SA260324R-0105_r1

SA260324R-0105

① SA Source

• Source: 開啟完整 SA 文章
• Section: GPU and LPU Integration: Attention FFN Disaggregation (AFD)
• Line hint: 111

Context Before

Attention and FFN are subsets of operations in a model. In a model forward pass, attention’s output feeds into a token router, and the token router assigns each token to k experts, where each expert is an FFN. Attention and FFN have very different performance properties. During decode phase, the GPU utilization of attention barely improves when scaling batch size due to being bounded by loading KV cache.

Evidence

In contrast, the GPU utilization of FFN scales with batch size comparatively better.

Context After

This is something we have worked with certain hardware vendors and memory companies on with our inference simulator for more than 6 months.

② Atomic Claim

相較之下，GPU 執行 FFN 時的 utilization，隨 batch size 增加而改善的幅度相對更好。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "attribute": "FFN_UTILIZATION_SCALING",
  "context_nodes": [
    {
      "id": "04_knowledge_base/Feed-forward Network",
      "label": "FFN"
    },
    {
      "id": "04_knowledge_base/Batch size",
      "label": "batch size"
    }
  ],
  "entity": {
    "id": "04_knowledge_base/GPU",
    "label": "GPU"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [],
    "value_text": "執行 FFN 時，utilization 隨 batch size 增加而改善的幅度相對更好。"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	GPU	GPU
context_node	FFN	04_knowledge_base/Feed-forward Network
context_node	batch size	04_knowledge_base/Batch size

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。