clm_404dd328b4924016a85a1d0a466dd254_r1

IX2-0170

① SA Source

• Source: 開啟完整 SA 文章
• Section: Nvidia versus AMD Disagg Prefill
• Line hint: 282

Context Before

Source: SemiAnalysis InferenceX ↗

Evidence

The MI355X can only match the B200 (without MTP) in performance by using MTP, and only for a range of interactivities from ~60 tok/s/user through ~120 tok/s/user

Context After

Source: SemiAnalysis InferenceX ↗

② Atomic Claim

MI355X 只有在開啟 MTP 後，且約 60～120 tok/s/user 的 interactivity 區間內，才能匹配未開啟 MTP 的 B200。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "attribute": "THROUGHPUT",
  "context_nodes": [
    {
      "id": "04_knowledge_base/Multi-Token Prediction",
      "label": "MTP"
    },
    {
      "id": "04_knowledge_base/NVIDIA B200",
      "label": "B200"
    }
  ],
  "entity": {
    "id": "04_knowledge_base/MI355X",
    "label": "MI355X"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "60",
      "120 tok/s/user"
    ],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [
      "60",
      "120 tok/s/user"
    ],
    "value_text": "MI355X 只有在開啟 MTP 後，且約 60～120 tok/s/user 的 interactivity 區間內，才能匹配未開啟 MTP 的 B200。"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	MI355X	MI355X
context_0	MTP	04_knowledge_base/Multi-Token Prediction
context_1	B200	04_knowledge_base/NVIDIA B200

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。