RHEL gives Linux a much-needed AI update

A few days ago, Dell Technologies teamed up with Red Hat Inc. to bring the Red Hat Enterprise Linux AI (RHEL AI) platform to its popular PowerEdge servers, paving the way for its hardware to serve as the foundation for can serve AI development.

The idea behind this partnership was to make it easier for companies to scale their IT infrastructure to support successful AI and ML strategies without having to host these workloads in the cloud. They can deploy Dell’s PowerEdge servers in their own on-premises data centers or, alternatively, use them as part of a larger hybrid cloud setup.

In addition to its partnership with Dell, Red Hat recently introduced Enterprise Linux AI, aimed primarily at developers. A good example is a bootable RHEL image with pre-configured AI libraries like PyTorch, which allows users to quickly set up an AI-ready environment without having to go through complex installation and configuration processes.

AMD is also very active in supporting AMD GPUs on Linux and recently released the AMD XDNA Linux Driver v3, which is crucial for enabling the Ryzen AI Neural Processing Unit (NPU) on Linux systems. It will likely be integrated into the Linux kernel 6.13.

Linux-centric companies like openSUSE are helping keep AI accessible. For example, openSUSE was listed on Hugging Face and made the first contribution to a dataset called cavil-licence-patterns, which aims to provide more advanced and accurate detection of license issues and compliance.

Red Hat’s contribution

RHEL AI combines several key components to create a powerful foundation for AI innovation. The focus is on the open source Granite models, a family of LLMs developed by IBM Research. These models are complemented by InstructLab, an open source project that makes it easier to experiment and fine-tune models.

This allows domain experts without extensive data science knowledge to contribute to AI models. All of these components are packaged in a bootable Red Hat Enterprise Linux image, streamlining deployment in hybrid cloud environments.

This approach leans more towards ethical AI. Many of Red Hat’s customers couldn’t get anywhere near AI due to copyright issues, but this is essentially the most ethical form AI can take.

Red Hat’s approach addresses several challenges in adopting AI in enterprises. By leveraging open source principles, RHEL AI lowers the barriers to entry for AI innovations and makes them more accessible to a broader range of organizations. The platform offers up to 50% lower costs compared to similar solutions, making AI development more economical for companies.

One Reddit user praised the closed integration with CI/CD, mentioning that you can create and share host images the same way you would container images, and that developers and operators can now use the exact same image in a container or as a bare-bones version. Can run Metal on the host.

“The technology is great, but it’s when it’s integrated into your development, build and deployment pipeline that the magic happens. “It’s not a huge leap forward, just a few modest but extremely useful steps forward,” he added.

Linux is very important for AI developers

Developers advocate using Linux to train models. A developer on Reddit mentioned that compiling applications is much easier with Linux and if he uses Windows for the same task, he has to try to find the VS 2022 launcher, download it and run it, and Google to see which one Options he has to tick. All this coupled with several GB of unnecessary dependencies that he doesn’t want to have anyway.

“Then I have to go to the NVIDIA website, download and install the CUDA Toolkit. Then I need CMAKE, download it and install it. If everything works, compiling will take about 25 minutes. On Arch Linux, just type pacman -S base-devel cuda and you’re good to go. Compiling takes about 5-10 minutes. and the inference is also ~25% faster” he added, suggesting that Linux is efficient not only at compiling but also at inferencing.

Efforts like RHEL AI are important to the Linux community because almost everything related to AI is researched and developed on Linux, ultimately making it a more efficient platform for training and using AI models.

In most cases, you are just one command away from setting up the AI ​​development environment and it works simply by spending hours on Windows to configure the AI ​​development environment.