[Bugfix] Do not use resource module on Windows (#12858)

[MoonRide303]

Python resource module is supported only on *nix OSes - it should never be imported on Windows.

FIX #12858

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[mgoin]

LGTM, FYI @robertgshaw2-redhat

[mgoin]

I'm curious - have you actually tried to get vllm to run on windows? I have only used it on WSL

[MoonRide303]

I'm curious - have you actually tried to get vllm to run on windows? I have only used it on WSL

Yup - other apps work just fine, so I see no real reason why vLLM shouldn't :). This was just first obstacle and not the only problem, though - but one thing at a time.

As of WSL - it's fine for running some *nix-only tools, but it's not really that good for more demanding apps. CUDA via WSL worked much worse for me than running it natively on Windows.

[levitation]

Regarding the motivation why resource module was originally imported - you can use memory limits under Windows as well, but you need a separate implementation for that. One implementation example is available here:
https://github.com/levitation-opensource/python_memory_limit/blob/main/windows_jobobject.py