clm_5707c6c00c25409d94ecdd365f8012b8_r1

VR2-0575

① SA Source

• Source: 開啟完整 SA 文章
• Section: Scale-Up Network – NVLink 6
• Line hint: 687

Context Before

Scale-Up Network – NVLink 6

The bidirectional SerDes technology employed on the copper backplane is used for NVIDIA’s scale-up network. For Vera Rubin NVL72, the scale-up network continues to be rail-optimized with all-to-all connectivity between each GPU and switch ASIC in the scale-up domain.

Evidence

Last year, we discussed NVIDIA’s GB200 scale-up architecture, which consists of 18 NVLink 5 Switch chips in a single rack for the NVL72 system

Context After

GB200 Hardware Architecture - Component Supply Chain & BOM

② Atomic Claim

SemiAnalysis 去年討論過 NVIDIA GB200 scale-up architecture；NVL72 單 rack 中含 18 顆 NVLink 5 Switch chips。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "attribute": "COUNT",
  "context_nodes": [
    {
      "id": "04_knowledge_base/GB200",
      "label": "GB200"
    },
    {
      "id": "04_knowledge_base/NVLink5",
      "label": "NVLink 5"
    }
  ],
  "entity": {
    "id": "02_companies/NVDA",
    "label": "NVIDIA"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "18",
      "5 S"
    ],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [
      "18",
      "5 S"
    ],
    "value_text": "SemiAnalysis 去年討論過 NVIDIA GB200 scale-up architecture；NVL72 單 rack 中含 18 顆 NVLink 5 Switch chips。"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	NVIDIA	NVDA
context_0	GB200	GB200
context_1	NVLink 5	NVLink5

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。