    }'

    4. Decode node 2

      nic_name="xxxx" # change to your own nic name
local_ip="xxxx" # 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=1

export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
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

export VLLM_ASCEND_ENABLE_FUSED_MC2=1
export VLLM_ASCEND_ENABLE_MLAPO=1
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w8a8 \
    --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": 3,  "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name glm-5 \
    --max-model-len 200000 \
    --max-num-batched-tokens 32 \
    --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY"}' \
    --additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true, "recompute_scheduler_enable": true}' \
    --trust-remote-code \
    --max-num-seqs 8 \
    --gpu-memory-utilization 0.92 \
    --quantization ascend \
    --enable-auto-tool-choice \
    --tool-call-parser glm47 \
    --reasoning-parser glm45 \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeConnectorV1",
    "kv_role": "kv_consumer",
    "kv_port": "30100",
    "kv_connector_extra_config": {
                "use_ascend_direct": true,
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 16,
                        "tp_size": 4
                }
        }
    }'

    5. Decode node 3

      nic_name="xxxx" # change to your own nic name
local_ip="xxxx" # 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=1

export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
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

export VLLM_ASCEND_ENABLE_FUSED_MC2=1
export VLLM_ASCEND_ENABLE_MLAPO=1
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib

vllm serve /root/.cache/modelscope/hub/models/vllm-ascend/GLM5-w8a8 \
    --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": 3,  "method":"deepseek_mtp"}' \
    --profiler-config \
    '{"profiler": "torch",
    "torch_profiler_dir": "./vllm_profile",
    "torch_profiler_with_stack": false}' \
    --seed 1024 \
    --served-model-name glm-5 \
    --max-model-len 200000 \
    --max-num-batched-tokens 32 \
    --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY"}' \
    --additional-config '{"fuse_muls_add": true, "multistream_overlap_shared_expert": true, "recompute_scheduler_enable": true}' \
    --trust-remote-code \
    --max-num-seqs 8 \
    --gpu-memory-utilization 0.92 \
    --quantization ascend \
    --enable-auto-tool-choice \
    --tool-call-parser glm47 \
    --reasoning-parser glm45 \
    --kv-transfer-config \
    '{"kv_connector": "MooncakeConnectorV1",
    "kv_role": "kv_consumer",
    "kv_port": "30100",
    "kv_connector_extra_config": {
                "use_ascend_direct": true,
                "prefill": {
                        "dp_size": 2,
                        "tp_size": 16
                },
                "decode": {
                        "dp_size": 16,
                        "tp_size": 4
                }
        }
    }'

Once the preparation is done, you can start the server with the following command on each node:

  1. Prefill node 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 $node_p0_ip --dp-rpc-port 10521 --vllm-start-port 6700

  2. Prefill node 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 $node_p0_ip --dp-rpc-port 10521 --vllm-start-port 6700

  3. Decode node 0

    # change ip to your own
python launch_online_dp.py --dp-size 16 --tp-size 4 --dp-size-local 4 --dp-rank-start 0 --dp-address $node_d0_ip --dp-rpc-port 10523 --vllm-start-port 6721

  4. Decode node 1

    # change ip to your own
python launch_online_dp.py --dp-size 16 --tp-size 4 --dp-size-local 4 --dp-rank-start 4 --dp-address $node_d0_ip --dp-rpc-port 10523 --vllm-start-port 6721

  5. Decode node 2

    # change ip to your own
python launch_online_dp.py --dp-size 16 --tp-size 4 --dp-size-local 4 --dp-rank-start 8 --dp-address $node_d0_ip --dp-rpc-port 10523 --vllm-start-port 6721

  6. Decode node 3

    # change ip to your own
python launch_online_dp.py --dp-size 16 --tp-size 4 --dp-size-local 4 --dp-rank-start 12 --dp-address $node_d0_ip --dp-rpc-port 10523 --vllm-start-port 6721

5.4 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_server_example.py

unset http_proxy
unset https_proxy

python load_balance_proxy_server_example.py \
    --port 8000 \
    --host 0.0.0.0 \
    --prefiller-hosts \
       $node_p0_ip \
       $node_p1_ip \
    --prefiller-ports \
       6700 \
       6700 \
    --decoder-hosts \
      $node_d0_ip \
      $node_d0_ip \
      $node_d0_ip \
      $node_d0_ip \
      $node_d1_ip \
      $node_d1_ip \
      $node_d1_ip \
      $node_d1_ip \
      $node_d2_ip \
      $node_d2_ip \
      $node_d2_ip \
      $node_d2_ip \
      $node_d3_ip \
      $node_d3_ip \
      $node_d3_ip \
      $node_d3_ip \
    --decoder-ports \
      6721 6722 6723 6724 \
      6721 6722 6723 6724 \
      6721 6722 6723 6724 \
      6721 6722 6723 6724

注意：

一些优化配置如下所示：

  • VLLM_ASCEND_ENABLE_FLASHCOMM1：启用 FlashComm 优化以减少预填充节点上的通信和计算开销。启用 FlashComm 后，layer_sharding 列表不能包含 o_proj 作为元素。
  • VLLM_ASCEND_ENABLE_FUSED_MC2：启用以下融合算子：dispatch_gmm_combine_decode 和 dispatch_ffn_combine 算子。
  • VLLM_ASCEND_ENABLE_MLAPO：启用融合算子 MlaPreprocessOperation。

有关上述环境变量的进一步说明和限制，请参考以下 Python 文件：envs.py

6 功能验证

服务器启动后，您可以使用输入提示词查询模型：

curl http://<node0_ip>:<port>/v1/completions \
    -H "Content-Type: application/json" \
    -d '{
        "model": "glm-5",
        "prompt": "The future of AI is",
        "max_completion_tokens": 15,
        "temperature": 0
    }'

预期结果：

{"id": "chatcmlib-bc44ad093dec79a2", "object": "chat.completion", "created": "1770903266", "model": "glm-5", "choices": [{ "index": 0, "message": {"role": "assistant", "content": "The future of AI is not one thing, but a convergence of several powerful trends.", "annotations": "null", "audio": "null", "function_call": "null", "tool_calls": [], "reasoning": "null"}, "logprobs": "null", "finish_reason": "length", "stop_reason": "null", "token_ids": null}], "service_tier": "null", "system fingerprint": "null", "usage": {"prompt_tokens": 5, "total_tokens": 20, "completion_tokens": 15, "prompt_tokens_details": null}, "prompt_logprobs": "null", "prompt_token_ids": "null", "kv_transfer_params": null}

7 精度评估

7.1 使用 AISBench

  1. 详细信息请参考使用 AISBench 进行精度评估。

  2. 执行后即可获得结果。

8 性能评估

8.1 使用 AISBench

详细信息请参考使用 AISBench 进行性能评估。

8.2 使用 vLLM 基准测试

更多详细信息请参考 vLLM 基准测试。

9 性能调优

9.1 推荐配置

注意：以下配置在特定测试环境中验证，仅供参考。最佳配置取决于最大输入/输出长度、前缀缓存命中率、精度要求和部署机器比例等因素。建议根据实际条件参考第 9.2 节进行调优。

表 1：场景概览

场景 部署模式 *NPU 总数 权重版本 关键考量
高吞吐 1P1D 部署 32 (A3) GLM5-w8a8/GLM5.1-w8a8 P 节点上 dp4 tp8，D 节点上 dp8 tp4，以平衡延迟和吞吐
低延迟 1P1D 部署 32 (A3) GLM5-w8a8/GLM5.1-w8a8 P 和 D 节点上均使用 dp4 tp8 以降低延迟

*Total NPUs 表示所有节点上使用的 NPU 总数。

表 2：详细节点配置

场景 配置 NPU 数量 TP DP 最大序列数 最大批处理 Token 数 最大模型长度 MTP 推测数量
高吞吐 (A3) 1P1D 部署 32 P:8 D:4 P:4 D:8 P:64 D:128 P:4096 D:32 P:133120 D:150000 3
低延迟 (A3) 1P1D 部署 32 4 8 P:64 D:128 P:4096 D:32 P:133120 D:150000 3

10 常见问题

  • 常见问题提示：如果遇到问题，请参考常见问题解答。

  • 问：如何解决 ValueError：Tokenizer class TokenizersBackend does not exist or is not currently imported？

答：请将 transformers 版本升级到 5.2.0

  • 问：如何为 GLM-5 启用函数调用功能？

答：请在 vLLM 启动命令中添加以下配置

--tool-call-parser glm47 \
--reasoning-parser glm45 \
--enable-auto-tool-choice \