VR2R-0721_r1

VR2R-0721

① SA Source

• Source: 開啟完整 SA 文章
• Section: VR NVL72 TCO: BoM and Power Budget Analysis
• Line hint: 631

Context Before

The MI4XX currently has a performance per TCO advantage based on marketed dense FLOPS over VR.

Evidence

VR has a higher TCO per GPU compared to MI4XX, yet VR marketed PFLOPs are lower vs MI4XX resulting in a TCO per PFLOP disadvantage for the VR vs MI4XX.

Context After

NVIDIA is offering sparsity for FP4, marketing 50 PFLOPS of FP4, while AMD has opted to remove sparsity support since CDNA4 for inference dtypes. Harnessing the 50 PFLOPS sparse vs 35 PFLOPS dense drops the cost per perf in units of $/hr per Marketed PFLOP by 35%—a valid comparison if AI Labs can indeed successfully harness Sparse FP4 on the VR NVL72.

② Atomic Claim

VR 每顆 GPU 的 TCO 高於 MI4XX，但 marketed PFLOPs 又低於 MI4XX，因此 VR 在 TCO per PFLOP 上處於劣勢。

• Epistemic Mode: ASSERTED
• Mapping Status: PARTIAL

③ Semantic Frame

{
  "frame_type": "NARY_RELATION",
  "participants": [
    {
      "node": {
        "id": "04_knowledge_base/MI400",
        "label": "MI4XX"
      },
      "role": "comparison_platform"
    },
    {
      "node": {
        "id": "04_knowledge_base/GPU",
        "label": "GPU"
      },
      "role": "cost_unit"
    }
  ],
  "qualifiers": {
    "condition_text": "VR is plain text shorthand; higher TCO/GPU and lower marketed PFLOPs",
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "relation_type": "VR_TCO_PER_PFLOP_DISADVANTAGE_VS_MI4XX"
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
comparison_platform	MI4XX	MI400
cost_unit	GPU	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。