✨[Feature] Key configuration quick feedback workflows for PRs

[narendasan]

Is your feature request related to a problem? Please describe.

We need faster turnaround time on PRs being validated in CI. Right now builds take 20mins + 40min - 1.5hrs for tests to run assuming full runner availability

Describe the solution you'd like

We want a subset of tests that will return results in less than one 1hr that will check the key configurations of Torch-TensorRT

• Python 3.12 (or latest Ubuntu LTS py version) + latest cuda (cu128) + x86-64 linux only
  • 1 job that takes 20mins
  • sub 30m allocated for min-spanning test set (i.e. a set of tests that will exercise critical parts of the library)

We can also use this workflow template to test edge case configs i.e python 3.13 or cuda 13

Note this set is not sufficient to merge a PR, just to give devs feedback quick. All tests in Build-Test [Linux/Windows] must pass for a PR to merge.

Describe alternatives you've considered

Additional context

[narendasan]

@lanluo-nvidia can you determine a min test set? https://coverage.readthedocs.io/en/7.8.0/ might help