[Power] Bulk-enable GB multinode telemetry / 批量启用 GB 多节点功耗采集

[edwingao28]

What changed

• Evaluate 61 GB200/GB300 multinode recipes against the existing power-lane eligibility contract.
• Enable dcgm-power on 59 eligible recipes: retain required: true for 8 previously hardware-validated shapes and start 51 additional recipes with required: false.
• Leave 2 dedicated-infrastructure recipes disabled because their current topology is not eligible for this launcher path.
• Add dynamic contract coverage and CI path filters so every candidate and eligible recipe remains covered when master config mappings change.
• Append the required performance changelog entry without changing runner behavior.

Why

The producer and dashboard paths already support GB telemetry, but most eligible recipes were not opted in. Starting new shapes as optional exposes telemetry without turning an unverified power failure into a benchmark failure.

Validation

• CI-equivalent power and result-processing suite: 205 passed.
• Official performance changelog validator passed against origin/main.
• Contract coverage includes 61/61 candidates and 59/59 eligible recipes.
• git diff --check passed.

Hardware status

No GB200, GB300, B200, or B300 job was dispatched for this PR. The current validation environment is limited to H100/H200. The 51 new recipes therefore remain required: false; hardware artifact and dashboard verification is pending on an authorized GB environment.

中文说明

变更内容

• 按现有功耗 lane 资格契约检查 61 个 GB200/GB300 多节点 recipe。
• 为 59 个符合条件的 recipe 启用 dcgm-power：8 个已有真机验证的拓扑继续使用 required: true，另外 51 个新增 recipe 从 required: false 起步。
• 2 个使用独立基础设施节点且不符合当前 launcher 资格条件的 recipe 保持关闭。
• 增加动态 contract 覆盖与 CI path 过滤，确保 master config 映射变化后，所有候选及 eligible recipe 仍会被测试覆盖。
• 追加必需的 performance changelog，不修改 runner 行为。

原因

producer 和 dashboard 链路已经支持 GB 功耗数据，但大部分符合条件的 recipe 尚未显式开启。新增拓扑先采用 optional 模式，可以暴露遥测数据，同时避免尚未验证的功耗失败导致基准测试失败。

验证

• CI 同款功耗与结果处理测试：205 个通过。
• 官方 performance changelog validator 基于 origin/main 验证通过。
• contract 覆盖 61/61 个候选及 59/59 个 eligible recipe。
• git diff --check 通过。

硬件状态

本 PR 未派发任何 GB200、GB300、B200 或 B300 作业；当前可用验证硬件仅限 H100/H200。因此 51 个新增 recipe 保持 required: false，等待未来在获准的 GB 环境中完成产物与 dashboard 验证。