cuda: Verify Cuda Toolkit is Supported by the NVIDIA Architecture

[Treece-Burgess]

Pull Request Description

This PR adds the functionality to check a users Cuda Toolkit to verify that it can be used with the NVIDIA architecture on the machine and provide a helpful disabled message. This is important as:

1. Older Cuda Toolkits do not support the newer architectures
2. Newer Cuda Toolkits do not support the older architectures

Testing

Setup

Testing was done on Illyad and Voltar at Oregon.

Illyad:
OS: RHEL 8.10
CPU: AMD EPYC 7402
GPU: 1 * H100
Cuda Toolkit: 11.5.2 and 12.9.0

Voltar:
OS: RHEL 8.10
CPU: Intel Xeon Gold 6226R
GPU: 1 * A100, 1 * V100, and 1 * P100
Cuda Toolkit: 13.0.0

Results

Illyad:

• With Cuda Toolkit 11.5.2 the cuda component is disabled and this is the correct behavior as the H100 does not support this Cuda Toolkit version
• With Cuda Toolkit 12.9.0:
  • PAPI Utilities*: ✅
  • Component tests: ✅

Voltar:
The V100 and P100 cause the cuda component to be disabled with Cuda Toolkit 13.0, but setting CUDA_VISIBLE_DEVICES=1,2 will allow for it to be active as this would only "show" the A100.

• PAPI Utilities*: ✅

* - papi_component_avail, papi_native_avail, and papi_command_line

Author Checklist

• Description
  Why this PR exists. Reference all relevant information, including background, issues, test failures, etc
• Commits
  Commits are self contained and only do one thing
  Commits have a header of the form: module: short description
  Commits have a body (whenever relevant) containing a detailed description of the addressed problem and its solution
• Tests
  The PR needs to pass all the tests