VR2-0632
① SA Source
- Source: 開啟完整 SA 文章
- Section:
Connecting GPUs in the Scale-Out InfiniBand Network - Line hint:
759
Context Before
Source: SemiAnalysis AI Networking Model ↗
Evidence
For the Vera Rubin NVL72 deployment however, there are two 800G OSFP cages per GPU for a total of 1.6T per GPU bandwidth. Having two 800G logical ports per GPU is advantageous because it allows multi-plane network deployments without complex fiber management – by splitting one logical GPU two ways to two different leaf switches. As such, larger network clusters can be built with two 800G logical ports than if only one 1.6T logical port were used. In fact, as we have explained in multiple prior articles such as the networking sections of NVIDIA’s Optical Boogeyman ↗ and Microsoft’s AI Strategy Deconstructed ↗, this relationship is dictated by a simple formula for the maximum number of hosts that can be supported using a switch of k ports on an L-layer:
Context After
2(k2)L
Illustratively, consider two hypothetical VR NVL deployments with 1.6T and 800G logical ports respectively. A 1-plane, 3-layer network with one 1.6T logical ports achieves only a maximum cluster size of 93,312 GPUs, or:
② Atomic Claim
如 SemiAnalysis 過去在 NVIDIA Optical Boogeyman、Microsoft AI Strategy Deconstructed 等 networking 章節所述，此關係可由一個簡單公式決定：在 L-layer network 中，使用 k-port switch 可支援的最大 hosts 數量。
- Epistemic Mode:
ASSERTED - Mapping Status:
PARTIAL
③ Semantic Frame
{
"additional_nodes": [],
"frame_type": "RELATION",
"object": {
"id": "02_companies/MSFT",
"label": "Microsoft"
},
"predicate": "USES",
"qualifiers": {
"condition_text": null,
"numeric_mentions": [],
"temporal_mentions": []
},
"subject": {
"id": "02_companies/NVDA",
"label": "NVIDIA"
}
}④ Canonical Entity Mapping
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。