clm_fbfa23b97a984f21911d842e6c6e6099_r1

GB2-0024

① SA Source

• Source: 開啟完整 SA 文章
• Section: Networking
• Line hint: 261

Context Before

Out of Band Networking

As a quick refresher, the frontend networking is just your normal ethernet network that you use to connect to the internet, SLURM/Kubernetes, networked storage, data loading, model checkpoints. This network is typically 25-50Gb/s per GPU, so on a HGX H100 server, it will be 200-400Gb/s per server, while on a GB200 computer tray node, it will be 200-800Gb/s per server depending on the configuration.

Evidence

Your backend network is used to scale out GPU-GPU communications across hundred to thousands of racks. This network could either be Nvidia’s Infiniband or Nvidia Spectrum-X Ethernet or Broadcom Ethernet.

Context After

The scale-up accelerator interconnect (NVLink on Nvidia ↗, Infinity Fabric/UALink on AMD, ICI on Google TPU ↗, NeuronLink on Amazon Trainium 2) is an ultra-high speed network that connects GPUs together within a system. On Hopper, this network connected 8 GPUs together at 450GB/s each while on Blackwell NVL72, it will connect 72 GPUs together at 900GB/ each. There is a variant of Blackwell called NVL576 that will connect 576 GPUs together but basically no customers will opt for it. In general, your accelerator interconnect is 8-10x faster than your backend networking.

Lastly, there is your out of band management network which is used for re-imaging your operating system, monitor node health such as fan speed, temperatures, power draw, etc. Your baseboard management controller (BMC) on servers, PDUs, switches, CDUs are usually connected to this network to monitor and control these IT equipment.

② Atomic Claim

GB200 類系統的 backend network 用於跨數百至數千個 racks 的 GPU-to-GPU scale-out，選項包括 Nvidia InfiniBand、Spectrum-X Ethernet 或 Broadcom Ethernet。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "frame_type": "NARY_RELATION",
  "participants": [
    {
      "node": {
        "id": "04_knowledge_base/GPU",
        "label": "GPU"
      },
      "role": "workload"
    },
    {
      "node": {
        "id": "04_knowledge_base/InfiniBand",
        "label": "InfiniBand"
      },
      "role": "fabric"
    },
    {
      "node": {
        "id": "04_knowledge_base/NVIDIA Spectrum-X",
        "label": "NVIDIA Spectrum-X"
      },
      "role": "fabric"
    },
    {
      "node": {
        "id": "04_knowledge_base/Ethernet",
        "label": "Ethernet"
      },
      "role": "fabric"
    }
  ],
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "relation_type": "BACKEND_NETWORK_OPTIONS"
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
workload	GPU	GPU
fabric	InfiniBand	InfiniBand
fabric	NVIDIA Spectrum-X	04_knowledge_base/NVIDIA Spectrum-X
fabric	Ethernet	Ethernet

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。