clm_8a771a5c565a4daf9eb8d2091853e7f9_r1

NIEK2-0117

① SA Source

• Source: 開啟完整 SA 文章
• Section: GPU and LPU Integration: Attention FFN Disaggregation (AFD)
• Line hint: 107

Context Before

Source: SemiAnalysis

Evidence

For FFNs, we map them to LPUs, since LPU architecture is inherently deterministic and benefits from static compute workloads

Context After

Source: SemiAnalysis, MegaScale-Infer

② Atomic Claim

FFN 則映射到 LPU，因為 LPU 架構本身具有確定性，較適合靜態運算工作負載。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "attribute": "DESIGN_PROPERTY",
  "context_nodes": [
    {
      "id": "04_knowledge_base/Groq LPU",
      "label": "LPU"
    }
  ],
  "entity": {
    "id": "04_knowledge_base/Feed-forward Network",
    "label": "FFN"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [],
    "value_text": "FFN 則映射到 LPU，因為 LPU 架構本身具有確定性，較適合靜態運算工作負載。"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	FFN	04_knowledge_base/Feed-forward Network
context_0	LPU	04_knowledge_base/Groq LPU

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。