Isaac Sim 安装与使用

0. Isaac Sim

Isaac Sim是一个从头开始构建的软件平台,旨在支持日益自动化的世界。目标是使您尽可能容易地为真实的物理机器人设计、调整、训练和部署自主控制代理。
请添加图片描述
请添加图片描述

电脑配置要求

检测自己的电脑满不足满足最低要求,最低GPU显卡最低是3070 8GB显存,英特尔酷睿i7(第7代),32GB内存。操作系统可以支持Ubuntu和Windows系统,但是The Isaac Sim container 只支持Linux,所以推荐安装Ubuntu 22.04系统,在Linux上运行Isaac Sim。

https://docs.omniverse.nvidia.com/isaacsim/latest/installation/requirements.html

请添加图片描述

1.双系统的安装

1.1压缩硬盘

Isaac Sim 最低要求是50GB左右空间。所以我们需要准备预留50G进行系统安装,找到计算机管理->磁盘管理

请添加图片描述

找到一个空余磁盘,例如D盘,右键压缩卷,分出来50GB就行。

请添加图片描述
按需求分配即可。
请添加图片描述

现在固态硬盘比以前便宜很多了,国产的硬盘性价比也高,我这里是准备了一块1tb的硬盘给到ubuntu系统,就是953GB未分配。

1.2 下载Ubuntu镜像系统

需要注意一点,Isaac Sim对ubuntu 20.4的ROS1和ROS2不再支持,并将来打算删除,推荐安装ubuntu 22.04,别下载成最新的ubuntu 24.04, Isaac Sim暂时不支持哈。

https://www.releases.ubuntu.com/22.04/

请添加图片描述

请添加图片描述

1.3 把U盘制作系统盘

https://rufus.ie/en/

下载rufus软件

请添加图片描述

将准备好的u盘插在电脑上
打开刚刚下载的rufus软件

请添加图片描述
请添加图片描述
请添加图片描述

请添加图片描述
请添加图片描述
请添加图片描述

完成后,关闭窗口然后电脑重启就行了。

1.4 通过BIOS进入系统盘

上网搜索一下,你的主板如何进入BIOS的,不同主板的操作会有略微不同,但是思路是一样的,我的主板是微星,操作就是就是开机后,一直敲键盘上的"Del"按键进入BIOS,然后要设置两个地方:1.将U盘设置为开机启动项。2.将“安全启动”关闭。

1.微星的开机启动项设置比较简单,用鼠标把UEFI USB往前拖动就行了。

请添加图片描述

2.将“安全启动”关闭,这里顺序"SETTINGS"->“安全”->“安全引导”->“安全启动”-> “允许改为禁止”。

然后F10保存重启就行了。如果你实在不知道自己电脑主板怎么进入BIOS设置,可以去某东或者某宝找到主板品牌的人工客服问问怎么设置。反正我一开始找不到安全启动就是问客服怎么搞的,哈哈。

请添加图片描述
请添加图片描述
请添加图片描述

1.5 设置分区

重启后就会进入U盘的系统引导,我们需要将U盘的系统引导复制到刚刚分配硬盘中。

1.系统我会设置为英文。也建议系统使用英文,可以避免中文路径编码的问题。
请添加图片描述
请添加图片描述

2.会让你选择连接一个网络,我这里插着网线,就自动帮我跳过了。

3.常规软件的安装,这个默认选项继续就行。

请添加图片描述

4.选择其他安装选项

请添加图片描述

5.分区思路

看了很多博客一般分区思路如下,需要划分四个区域

  • /根目录 跟/home平分
  • /biosgrub 分配500MB-2GB
  • /swap 分配内存大小一样
  • /home 跟/根目录平分

那么举个例子,我的1000G分配如下

  • / 分配 460GB
  • /biosgrub 分配1GB
  • /swap 分配64GB
  • /home 分配475GB

6.分区操作

找到free space,鼠标双击后,/根目录分配460GB,选择逻辑分区,盘符填写"/"

### IsaacLab Installation Guide for Ubuntu 20.04 For installing IsaacLab on an Ubuntu 20.04 system, a series of preparatory steps must be completed to ensure compatibility and functionality. #### Preparing the System Environment Ensure that all necessary software is installed as described in tutorials covering essential installations for Ubuntu 20.04[^2]. This includes setting up the correct time zone, configuring input methods such as Sogou Pinyin, deploying Anaconda for Python development environments, using Terminator terminal emulator, ensuring Nvidia drivers are properly set up which is crucial since IsaacLab relies heavily on GPU acceleration capabilities provided by Nvidia hardware. #### Installing CUDA Toolkit Since IsaacLab requires CUDA support, follow guidelines similar to those outlined for WSL 2 but adapted specifically for native Linux distributions like Ubuntu 20.04[^4]. Install the appropriate version of the CUDA toolkit compatible with your graphics card model through either package managers or direct downloads from Nvidia's official website. ```bash sudo apt-get update && sudo apt-get install cuda ``` After installation completes successfully, verify it works correctly: ```bash nvcc --version ``` #### Setting Up cuDNN Following successful setup of CUDA, proceed to configure cuDNN according to instructions found within documentation related to Autoware environment configuration under Ubuntu 20.04[^3]: Download `libcudnn8` and its developer library corresponding to the chosen CUDA version before executing commands below: ```bash sudo dpkg -i libcudnn8_*.deb sudo dpkg -i libcudnn8-dev_*.deb ``` Confirm proper integration via running tests included inside sample directories bundled alongside cuDNN packages. #### Obtaining and Configuring IsaacLab Once prerequisites have been met, obtain access rights to download IsaacSim/IsaacLab SDK directly from Nvidia’s developer portal after registering there first if not done already. Follow detailed guides available online regarding unpacking archives, placing files into designated locations while adhering closely to any specific requirements mentioned during this process concerning dependencies management tools used throughout projects built around Omniverse platform components including IsaacLab itself. Finally, customize settings based upon personal preferences or project needs following post-installation procedures recommended officially by developers behind these applications. --related questions-- 1. What versions of CUDA should one choose when preparing their machine learning workstation? 2. How does adjusting system locale affect non-English speaking users' experiences on fresh installs of Ubuntu-based systems? 3. Can you explain how to troubleshoot common issues encountered after updating Nvidia drivers on Linux machines? 4. In what ways can integrating Jupyter notebooks enhance productivity within data science workflows conducted over Conda-managed virtual environments? 5. Are there alternative text editors besides Vim/GNU Emacs suitable for beginners who wish to write scripts efficiently without steep learning curves?
评论 1
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值