clm_7626caf939714ae2b8f277c51e9e7a83_r1

IX2-0151

① SA Source

• Source: 開啟完整 SA 文章
• Section: Nvidia TensorRT LLM and NVL72
• Line hint: 250

Context Before

Source: SemiAnalysis InferenceX ↗

Evidence

At low batch sizes, the inference workload shrinks enough to fit within a single HGX node’s NVLink domain (i.e. 8 GPUs)

Context After

Source: SemiAnalysis InferenceX ↗

② Atomic Claim

在低 batch size 下，推論 workload 會縮小到足以放進單一 HGX node 的 NVLink domain，也就是 8 顆 GPUs。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "attribute": "COUNT",
  "context_nodes": [
    {
      "id": "04_knowledge_base/NVLink",
      "label": "NVLink"
    },
    {
      "id": "04_knowledge_base/GPU",
      "label": "GPUs"
    }
  ],
  "entity": {
    "id": "04_knowledge_base/HGX",
    "label": "HGX"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "8"
    ],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [
      "8"
    ],
    "value_text": "在低 batch size 下，推論 workload 會縮小到足以放進單一 HGX node 的 NVLink domain，也就是 8 顆 GPUs。"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	HGX	HGX
context_0	NVLink	NVLink
context_1	GPUs	GPU

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。