VR2R-0264
① SA Source
- Source: 開啟完整 SA 文章
- Section:
Compute Tray Topology - Line hint:
371
Context Before
Source: Nvidia VR NVL72 BoM and Power Budget Model ↗
Firstly, the connection between the GPU and the ConnectX NICs evolved from GB200 to GB300 then to Vera Rubin. For GB200, the GPU does not have direct access to the ConnectX-7. Instead, B200 connects to Grace CPU via C2C connection then the Grace CPU talks to the ConnectX-7 over PCIe 5. For GB300, Nvidia introduced NIC direct to ConnectX-8, which allows the B300 GPU to communicate directly with the ConnectX-8 NIC without going through the Grace CPU.
Evidence
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 Connect-9. Rubin connects to Vera via C2C link, then Vera will connect to ConnectX-9 via PCIe6 lanes.
Context After
Secondly, the local NVMe storage for Rubin has been moved to a different location from that of NVMe storage in Grace Blackwell. 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
但在 VR NVL72 中，Rubin GPU 與 ConnectX-9 的 direct connection 又回到與 GB200 相同的設計，因為 Rubin 沒有足夠的 PCIe bandwidth 供兩個 ConnectX-9 使用。
- Epistemic Mode:
ASSERTED - Mapping Status:
COMPLETE
③ Semantic Frame
{
"frame_type": "NARY_RELATION",
"participants": [
{
"node": {
"id": "04_knowledge_base/VR NVL72",
"label": "VR NVL72"
},
"role": "system"
},
{
"node": {
"id": "04_knowledge_base/Vera Rubin",
"label": "Rubin"
},
"role": "accelerator"
},
{
"node": {
"id": "04_knowledge_base/GPU",
"label": "GPU"
},
"role": "accelerator_type"
},
{
"node": {
"id": "04_knowledge_base/ConnectX-9",
"label": "ConnectX-9"
},
"role": "network_device"
},
{
"node": {
"id": "04_knowledge_base/GB200",
"label": "GB200"
},
"role": "reference_system"
},
{
"node": {
"id": "04_knowledge_base/PCIe",
"label": "PCIe"
},
"role": "interface"
}
],
"qualifiers": {
"condition_text": "Rubin lacks PCIe bandwidth for two ConnectX-9 devices",
"numeric_mentions": [
"two"
],
"temporal_mentions": []
},
"relation_type": "VR_REVERTS_GPU_CONNECTX9_PATH_TO_GB200_STYLE"
}④ Canonical Entity Mapping
| Role | Surface Label | Canonical Target |
|---|---|---|
| system | VR NVL72 | 04_knowledge_base/VR NVL72 |
| accelerator | Rubin | 04_knowledge_base/Vera Rubin |
| accelerator_type | GPU | GPU |
| network_device | ConnectX-9 | ConnectX-9 |
| reference_system | GB200 | GB200 |
| interface | PCIe | PCIe |
⑤ V2 Draft（需要修正時直接編輯）
Inline repair
只編輯以下三個 Draft 區塊。完成後將
review_action設為REQUEST_REPAIR，再手動執行 Repair Preview。
執行腳本前，Draft 不會寫入 Claim Store 或 Graph。
V2 Atomic Claim Draft
但在 VR NVL72 中，Rubin GPU 與 ConnectX-9 的 direct connection 又回到與 GB200 相同的設計，因為 Rubin 沒有足夠的 PCIe bandwidth 供兩個 ConnectX-9 使用。
V2 Semantic Frame Draft
{
"frame_type": "NARY_RELATION",
"participants": [
{
"node": {
"id": "04_knowledge_base/VR NVL72",
"label": "VR NVL72"
},
"role": "system"
},
{
"node": {
"id": "04_knowledge_base/Vera Rubin",
"label": "Rubin"
},
"role": "accelerator"
},
{
"node": {
"id": "04_knowledge_base/GPU",
"label": "GPU"
},
"role": "accelerator_type"
},
{
"node": {
"id": "04_knowledge_base/ConnectX-9",
"label": "ConnectX-9"
},
"role": "network_device"
},
{
"node": {
"id": "04_knowledge_base/GB200",
"label": "GB200"
},
"role": "reference_system"
},
{
"node": {
"id": "04_knowledge_base/PCIe",
"label": "PCIe"
},
"role": "interface"
}
],
"qualifiers": {
"condition_text": "Rubin lacks PCIe bandwidth for two ConnectX-9 devices",
"numeric_mentions": [
"two"
],
"temporal_mentions": []
},
"relation_type": "VR_REVERTS_GPU_CONNECTX9_PATH_TO_GB200_STYLE"
}V2 Revision Metadata Draft
{
"epistemic_mode": "ASSERTED",
"mapping_status": "COMPLETE"
}⑥ 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。