clm_a42af73305ba41b08640b4011d9cb1e4_r1

IX2-0129

① SA Source

• Source: 開啟完整 SA 文章
• Section: Disaggregated Inference Frameworks
• Line hint: 200

Context Before

Source: SemiAnalysis InferenceX ↗

Disaggregated Inference Frameworks

Evidence

For AMD, we use SGLang with two different KV cache transfer frameworks: MoRI and Mooncake

Context After

DeepSeek Disagg +WideEP Results Deep Dive

At almost all interactivity levels, disagg outperform aggregated inference (grey lines) in terms of total token throughput per GPU. Multi-node disaggregrated prefill framemogs single node aggregrated serving.

② Atomic Claim

對 AMD，InferenceX 使用 SGLang 搭配兩種不同的 KV cache transfer frameworks：MoRI 與 Mooncake。

• Epistemic Mode: ASSERTED
• Mapping Status: PARTIAL

③ Semantic Frame

{
  "frame_type": "NARY_RELATION",
  "participants": [
    {
      "node": {
        "id": "02_companies/AMD",
        "label": "AMD"
      },
      "role": "user_or_subject"
    },
    {
      "node": {
        "id": "04_knowledge_base/SGLang",
        "label": "SGLang"
      },
      "role": "used_entity"
    },
    {
      "node": {
        "id": "04_knowledge_base/KV cache",
        "label": "KV cache"
      },
      "role": "used_entity"
    },
    {
      "node": {
        "id": "04_knowledge_base/AMD MoRI backend",
        "label": "MoRI"
      },
      "role": "used_entity"
    },
    {
      "node": {
        "id": "04_knowledge_base/Mooncake",
        "label": "Mooncake"
      },
      "role": "used_entity"
    }
  ],
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "relation_type": "USES"
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
user_or_subject	AMD	AMD
used_entity	SGLang	SGLang
used_entity	KV cache	04_knowledge_base/KV cache
used_entity	MoRI	04_knowledge_base/AMD MoRI backend
used_entity	Mooncake	Mooncake

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。