Optimizing Linux for AI
Published by Pearson
Tune Linux to get the most out of your on-premises GenAI solutions
Course Outcomes
- Prepare a production on-prem platform for hosting AI.
- Get real-time insight into the current handling of AI workloads.
- Tweak your system in real time to see immediate results.
Hosting GenAI on premises is imperative if you want to use AI capabilities to work with your private data. Linux is the most effective operating system to host Generative AI solutions on- prem. This flexible operating system has all the tools that allow for optimal tuning to get the best possible performance for all key hardware components involved in AI.
This 3-hour course introduces best practices for optimizing Linux for AI. You’ll learn how to install and configure the right drivers for your hardware, analyze current performance, and optimize the Linux kernel to get the best possible performance for managing the LLM lifecycle on Linux.
What you’ll learn and how you can apply it
- Linux kernel tunables for AI optimization
- The role and benefits of using the right driver
- Open-source and proprietary Linux tools and when to use them for specific tasks
- How to analyze current Linux performance while managing the AI lifecycle
This live event is for you because...
- You are a Linux Administrator and need to get the best possible performance out of your Linux system
- You are a data scientist who wants to secure your GenAI system by hosting it on premises
- You are a developer who wants to tweak drivers for handling AI workloads for better performance
Prerequisites
- Students need to be familiar with the Linux operating system and know how to configure it for handling common tasks
Course Set-up
- Students should have a Linux system to follow along with the demos in this course. A generic installation of any Red Hat related distribution or Ubuntu will do. Specific AI-related hardware such as GPUs are not required, but it does make following along easier.
Recommended Preparation
- Attend: Linux Fundamentals bootcamp by Sander van Vugt
- Watch: Linux Fundamentals by Sander van Vugt
Recommended Follow-up
- Watch: Linux under the Hood by Sander van Vugt
- Attend: Open Source GenAI with InstructLab and Granite by Sander van Vugt
Schedule
The time frames are only estimates and may vary according to how the class is progressing.
Segment 1: Introduction (5 minutes)
Segment 2: Configuring Drivers for AI (40 minutes)
- Demo: Replacing the Nvidia nouveau driver with the proprietary driver
- Demo: Managing CUDA parameters
Q&A (5 minutes)
Break (5 minutes)
Segment 3: Analyzing Linux Hardware Configuration (25 minutes)
- Demo: Verifying installed hardware devices
- Demo: Checking current driver configuration
Segment 4: Analyzing Linux System Performance (25 minutes)
- Using common tools to monitor performance parameters
- Focusing on AI-related performance parameters
- Demo: Using GPU tools to analyze GPU hardware
Q&A (5 minutes)
Break (5 minutes)
Segment 5: Optimizing the Linux kernel for AI (60 minutes)
- Identifying relevant kernel parameters
- Testing new kernel parameters
- Making tweaked kernel parameters persistent
Q&A (5 minutes)
Your Instructor
Sander van Vugt