VR2R-0592_r1

VR2R-0592

① SA Source

• Source: 開啟完整 SA 文章
• Section: Connecting GPUs in the Scale-Out InfiniBand Network
• Line hint: 492

Context Before

For the Vera Rubin NVL72 deployment however, there are two 800G OSFP cages per GPU for a total of 1.6T per GPU bandwidth.

Evidence

Having two 800G logical ports per GPU is advantageous because it allows multi-plane network deployments without complex fiber management—by splitting one logical GPU two ways to two different leaf switches.

Context After

As such, larger network clusters can be built with two 800G logical ports than if only one 1.6T logical port were used.

② Atomic Claim

每顆 GPU 有 2 個 800G logical ports 的優點，是能把一顆 logical GPU 分到兩個不同 leaf switches，支援 multi-plane deployment，而不需要複雜的 fiber management。

• Epistemic Mode: ASSERTED
• Mapping Status: PARTIAL

③ Semantic Frame

{
  "attribute": "DUAL_800G_PORT_MULTIPLANE_ADVANTAGE",
  "context_nodes": [],
  "entity": {
    "id": "04_knowledge_base/GPU",
    "label": "GPU"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": "leaf switches and fiber management are plain text",
    "numeric_mentions": [
      "2",
      "800G"
    ],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [
      "2",
      "800G"
    ],
    "value_text": "two 800G logical ports allow one logical GPU to split across two leaf switches for multi-plane deployment without complex fiber management."
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	GPU	GPU

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。