优化与调优

优化与调优#

本指南旨在帮助用户在系统层面提升 vllm-ascend 的性能。内容涵盖操作系统配置、库优化、部署指南等。欢迎任何反馈。

准备工作#

运行容器：

# Update DEVICE according to your device (/dev/davinci[0-7])
export DEVICE=/dev/davinci0
# Update the cann base image
export IMAGE=m.daocloud.io/quay.io/ascend/cann:8.3.rc2-910b-ubuntu22.04-py3.11
docker run --rm \
--name performance-test \
--device $DEVICE \
--device /dev/davinci_manager \
--device /dev/devmm_svm \
--device /dev/hisi_hdc \
-v /usr/local/dcmi:/usr/local/dcmi \
-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

配置您的环境：

# Configure the mirror
echo "deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu-ports/ jammy main restricted universe multiverse" > /etc/apt/sources.list && \
echo "deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu-ports/ jammy main restricted universe multiverse" >> /etc/apt/sources.list && \
echo "deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu-ports/ jammy-updates main restricted universe multiverse" >> /etc/apt/sources.list && \
echo "deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu-ports/ jammy-updates main restricted universe multiverse" >> /etc/apt/sources.list && \
echo "deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu-ports/ jammy-backports main restricted universe multiverse" >> /etc/apt/sources.list && \
echo "deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu-ports/ jammy-backports main restricted universe multiverse" >> /etc/apt/sources.list && \
echo "deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu-ports/ jammy-security main restricted universe multiverse" >> /etc/apt/sources.list && \
echo "deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu-ports/ jammy-security main restricted universe multiverse" >> /etc/apt/sources.list

# Install os packages
apt update && apt install wget gcc g++ libnuma-dev git vim -y

安装 vllm 和 vllm-ascend：

# Install necessary dependencies
pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
pip install modelscope pandas datasets gevent sacrebleu rouge_score pybind11 pytest

# Configure this var to speed up model download
VLLM_USE_MODELSCOPE=true

请按照 安装指南 操作，确保 vLLM、vllm-ascend 和 MindIE Turbo 已正确安装。

备注

请确保在完成 Python 配置后再安装 vLLM 和 vllm-ascend，因为这些包会使用当前环境中的 Python 构建二进制文件。如果您在完成 1.1 节之前安装了 vLLM、vllm-ascend 和 MindIE Turbo，则生成的二进制文件将无法利用优化后的 Python。

优化项#

1. Compilation Optimization#

1.1. 安装优化版 python#

Python 从 3.6 及以上版本开始支持 LTO 和 PGO 优化，这些优化可以在编译时启用。为了方便起见，我们直接向用户提供优化后的 python 软件包。您也可以参考此 教程 根据具体场景自行构建。

mkdir -p /workspace/tmp
cd /workspace/tmp

# Download prebuilt lib and packages
wget https://repo.oepkgs.net/ascend/pytorch/vllm/lib/libcrypto.so.1.1
wget https://repo.oepkgs.net/ascend/pytorch/vllm/lib/libomp.so
wget https://repo.oepkgs.net/ascend/pytorch/vllm/lib/libssl.so.1.1
wget https://repo.oepkgs.net/ascend/pytorch/vllm/python/py311_bisheng.tar.gz

# Configure python and pip
cp ./*.so* /usr/local/lib
tar -zxvf ./py311_bisheng.*  -C /usr/local/
mv  /usr/local/py311_bisheng/  /usr/local/python
sed -i "1c#\!/usr/local/python/bin/python3.11" /usr/local/python/bin/pip3
sed -i "1c#\!/usr/local/python/bin/python3.11" /usr/local/python/bin/pip3.11
ln -sf  /usr/local/python/bin/python3  /usr/bin/python
ln -sf  /usr/local/python/bin/python3  /usr/bin/python3
ln -sf  /usr/local/python/bin/python3.11  /usr/bin/python3.11
ln -sf  /usr/local/python/bin/pip3  /usr/bin/pip3
ln -sf  /usr/local/python/bin/pip3  /usr/bin/pip

export PATH=/usr/bin:/usr/local/python/bin:$PATH

2. OS Optimization#

2.1. jemalloc#

jemalloc 是一种内存分配器，可提升多线程场景下的性能并减少内存碎片。jemalloc 使用线程本地内存管理来分配变量，从而避免线程间的锁竞争，能极大优化性能。

# Install jemalloc
sudo apt update
sudo apt install libjemalloc2

# Configure jemalloc
export LD_PRELOAD=/usr/lib/"$(uname -i)"-linux-gnu/libjemalloc.so.2 $LD_PRELOAD

2.2. Tcmalloc#

Tcmalloc (Thread Counting Malloc) 是一种通用内存分配器，它通过引入多级缓存结构、减少互斥锁竞争并优化大对象处理流程，在确保低延迟的同时提升整体性能。点击 此处 了解更多细节。

# Install tcmalloc
sudo apt update
sudo apt install libgoogle-perftools4 libgoogle-perftools-dev

# Get the location of libtcmalloc.so*
find /usr -name libtcmalloc.so*

# Make the priority of tcmalloc higher
# The <path> is the location of libtcmalloc.so we get from the upper command
# Example: "$LD_PRELOAD:/usr/lib/aarch64-linux-gnu/libtcmalloc.so"
export LD_PRELOAD="$LD_PRELOAD:<path>"

# Verify your configuration
# The path of libtcmalloc.so will be contained in the result if your configuration is valid
ldd `which python`

3. torch_npu Optimization#

torch_npu 中的某些性能调优特性受环境变量控制。部分特性及其相关的环境变量如下所示。

内存优化：

# Upper limit of memory block splitting allowed (MB): Setting this parameter can prevent large memory blocks from being split.
export PYTORCH_NPU_ALLOC_CONF="max_split_size_mb:250"

# When operators on the communication stream have dependencies, they all need to be ended before being released for reuse. The logic of multi-stream reuse is to release the memory on the communication stream in advance so that the computing stream can be reused.
export PYTORCH_NPU_ALLOC_CONF="expandable_segments:True"

调度优化：

# Optimize operator delivery queue. This will affect the memory peak value, and may degrade if the memory is tight.
export TASK_QUEUE_ENABLE=2

# This will greatly improve the CPU bottleneck model and ensure the same performance for the NPU bottleneck model.
export CPU_AFFINITY_CONF=1

4. CANN Optimization#

4.1. HCCL 优化#

HCCL 中包含一些性能调优特性，这些特性通过环境变量进行控制。

您可以通过设置下方的环境变量，配置 HCCL 使用 "AIV" 模式来优化性能。在 "AIV" 模式下，通信由 AI Vector Core 配合 RoCE 直接调度，不再由 AI CPU 调度。

export HCCL_OP_EXPANSION_MODE="AIV"

此外，针对特定场景还有更多性能优化特性，如下所示。

  • HCCL_INTRA_ROCE_ENABLE: 在两台 8P 机器之间使用 RDMA 链路代替 SDMA 链路作为 Mesh 互联链路。点击 此处 了解更多细节。

  • HCCL_RDMA_TC: 使用该变量配置 RDMA 网卡的流量类别 (Traffic Class)。点击 此处 了解更多细节。

  • HCCL_RDMA_SL: 使用该变量配置 RDMA 网卡的服务级别 (Service Level)。点击 此处 了解更多细节。

  • HCCL_BUFFSIZE: 使用该变量控制两个 NPU 之间共享数据的缓存大小。点击 此处 了解更多细节。

目录