[V1] Enable V1 for compute capability < 8.0 + FP32

[DarkLight1337]

Purpose

Since we now support FlexAttention: https://github.com/vllm-project/vllm/blob/main/vllm/platforms/cuda.py#L332

V1 Engine should be allowed for older devices as well.

FIX #23531

Test Plan

Test Result

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
• (Optional) Release notes update. If your change is user facing, please update the release notes draft in the Google Doc.

[gemini-code-assist]

Code Review

This pull request enables the V1 engine for GPUs with a compute capability of less than 8.0. This is achieved by removing a check that previously restricted V1 to newer GPUs. The change is justified by the integration of the FlexAttention backend, which is designed to support these older architectures within the V1 engine. The modification is localized and appears consistent with the existing attention backend selection logic. I have not identified any issues with this change.

[mgoin]

Seems reasonable to me, but can we test this?

[DarkLight1337]

I don't have access to such a device, so I'll wait for the OP in #23531 to comment on this

[Isotr0py]

I have tested this on a T4 machine earlier, can confirm flex attention working with fp32.