cuda: Create a component test to enumerate all base native events and add them to an eventset

[Treece-Burgess]

Pull Request Description

This new cuda component test will enumerate all events on the machine and attempt to add them to a PAPI Eventset one by one. For each added event, a PAPI_start and PAPI_stop call follows. Output is as follows:

Base Event#: 2875
PAPI Eventcode: 0x4000df5e
PAPI Eventname: cuda:::smsp__warp_issue_stalled_misc_per_warp_active
PAPI Longdescr: proportion of warps per cycle, waiting on a miscellaneous hardware reason. Units=(unitless). Numpass=1.
PAPI Countervalue: 1

Base Event#: 2876
PAPI Eventcode: 0x4000df60
PAPI Eventname: cuda:::smsp__warp_issue_stalled_no_instruction_per_warp_active
PAPI Longdescr: proportion of warps per cycle, waiting to be selected for instruction fetch, or waiting on an instruction cache miss. Units=(unitless). Numpass=1.
PAPI Countervalue: 1

Note: This test can take up to 12 minutes of wall clock time to complete; therefore, it will not be added to the run_cuda_tests.sh script.

Testing

Testing was done on Voltar at Oregon (1 * A100, 1 * V100) with Cuda Toolkit 12.9.

• cuda_add_all_events.cu: ✅

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

[Treece-Burgess]

This PR should not be reviewed/merged until PR #597 and PR #599 has been reviewed/merged.