Add GPU testing workflows and add gpu tox configuration

[fabianliebig]

This PR adds a dedicated Tox environment for GPU-related tests and one manual workflow as well as additional steps to regular and CI for installing and starting those tests. At the moment, they are only executed if triggered manually and if core tests have passed. Support for other triggers is WIP. Additionally, I've changed the Benchmark installation to UV and added terminal outputs of standard resource commands for CPU, RAM and so on.

[AVHopp]

First round of comments.

[Copilot]

Pull Request Overview

This PR adds GPU testing capabilities to the project by introducing a dedicated Tox environment for GPU tests and workflow configurations. The changes include manual workflow triggers for GPU testing, system information outputs for benchmarking, and modernization of the benchmark installation process.

• Adds a new gputest Tox environment for GPU-specific testing
• Introduces manual GPU testing workflows that require core tests to pass first
• Modernizes benchmark installation by switching from pip to uv package manager

Reviewed Changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tox.ini	Adds gputest environment configuration with GPU availability check
.github/workflows/regular.yml	Adds GPU testing job triggered by manual workflow dispatch
.github/workflows/gpu_tests.yml	New dedicated workflow for GPU tests with AWS Lambda runner provisioning
.github/workflows/ci.yml	Adds GPU testing job to CI workflow with manual trigger
.github/workflows/benchmark.yml	Adds system information output and migrates to uv package manager

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[AVHopp]

LGTM, only thing to do is to remove the single precision env variables, but there is already a comment for that.

[AdrianSosic]

Thanks, @fabianliebig, for taking the initiative here. I've gone through the changes and comments by others and there's no obvious reason to block anything from my side. However, it's one of the typical fabian knows this stuff so much better than me PRs that I'm more or less trusting you blindly here and will only potentially complain once I actually start using it and encounter issues, i.e. once our GPU code is ready 😄

[fabianliebig]

Thanks, @fabianliebig, for taking the initiative here. I've gone through the changes and comments by others and there's no obvious reason to block anything from my side. However, it's one of the typical fabian knows this stuff so much better than me PRs that I'm more or less trusting you blindly here and will only potentially complain once I actually start using it and encounter issues, i.e. once our GPU code is ready 😄

Thanks for the review and trust :). Please don't hesitate to reach out anytime if you run into problems or even just have questions. I'm happy to help troubleshoot. Working with GPU can be quite messy and there are nearly always issues and synergies that weren't foreseeable :D

I think if the pipeline code and TOX environment work for you, and since Torch can already detect the GPU, that's a good starting point to explore GPU usage :) Feel free to merge once you think it fits and as said ping me if you need me :D