Remove timing-sensitive NVML total-energy test

[rwgk]

Description

Addresses cuda-python-private issue 509.

Removes the timing-sensitive test_device_get_total_energy_consumption bootstrap test. NVML does not guarantee a refresh cadence, so requiring the cumulative energy counter to advance within 150 ms can fail even when the binding is working correctly.

Checklist

• New or existing tests cover these changes. (Not applicable: this removes a timing-sensitive bootstrap test without changing product code.)
• The documentation is up to date with these changes. (No documentation changes are required.)

[copy-pr-bot]

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

[rwgk]

I also looked into whether removing this test should leave behind a comment such as # [Skipping] pynvml.nvmlDeviceGetTotalEnergyConsumption.

I found that the existing [Skipping] comments in this file were copied verbatim from the upstream pynvml 11.5.3 test when the file was introduced; they are not maintained as a ledger of tests later removed. As direct precedent, #2483 removed test_device_get_nvlink_capability from this same file without adding a corresponding [Skipping] comment. I therefore left the removal as-is.

[github-actions]

Doc Preview CI
🚀 View preview at https://nvidia.github.io/cuda-python/pr-preview/pr-2644/
https://nvidia.github.io/cuda-python/pr-preview/pr-2644/cuda-core/
https://nvidia.github.io/cuda-python/pr-preview/pr-2644/cuda-bindings/
https://nvidia.github.io/cuda-python/pr-preview/pr-2644/cuda-pathfinder/
Preview will be ready when the GitHub Pages deployment is complete.