SA260324R-0115_r1

SA260324R-0115

① SA Source

• Source: 開啟完整 SA 文章
• Section: GPU and LPU Integration: Attention FFN Disaggregation (AFD)
• Line hint: 125

Context Before

Comparing the two operations, we see attention is stateful due to dynamic KV cache loading patterns, whereas FFN is stateless since the computation only depends on the token inputs. Thus, we disaggregate the computation of attention and FFN. We map attention computations to GPUs, which handle dynamic workloads well.

Evidence

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

Context After

With AFD, token routing from GPUs to LPUs can become the bottleneck, especially under strict latency constraints. The token routing flow involves two operations: dispatch and combine. In the dispatch step, we route each token to their top k experts with an All-to-All collective operation. After experts complete their computation, we perform the combine step, where the outputs are sent back to the source location with a reverse All-to-All collective, continuing the next layer’s computation.

② Atomic Claim

FFN 會被分配到 LPU，因為 LPU 架構本質上是 deterministic，並能受益於 static compute workload。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "additional_nodes": [],
  "frame_type": "RELATION",
  "object": {
    "id": "04_knowledge_base/Groq LPU",
    "label": "LPU"
  },
  "predicate": "MAPS_TO",
  "qualifiers": {
    "condition_text": "LPU 架構本質上 deterministic，並受益於 static compute workload",
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "subject": {
    "id": "04_knowledge_base/Feed-forward Network",
    "label": "FFN"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
subject	FFN	04_knowledge_base/Feed-forward Network
object	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。