---
title: 环境要求
description: "硬件要求如表1所示。"
url: https://www.hikunpeng.com/document/detail/zh/kunpenghpcs/hpcindapp/prtg-osc/kunpengpytorch_02_0002.html
sourcePath: /source/zh/kunpenghpcs/hpcindapp/prtg-osc/kunpengpytorch_02_0002.html
indexId: 4ffd561a8eff32939c51695f0d1699df90deaaf8919d92a496d27917429ae58f69
---
# 环境要求

#### 硬件要求

硬件要求如表1所示。


**表1 硬件要求**

| 项目 | 说明 |
| --- | --- |
| CPU | 鲲鹏920处理器 |
| GPU | NVIDIA Tesla V100 |


#### 软件要求

软件要求如表2所示。


**表2 软件要求**

| 项目 | 版本 | 下载地址 |
| --- | --- | --- |
| Docker | 19.03.15 | https://download.docker.com/linux/static/stable/aarch64/docker\-19.03.15.tgz(https://download.docker.com/linux/static/stable/aarch64/docker\-19.03.15.tgz) |
| NVIDIA Container Toolkit | Yum安装 | https://docs.nvidia.com/datacenter/cloud\-native/container\-toolkit/install\-guide.html(https://docs.nvidia.com/datacenter/cloud\-native/container\-toolkit/install\-guide.html) |
| PyTorch | 1.9 | https://github.com/pytorch/pytorch(https://github.com/pytorch/pytorch) |
| Cuda | 11.1 | https://registry.hub.docker.com/r/nvidia/cuda\-arm64/tags?page=1&ordering=last\_updated&name=11.1(https://registry.hub.docker.com/r/nvidia/cuda\-arm64/tags?page=1&ordering=last\_updated&name=11.1) |
| Cudnn | 8 | https://developer.download.nvidia.cn/compute/machine\-learning/repos/rhel8/sbsa/(https://developer.download.nvidia.cn/compute/machine\-learning/repos/rhel8/sbsa/) |


#### 操作系统要求

操作系统要求如表3所示。


**表3 操作系统要求**

| 项目 | 版本 | 下载地址 |
| --- | --- | --- |
| CentOS | 8.2 | https://vault.centos.org/8.2.2004/isos/aarch64/(https://vault.centos.org/8.2.2004/isos/aarch64/) |