SA260324R-0106_r1

SA260324R-0106

① SA Source

• Source: 開啟完整 SA 文章
• Section: GPU and LPU Integration: Attention FFN Disaggregation (AFD)
• Line hint: 113

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

Evidence

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

Context After

As state-of-the-art mixture-of-expert (MoE) models grow increasingly sparse, tokens can choose experts from a larger expert pool. As a result, each expert receives fewer tokens, leading to lower utilization. This motivates attention and FFN disaggregation. If a GPU only performs attention operations, its HBM capacity can be fully allocated to KV cache, increasing the total number of tokens it can process, which then increases the tokens each expert processes on average.

② Atomic Claim

SemiAnalysis 表示，他們已透過 inference simulator 與部分 hardware vendor 及 memory company 共同研究這個問題超過 6 個月。

• Epistemic Mode: ASSERTED
• Mapping Status: PARTIAL

③ Semantic Frame

{
  "frame_type": "CLAIM_ONLY",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "more than 6 months"
    ],
    "temporal_mentions": [
      "more than 6 months"
    ]
  },
  "reason": "META_RESEARCH_PROCESS_WITH_UNMAPPED_PARTIES",
  "statement": "SemiAnalysis 表示，他們已透過 inference simulator 與部分 hardware vendor 及 memory company 共同研究這個問題超過 6 個月。"
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
—	—	—

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。