VR2R-0266_r1

VR2R-0266

① SA Source

• Source: 開啟完整 SA 文章
• Section: Compute Tray Topology
• Line hint: 232

Context Before

Essentially, this means ConnectX-8 has two hosts, Grace CPU and B300 GPU. This improves latency in the backend network. However, for VR NVL72, the direct connection between Rubin GPU and ConnectX-9 is reverted back to the same design as GB200, as Rubin does not have PCIe bandwidth for two ConnectX-9. Rubin connects to Vera via C2C link, then Vera will connect to ConnectX-9 via PCIe6 lanes.

Evidence

Secondly, the local NVMe storage for Rubin has been moved to a different location from that of NVMe storage in Grace Blackwell.

Context After

Previously, local NVMe storage was managed by BlueField-3. For VR NVL72, the local NVMe storage is physically on the Orchid module managed by the ConnectX-9.

② Atomic Claim

第二，Rubin 的 local NVMe storage 已移到與 Grace Blackwell 中 NVMe storage 不同的位置。

• Epistemic Mode: ASSERTED
• Mapping Status: PARTIAL

③ Semantic Frame

{
  "frame_type": "NARY_RELATION",
  "participants": [
    {
      "node": {
        "id": "04_knowledge_base/Vera Rubin",
        "label": "Rubin"
      },
      "role": "current_platform"
    },
    {
      "node": {
        "id": "04_knowledge_base/NVIDIA Grace",
        "label": "Grace"
      },
      "role": "reference_cpu"
    },
    {
      "node": {
        "id": "04_knowledge_base/Blackwell",
        "label": "Blackwell"
      },
      "role": "reference_gpu_generation"
    }
  ],
  "qualifiers": {
    "condition_text": "local NVMe storage is plain text in source claim",
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "relation_type": "RUBIN_LOCAL_NVME_LOCATION_DIFFERS_FROM_GRACE_BLACKWELL"
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
current_platform	Rubin	04_knowledge_base/Vera Rubin
reference_cpu	Grace	04_knowledge_base/NVIDIA Grace
reference_gpu_generation	Blackwell	Blackwell

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。