AMDGPU.jl VS distrobox

For myself, I use Julia to write my own software (that is run on AMD supercomputer) on Fedora system, using 6800XT. For my experience, everything worked nicely. To install you need to install rocm-opencl package with dnf, AMD Julia package (AMDGPU.jl), add yourself to video group and you are good to go. Also, Julia's KernelAbstractions.jl is a good to have, when writing portable code.

The Julia package AMDGPU.jl provides a Julia interface for AMD GPU (ROCm) programming. As they say, the package is being developed for Julia 1.7, 1.9 and above, but not 1.8. Therefore I downloaded the Julia binary of version 1.7.3 from the older releases Julia page.

This is exciting news! What's also exciting is that it's not just C++ that can run on this supercomputer; there is also good (currently unofficial) support for programming those GPUs from Julia, via the AMDGPU.jl library (note: I am the author/maintainer of this library). Some of our users have been able to run AMDGPU.jl's testsuite on the Crusher test system (which is an attached testing system with the same hardware configuration as Frontier), as well as their own domain-specific programs that use AMDGPU.jl.

What's nice about programming GPUs in Julia is that you can write code once and execute it on multiple kinds of GPUs, with excellent performance. The KernelAbstractions.jl library makes this possible for compute kernels by acting as a frontend to AMDGPU.jl, CUDA.jl, and soon Metal.jl and oneAPI.jl, allowing a single piece of code to be portable to AMD, NVIDIA, Intel, and Apple GPUs, and also CPUs. Similarly, the GPUArrays.jl library allows the same behavior for idiomatic array operations, and will automatically dispatch calls to BLAS, FFT, RNG, linear solver, and DNN vendor-provided libraries when appropriate.

I'm personally looking forward to helping researchers get their Julia code up and running on Frontier so that we can push scientific computing to the max!

Library link: <https://github.com/JuliaGPU/AMDGPU.jl>

GitHub - JuliaGPU/AMDGPU.jl: AMD GPU (ROCm) programming in Julia

https://github.com/JuliaGPU/AMDGPU.jl

https://github.com/JuliaGPU/oneAPI.jl

These are both less mature than CUDA.jl, but are in active development.

Regarding the stability issue on a dev machine - you may be interested in playing with one of the immutable-os distros, such as SilverBlue (fedora based).

The high-level take-away is you can't break your actual OS since it's root filesystem is read-only, and you use "pet" containers (on docker, podman, whatever) to do your work in. Applications are either sandboxed via Flatpak, or installed/run inside your pet containers. If your pet container dies, you cry about it for a moment, and when you're ready you get a new one - your actual os and other containers remain unaffected.

I use distrobox[1] to create/run the pet containers.

[1] https://github.com/89luca89/distrobox

Distrobox is a tool that enables us to try Linux distro CLI, including their package manager. This requires a containerization tool (e.g., Docker). In Windows, this can be achieved using WSL (Windows Subsystem for Linux)

I use containerized versions of things, ubuntu and chainguard images mostly.

You can always create containers with init if that's how you want to do that though. Some distros publish images that come that way: https://github.com/89luca89/distrobox/blob/main/docs/useful_...

https://news.ycombinator.com/item?id=38505448 ... https://github.com/89luca89/distrobox/blob/main/docs/useful_...

Yes, you can do that but I've seen others use something like distrobox to run linux inside of SteamOS: https://github.com/89luca89/distrobox/blob/main/docs/posts/steamdeck_guide.md

For such cases you might use something like this: https://github.com/89luca89/distrobox

Btw #3: Depending on what the user is trying to accomplish, e.g. maybe to make WSL(2) itself more of a "subsystem" than a "container engine", using something like Distrobox or nsbox.dev can be a good idea (along with Docker or Podman in Distrobox's case; the other one uses systemd-nspawn).

1. Find here in "Containers Distros" section the distro image that you want to install ("Toolbox" versions are better because they are configured for Distrobox) and get it URL: https://distrobox.it/compatibility/#containers-distros 2. Use that URL to create Distrobox: distrobox create -i registry.fedoraproject.org/fedora-toolbox:39 -n fedora_1_39 3. Enter Distrobox fedora_1_39: distrobox enter fedora_1_39 4. You are already in Distrobox console. Look at the name in console, it should be include the container name. 5. To exit Distrobox: exit 6. If you run: distrobox list you will see all distroboxes on the system. You will also see that distrobox that we exited is still running. 7. To stop distrobox use commands: distrobox stop fedora_1_39