clm_ddb8678cb11c4dcba17ca715eb075a8b_r1

VR2-0437

① SA Source

• Source: 開啟完整 SA 文章
• Section: Nvidia's Christmas Present: GB300 & B300 - Reasoning Inference, Amazon, Memory, Supply Chain
• Line hint: 535

Context Before

2024年12月25日

Read full story ↗

Evidence

In GPU demanding loads, 2300W goes to each GPU leaving 200W for the CPU

Context After

For the modules at the front of chassis - CX-9,BlueField-4, and management - the PDB feeds 12VDC of power to each of the modules. 50VDC l down to 12VDC at the PDB then, via a coper busbar device, the PDB feeds 12VDC to its neighboring modules. The power connector of CX-9 is located near the Paladin HD2 at the top of the modules.

Compute Tray: Mechanical

② Atomic Claim

在高 GPU 負載時，每顆 GPU 分配 2,300W，剩餘 200W 給 CPU。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "attribute": "COUNT",
  "context_nodes": [
    {
      "id": "04_knowledge_base/CPU",
      "label": "CPU"
    }
  ],
  "entity": {
    "id": "04_knowledge_base/GPU",
    "label": "GPU"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "2,300W",
      "200W"
    ],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [
      "2,300W",
      "200W"
    ],
    "value_text": "在高 GPU 負載時，每顆 GPU 分配 2,300W，剩餘 200W 給 CPU。"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	GPU	GPU
context_0	CPU	CPU

⑤ Human Review

請在 Properties 逐項確認：

• 原文 → Atomic Claim 是否忠實
• Atomic Claim → Semantic Frame 是否忠實
• Canonical Entity mapping 是否正確
• Epistemic mode 是否保留原文語氣
• 最後選擇 review_action

Review state

Markdown 內文不是正式 approval。只有 Apply bridge 寫入的 Decision Ledger event 才是正式決策。