VR2R-0173_r1

VR2R-0173

① SA Source

• Source: 開啟完整 SA 文章
• Section: CPX Form Factor
• Line hint: 143

Context Before

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.

Evidence

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

更重要的是，這正式形成 inference prefill（compute-bound）與 decode（bandwidth-bound）之間的 architectural disaggregation。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "comparison_expression": "prefill is compute-bound while decode is bandwidth-bound, enabling architectural disaggregation.",
  "entities": [
    {
      "id": "04_knowledge_base/Prefill",
      "label": "prefill"
    },
    {
      "id": "04_knowledge_base/Decode",
      "label": "decode"
    }
  ],
  "frame_type": "COMPARISON",
  "metric": "INFERENCE_RESOURCE_BOUND_CHARACTERISTIC",
  "operator": "DIFFERS_FROM",
  "qualifiers": {
    "condition_text": "dual-rack direction formalizes disaggregation",
    "numeric_mentions": [],
    "temporal_mentions": []
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	prefill	Prefill
entity	decode	Decode

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。