clm_4f7ae26a109641d98b128c2d41c84cd7_r1

VR2-0548

① SA Source

• Source: 開啟完整 SA 文章
• Section: Networking: NVLink 6, Rubin Scale Out
• Line hint: 637

Context Before

Networking: NVLink 6, Rubin Scale Out

NVIDIA’s gen-on-gen innovations are evolutionary and not revolutionary. With GPU scale-up and scale-out bandwidth doubling approximately every 18 months, the copper infrastructure in NVIDIA racks are being innovated to accommodate higher bandwidth workloads. Scale-up network infrastructure will eventually involve optics to build larger world sizes, but that is the topic of a separate article.

Evidence

NVLink 6 used in Vera Rubin doubles NVLink bandwidth by implementing bi-directional signaling over the same number of copper cables - effectively delivering 4 Lanes of 200G per NVLink

Context After

Source: Nvidia VR NVL72 Component BoM and Power Budget Model ↗

② Atomic Claim

Vera Rubin 使用的 NVLink 6，在相同數量的 copper cables 上採 bi-directional signaling，使 NVLink bandwidth 加倍，等效為每個 NVLink 提供 4 lanes × 200G。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "comparison_expression": "Vera Rubin 使用的 NVLink 6，在相同數量的 copper cables 上採 bi-directional signaling，使 NVLink bandwidth 加倍，等效為每個 NVLink 提供 4 lanes × 200G。",
  "entities": [
    {
      "id": "04_knowledge_base/Vera Rubin",
      "label": "Vera Rubin"
    },
    {
      "id": "04_knowledge_base/NVLink",
      "label": "NVLink"
    },
    {
      "id": "04_knowledge_base/Copper",
      "label": "copper"
    },
    {
      "id": "04_knowledge_base/200G",
      "label": "200G"
    }
  ],
  "frame_type": "COMPARISON",
  "metric": "BANDWIDTH",
  "operator": "MULTIPLE_OF",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "6",
      "4",
      "200"
    ],
    "temporal_mentions": []
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
comparison_entity_0	Vera Rubin	04_knowledge_base/Vera Rubin
comparison_entity_1	NVLink	NVLink
comparison_entity_2	copper	Copper
comparison_entity_3	200G	200G

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。