clm_39bd614b9b3b474d888b91fffe71bd81_r1

VR2-0046

① SA Source

• Source: 開啟完整 SA 文章
• Section: Extreme Co-design: 6 Silicon Product Offerings under the Rubin Platform - Chip Floorplans & Specifications
• Line hint: 85

Context Before

Source: SemiAnalysis, Nvidia

Evidence

Rubin’s dense FP4 and FP8 FLOPs increase by roughly ~3.5× versus GB200, while FP16 FLOPs rise by a more modest ~1.6x, underscoring NVIDIA’s continued emphasis on FP4/FP8 as the primary scaling vector

Context After

Source: SemiAnalysis, Nvidia

② Atomic Claim

相較 GB200，Rubin 的 dense FP4 與 FP8 FLOPs 約提高 3.5 倍，而 FP16 FLOPs 僅約提高 1.6 倍，顯示 NVIDIA 持續把 FP4／FP8 視為主要 scaling vector。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "comparison_expression": "相較 GB200，Rubin 的 dense FP4 與 FP8 FLOPs 約提高 3.5 倍，而 FP16 FLOPs 僅約提高 1.6 倍，顯示 NVIDIA 持續把 FP4／FP8 視為主要 scaling vector。",
  "entities": [
    {
      "id": "04_knowledge_base/GB200",
      "label": "GB200"
    },
    {
      "id": "04_knowledge_base/Vera Rubin",
      "label": "Rubin"
    },
    {
      "id": "04_knowledge_base/FP4",
      "label": "FP4"
    },
    {
      "id": "04_knowledge_base/FP8",
      "label": "FP8"
    },
    {
      "id": "04_knowledge_base/FP16",
      "label": "FP16"
    },
    {
      "id": "02_companies/NVDA",
      "label": "NVIDIA"
    }
  ],
  "frame_type": "COMPARISON",
  "metric": "COMPUTE_PERFORMANCE",
  "operator": "MULTIPLE_OF",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "3.5",
      "1.6"
    ],
    "temporal_mentions": []
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
comparison_entity_0	GB200	GB200
comparison_entity_1	Rubin	04_knowledge_base/Vera Rubin
comparison_entity_2	FP4	FP4
comparison_entity_3	FP8	FP8
comparison_entity_4	FP16	FP16
comparison_entity_5	NVIDIA	NVDA

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。