[Orin Nx] 如何跑满GPU和CPU，观察温度和散热性能？

1、环境说明

硬件: Nvidia Orin NX 16GB

软件：Jetson Linux R35.3.1

2、工具安装部署

GPU的压力测试主要使用工具：

https://github.com/anseeto/jetson-gpu-burn

CPU的压力测试主要使用工具 stress

注意安装 jetson-gpu-burn需要在 /etc/apt/sources.list 添加如下源：

root@w:~# cat /etc/apt/sources.list.d/nvidia-l4t-apt-source.list 
# SPDX-FileCopyrightText: Copyright (c) 2019-2021 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: LicenseRef-NvidiaProprietary
#
# NVIDIA CORPORATION, its affiliates and licensors retain all intellectual
# property and proprietary rights in and to this material, related
# documentation and any modifications thereto. Any use, reproduction,
# disclosure or distribution of this material and related documentation
# without an express license agreement from NVIDIA CORPORATION or
# its affiliates is strictly prohibited.

deb https://repo.download.nvidia.com/jetson/common r35.3 main
deb https://repo.download.nvidia.com/jetson/t234 r35.3 main

简单点操作：

 cat /etc/apt/sources.list.d/nvidia-l4t-apt-source.list >> /etc/apt/sources.list

然后执行如下安装命令进行安装

sudo apt-get update
sudo apt-get install cuda-toolkit-11-4
sudo apt-get install stress
sudo apt-get -y install pip
sudo apt-get -y install python-pip
sudo -H pip install -U jetson-stats

在设备上下载jetson-gpu-burn 源码进行编译

git clone https://github.com/anseeto/jetson-gpu-burn.git

cd jetson-gpu-burn

make

编译成功会生成一个 gpu_burn 的二进制可执行程序，执行该程序即可跑满gpu.

3、观测不同电源模式进行CPU、GPU性能和温度

跑满CPU：8个核

root@w:~# stress -c 8 &
[1] 3302
root@w:~# stress: info: [3302] dispatching hogs: 8 cpu, 0 io, 0 vm, 0 hdd

跑满GPU：

root@w:~/jetson-gpu-burn# while true;do ./gpu_burn ;done
Run length not specified in the command line.  Burning for 10 secs

执行jtop命令观察

可以看到我们设备的电源模式是最大功率MAXN的模式，GPU 918MHz和CPU 2.0GHz 8个核都已经全负荷工作，此时实时的功率24.8W左右，CPU和GPU的温度可以实时看到。

如果设备的散热性能不好的话，CPU和GPU的温度会很高，可能会影响性能表现，比如GPU和CPU会在高温时出现降频。

下面是Orin Nx模块的最大性能和工作温度范围：

如果要Orin Nx达到最大性能，设备的散热问题做好设计是不可忽略的。

4、trouble-shot:

root@w:~# jtop 

Traceback (most recent call last):
  File "/usr/local/bin/jtop", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.8/dist-packages/jtop/__main__.py", line 159, in main
    curses.wrapper(JTOPGUI, jetson, pages, init_page=args.page,
  File "/usr/lib/python3.8/curses/__init__.py", line 105, in wrapper
    return func(stdscr, *args, **kwds)
  File "/usr/local/lib/python3.8/dist-packages/jtop/gui/jtopgui.py", line 79, in __init__
    NColors(color_filter)
  File "/usr/local/lib/python3.8/dist-packages/jtop/gui/lib/colors.py", line 43, in __init__
    curses.init_pair(NColors.RED, curses.COLOR_RED if not color_filter else curses.COLOR_BLUE, curses.COLOR_BLACK)
_curses.error: init_pair() returned ERR

解决办法：

export TERM='xterm-256color'

root@w:~# jtop 
The jtop.service is not active. Please run:
sudo systemctl restart jtop.service

解决方法：

systemctl restart jtop.service