clm_3acb9375ea3a47499adec5589adc64d7_r1

VR2R-0171::RESEG01

回到這篇 SA Coverage

① SA Source

• Source: 開啟完整 SA 文章
• Section: CPX Form Factor
• Line hint: 229

Context Before

VR NVL72 CPX (Dual Rack): Rubin CPX deployed in a separate rack alongside the VR NVL72 rack

Evidence

The standalone/dedicated rack direction materially changes the deployment calculus. A dual-rack approach allows hyperscalers to scale prefill and decode capacity independently, optimize datacenter power envelopes, and reduce system-level failure domains versus tightly coupled trays. More importantly, it formalizes architectural disaggregation between inference prefill (compute-bound) and decode (bandwidth-bound).

Context After

Rubin CPX was originally architected as a GDDR7-based accelerator optimized for prefill, based on three key considerations:

② Atomic Claim

Dual-rack approach 讓 hyperscalers 可以獨立擴展 prefill 與 decode capacity。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "frame_type": "NARY_RELATION",
  "participants": [
    {
      "node": {
        "id": "04_knowledge_base/Hyperscaler",
        "label": "hyperscalers"
      },
      "role": "operator"
    },
    {
      "node": {
        "id": "04_knowledge_base/Prefill",
        "label": "prefill"
      },
      "role": "independently_scaled_workload"
    },
    {
      "node": {
        "id": "04_knowledge_base/Decode",
        "label": "decode"
      },
      "role": "independently_scaled_workload"
    }
  ],
  "qualifiers": {
    "condition_text": "dual-rack approach",
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "relation_type": "DUAL_RACK_INDEPENDENT_PREFILL_DECODE_SCALING"
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
operator	hyperscalers	Hyperscaler
independently_scaled_workload	prefill	Prefill
independently_scaled_workload	decode	Decode

⑤ V2 Draft（需要修正時直接編輯）

Inline repair

只編輯以下三個 Draft 區塊。完成後將 review_action 設為 REQUEST_REPAIR，再手動執行 Repair Preview。
執行腳本前，Draft 不會寫入 Claim Store 或 Graph。

V2 Atomic Claim Draft

Dual-rack approach 讓 hyperscalers 可以獨立擴展 prefill 與 decode capacity。

V2 Semantic Frame Draft

{
  "frame_type": "NARY_RELATION",
  "participants": [
    {
      "node": {
        "id": "04_knowledge_base/Hyperscaler",
        "label": "hyperscalers"
      },
      "role": "operator"
    },
    {
      "node": {
        "id": "04_knowledge_base/Prefill",
        "label": "prefill"
      },
      "role": "independently_scaled_workload"
    },
    {
      "node": {
        "id": "04_knowledge_base/Decode",
        "label": "decode"
      },
      "role": "independently_scaled_workload"
    }
  ],
  "qualifiers": {
    "condition_text": "dual-rack approach",
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "relation_type": "DUAL_RACK_INDEPENDENT_PREFILL_DECODE_SCALING"
}

V2 Revision Metadata Draft

{
  "epistemic_mode": "ASSERTED",
  "mapping_status": "COMPLETE"
}

⑥ 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。