Skip to main content

Isaac Sim Installation Guide

Learning Objectives​

By the end of this section, you will:

  • Understand the different Isaac Sim installation options (native, container, cloud)
  • Install Isaac Sim on your local machine or set up cloud access
  • Verify your NVIDIA GPU and driver installation
  • Obtain educational licensing for Isaac Sim
  • Configure ROS 2 integration with Isaac Sim
  • Troubleshoot common installation issues

Before You Begin: GPU Check​

Isaac Sim requires an NVIDIA GPU. Let's verify your hardware first.

Check GPU on Linux​

# Check if NVIDIA GPU is present
lspci | grep -i nvidia

# Check NVIDIA driver version
nvidia-smi

Expected output:

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.147.05   Driver Version: 525.147.05   CUDA Version: 12.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  NVIDIA GeForce ...  Off  | 00000000:01:00.0  On |                  N/A |
|  0%   45C    P8    10W / 200W |    500MiB /  8192MiB |      2%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+

Minimum Requirements:

  • GPU: NVIDIA GTX 1060 (6GB VRAM) or better
  • Driver: Version 525+ recommended
  • CUDA: Version 11.8 or 12.0+

Check GPU on Windows​

# Open PowerShell and run:
nvidia-smi

Or open NVIDIA Control Panel and check your GPU model under "System Information."

No GPU? Choose an Alternative​

If you don't have an NVIDIA GPU:

  1. Cloud GPU (recommended): Rent GPU compute ($1-2/hour, ~$15-25 total)
  2. Shared University Lab: Use GPU workstations on campus
  3. Pre-Recorded Demos: Learn concepts without hands-on execution

Continue reading for cloud setup instructions.

Installation Option 1: Native Install (Ubuntu 22.04)​

Best for: Students with local NVIDIA GPU who want full control.

Time: 45-60 minutes Difficulty: Intermediate

Step 1: Install NVIDIA Drivers​

# Update package lists
sudo apt update

# Install NVIDIA driver
sudo apt install nvidia-driver-525

# Reboot to load driver
sudo reboot

# Verify installation after reboot
nvidia-smi

Troubleshooting: If nvidia-smi shows errors, try:

sudo ubuntu-drivers autoinstall
sudo reboot

Step 2: Install Docker and NVIDIA Container Toolkit​

Isaac Sim can run in a Docker container (recommended for easier setup):

# Install Docker
curl -fsSL https://get.docker.com -o get-docker.sh
sudo sh get-docker.sh
sudo usermod -aG docker $USER
newgrp docker

# Install NVIDIA Container Toolkit
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | \
  sudo tee /etc/apt/sources.list.d/nvidia-docker.list

sudo apt update
sudo apt install -y nvidia-docker2
sudo systemctl restart docker

# Test GPU access in Docker
docker run --rm --gpus all nvidia/cuda:11.8.0-base-ubuntu22.04 nvidia-smi

Expected: You should see the same nvidia-smi output as before.

Step 3: Get Educational License​

NVIDIA Isaac Sim is free for students but requires registration:

  1. Go to NVIDIA Developer Program
  2. Click "Get Started" or "Download"
  3. Sign in with NVIDIA account (create if needed)
  4. Select "Educational Use" when prompted
  5. Accept terms and conditions
  6. You'll get access to download links and NGC credentials

Educational License Benefits:

  • Free access to Isaac Sim
  • Access to NVIDIA NGC container registry
  • Access to pre-trained models and datasets
  • Community forum support

Step 4: Download and Install Isaac Sim​

Option A: Omniverse Launcher (Easier)

  1. Download Omniverse Launcher
  2. Install and sign in with your NVIDIA account
  3. In Launcher, go to "Exchange" tab
  4. Search for "Isaac Sim"
  5. Click "Install" (select version 2023.1.1)
  6. Wait for download (10-20 GB)

Option B: Direct Download

# Download Isaac Sim (requires NGC credentials)
# This command downloads the standalone package
wget --content-disposition 'https://api.ngc.nvidia.com/v2/resources/nvidia/isaac-sim/versions/2023.1.1/zip' \
  -O isaac-sim-2023.1.1.zip

# Extract
unzip isaac-sim-2023.1.1.zip -d ~/isaac-sim

# Run setup script
cd ~/isaac-sim
./setup_python_env.sh

Step 5: Run Isaac Sim​

# Navigate to Isaac Sim directory
cd ~/isaac-sim

# Launch Isaac Sim
./isaac-sim.sh

# Or use Omniverse Launcher and click "Launch" button

First Launch:

  • May take 2-5 minutes to start
  • Will download additional dependencies
  • May show shader compilation progress

Expected: Isaac Sim GUI opens with the main viewport.

Step 6: Configure ROS 2 Integration​

# Install ROS 2 Humble if not already installed
sudo apt install ros-humble-desktop-full

# Source ROS 2
source /opt/ros/humble/setup.bash

# Install Isaac Sim ROS 2 bridge dependencies
sudo apt install ros-humble-ros-ign-gazebo ros-humble-ros-ign-bridge

# Add to your ~/.bashrc
echo "source /opt/ros/humble/setup.bash" >> ~/.bashrc

Step 7: Verify Installation​

Run the verification script:

# Download verification script
cd ~/isaac-sim
python -c "
from omni.isaac.kit import SimulationApp
simulation_app = SimulationApp({'headless': False})
from omni.isaac.core import World
world = World()
print('Isaac Sim installed successfully!')
simulation_app.close()
"

Expected output: Isaac Sim installed successfully!

Best for: Students who want easier setup with Docker.

Time: 20-30 minutes Difficulty: Beginner-Intermediate

Step 1: Set Up NGC Account​

  1. Go to NVIDIA NGC
  2. Sign up with your educational email
  3. Go to "Setup" → "Get API Key"
  4. Generate an API key (save it securely)

Step 2: Log in to NGC Registry​

# Log in to NGC container registry
docker login nvcr.io
# Username: $oauthtoken
# Password: <your-ngc-api-key>

Step 3: Pull Isaac Sim Container​

# Pull the Isaac Sim container (8-12 GB)
docker pull nvcr.io/nvidia/isaac-sim:2023.1.1

# Verify the image
docker images | grep isaac-sim

Step 4: Run Isaac Sim Container​

# Create directory for persistent data
mkdir -p ~/isaac-sim-data

# Run Isaac Sim container
docker run --name isaac-sim --entrypoint bash -it --gpus all \
  -e "ACCEPT_EULA=Y" \
  -v ~/isaac-sim-data:/root/workspace \
  -p 8211:8211 -p 8899:8899 \
  nvcr.io/nvidia/isaac-sim:2023.1.1

Inside the container:

# Launch Isaac Sim
./runapp.sh

# Or run headless (no GUI) for data generation
./runheadless.native.sh

Step 5: Access Web UI (Optional)​

Isaac Sim container supports web-based streaming:

  1. Run container with web streaming:
docker run --name isaac-sim --entrypoint bash -it --gpus all \
  -e "ACCEPT_EULA=Y" \
  -v ~/isaac-sim-data:/root/workspace \
  -p 8211:8211 -p 8899:8899 -p 8888:8888 \
  nvcr.io/nvidia/isaac-sim:2023.1.1

# Inside container, start streaming
./runapp.sh --enable-webrtc
  1. Open browser to http://localhost:8211
  2. You'll see Isaac Sim viewport in your browser

Benefits: Access Isaac Sim from any device on your network.

Installation Option 3: Cloud GPU (AWS)​

Best for: Students without local NVIDIA GPU.

Cost: $1-2/hour ($15-25 for full module) Time: 30-45 minutes setup Difficulty: Intermediate

Step 1: Create AWS Account​

  1. Go to AWS
  2. Sign up (requires credit card, free tier available)
  3. Verify email and complete account setup

Step 2: Request GPU Instance Quota​

AWS limits GPU instances by default:

  1. Go to Service Quotas in AWS Console
  2. Search for EC2
  3. Find "Running On-Demand G instances"
  4. Request quota increase to 4 vCPUs (allows 1× g5.xlarge)
  5. Wait for approval (usually 1-2 business days)

Step 3: Launch EC2 Instance​

  1. Go to EC2 Dashboard
  2. Click "Launch Instance"
  3. Configure:
    • Name: isaac-sim-instance
    • AMI: Ubuntu Server 22.04 LTS (HVM), SSD Volume Type
    • Instance Type: g5.xlarge (NVIDIA A10G GPU, 4 vCPUs, 16 GB RAM)
    • Storage: 50 GB gp3 (SSD)
    • Security Group: Allow SSH (port 22), HTTP (8211), custom ports if needed
  4. Click "Launch"
  5. Download the .pem key file

Step 4: Connect to Instance​

# Set permissions on key file
chmod 400 your-key.pem

# Connect to instance
ssh -i your-key.pem ubuntu@<instance-public-ip>

Step 5: Install Isaac Sim on AWS​

Once connected:

# Update system
sudo apt update && sudo apt upgrade -y

# Install NVIDIA drivers
sudo apt install nvidia-driver-525 -y
sudo reboot

Reconnect after reboot:

ssh -i your-key.pem ubuntu@<instance-public-ip>

# Verify GPU
nvidia-smi

Now follow Option 2 (NGC Container) steps to pull and run Isaac Sim.

Step 6: Access from Local Machine​

Option A: X11 Forwarding (if you have X server on local machine):

ssh -X -i your-key.pem ubuntu@<instance-public-ip>

Option B: Web Streaming (recommended):

  • Run Isaac Sim with --enable-webrtc
  • Access via http://<instance-public-ip>:8211

Step 7: Cost Management​

Important: Stop instance when not in use!

# From AWS Console: EC2 → Instances → Select → Instance State → Stop

Costs:

  • g5.xlarge: ~$1.00-1.50/hour (varies by region)
  • Storage: ~$5/month for 50 GB
  • Total for module: ~$15-25 if you stop instance when not using

Tips:

  • Use Spot Instances for 60-70% discount (may be interrupted)
  • Delete instance after completing module to avoid ongoing charges
  • Set up billing alerts to avoid surprises

Installation Option 4: Google Cloud Platform​

Similar to AWS, but with different pricing:

  1. Create GCP account
  2. Enable Compute Engine API
  3. Launch instance:
    • Machine type: n1-standard-4 + 1× NVIDIA T4
    • OS: Ubuntu 22.04 LTS
    • Storage: 50 GB SSD
  4. Install NVIDIA drivers and Isaac Sim (same as AWS)

Cost: ~$0.80-1.20/hour + storage

Installation Option 5: Shared University Lab​

Many universities provide GPU workstations:

  1. Contact Your Department:

    • Ask about GPU lab access for coursework
    • Request Isaac Sim installation if not already available

  2. Schedule Lab Time:

    • Reserve specific time slots for Isaac work
    • Coordinate with other students

  3. Remote Access:

    • Ask for SSH/VNC access to work remotely
    • Use web streaming if available

Post-Installation: Verify Setup​

Test 1: Launch Isaac Sim​

# Native or container
./isaac-sim.sh
# Or from Omniverse Launcher

Expected: GUI opens without errors.

Test 2: Create a Simple Scene​

In Isaac Sim:

  1. File → New Stage (or Ctrl+N)
  2. Create → Mesh → Cube
  3. Click Play button (bottom-left)
  4. Cube should fall due to gravity

Expected: Cube falls and hits ground plane.

Test 3: Check GPU Utilization​

While Isaac Sim is running:

nvidia-smi

Expected: GPU utilization > 0%, memory usage 1-4 GB.

Test 4: ROS 2 Bridge Test​

# Terminal 1: Source ROS 2
source /opt/ros/humble/setup.bash

# Terminal 2: Launch Isaac Sim with ROS 2 bridge
./isaac-sim.sh

In Isaac Sim:

  1. Isaac Examples → ROS2 → Clock
  2. Click Play

In Terminal 1:

ros2 topic list

Expected: You should see /clock topic published by Isaac Sim.

Troubleshooting Common Issues​

Issue 1: nvidia-smi Not Found​

Solution:

# Reinstall NVIDIA driver
sudo apt install --reinstall nvidia-driver-525
sudo reboot

Issue 2: Docker Permission Denied​

Solution:

sudo usermod -aG docker $USER
newgrp docker
# Or log out and log back in

Issue 3: Isaac Sim Crashes on Launch​

Solution:

# Check GPU memory
nvidia-smi

# Close other GPU applications
# Try launching with reduced resolution
./isaac-sim.sh --width 1280 --height 720

Issue 4: Container Cannot Access GPU​

Solution:

# Verify NVIDIA Container Toolkit
docker run --rm --gpus all nvidia/cuda:11.8.0-base-ubuntu22.04 nvidia-smi

# If fails, reinstall
sudo apt purge nvidia-docker2
sudo apt install nvidia-docker2
sudo systemctl restart docker

Issue 5: Slow Performance​

Possible causes:

  • Insufficient VRAM (< 4 GB)
  • CPU bottleneck
  • Thermal throttling

Solutions:

# Check GPU temperature
nvidia-smi

# Reduce physics rate in Isaac Sim
# Edit → Preferences → Physics → Simulation Rate → 30 Hz

# Disable ray tracing if enabled
# Viewport → Rendering → Real-Time

Issue 6: ROS 2 Bridge Not Working​

Solution:

# Ensure ROS 2 is sourced
source /opt/ros/humble/setup.bash

# Check ROS_DOMAIN_ID matches
echo $ROS_DOMAIN_ID
# Set if needed
export ROS_DOMAIN_ID=0

# Restart Isaac Sim

Hardware Recommendations by Budget​

Budget: $0 (No GPU)​

Options:

  1. Shared University Lab (free)
  2. Pre-Recorded Demos (learn concepts without hands-on)

Budget: $15-50​

Recommended: AWS/GCP Cloud GPU

  • Rent for 10-20 hours
  • Stop instance when not using
  • Good for learning and short projects

Budget: $200-500​

Recommended: Used NVIDIA GTX 1060/1070 or RTX 2060

  • Buy used on eBay/Craigslist
  • Sufficient for Isaac Sim
  • Useful beyond this course

Budget: $800-1500​

Recommended: New NVIDIA RTX 3060/4060 or better

  • Excellent performance
  • Future-proof for robotics work
  • Good for research and professional development

Summary​

You have multiple options for running Isaac Sim:

OptionTimeCostDifficultyBest For
Native Install45-60 min$0 (if GPU)MediumLocal GPU owners
NGC Container20-30 min$0 (if GPU)Easy-MediumDocker users
AWS Cloud30-45 min$15-25MediumNo local GPU
GCP Cloud30-45 min$15-25MediumNo local GPU
University LabVaries$0EasyStudents with lab access
Pre-Recorded0 min$0EasyLearn without hands-on

Recommended Path:

  • Have GPU? → Native Install or NGC Container
  • No GPU? → Cloud (AWS/GCP) or University Lab
  • Limited budget? → Pre-Recorded Demos (concept learning)

Next Steps:

  1. Choose your installation option
  2. Complete setup and verification
  3. Continue to Creating Photorealistic Scenes

Additional Resources​