Add timeouts to test_rocm_wheels.yml

[ScottTodd]

Motivation

Progress on #1559. When adding Python package tests to CI workflows on #3182 we noticed that this job hit a 6 hour timeout on the linux-strix-halo-gpu-rocm-5 runner: https://github.com/ROCm/TheRock/actions/runs/21533668625/job/62060067637?pr=3182.

Technical Details

I chose a 30 minute timeout for the overall job, to catch slow network issues and then a 5 minute timeout for just the test step, to give enough time for the packages to initialize and then catch any hung subprocesses.

Note that we currently skip tests on Linux gfx1151 here:

• TheRock/build_tools/github_actions/amdgpu_family_matrix.py

  Lines 77 to 87 in 55fa089

  	"gfx1151": {
  	"linux": {
  	"test-runs-on": "linux-gfx1151-gpu-rocm",
  	"test-runs-on-kernel": {
  	"oem": "linux-strix-halo-gpu-rocm-oem",
  	},
  	"family": "gfx1151",
  	"bypass_tests_for_releases": True,
  	"build_variants": ["release"],
  	"sanity_check_only_for_family": True,
  	},

• TheRock/build_tools/print_driver_gpu_info.py

  Lines 24 to 26 in 55fa089

  	AMDGPU_FAMILIES = os.getenv("AMDGPU_FAMILIES")
  	# TODO(#2964): Remove gfx950-dcgpu once amdsmi static does not timeout
  	unsupported_amdsmi_families = ["gfx1151", "gfx950-dcgpu"]

We may want to similarly filter Linux gfx1151 from parts of rocm-sdk test.

Test Plan and Results

Triggered a test run with ROCm version 7.12.0a20260202: https://github.com/ROCm/TheRock/actions/runs/21644961000

All tests passed in 1 minute, though the linux-strix-halo-gpu-rocm-7 runner was used for those runs.

Submission Checklist

• Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.