VR2R-0568
① SA Source
- Source: 開啟完整 SA 文章
- Section:
Scale-Up Network - NVLink 6 - Line hint:
474
Context Before
While the VR NVL72 system features GPUs and
Evidence
Context After
The second meaningful difference between the NVL72 and HGX deployments is that the former has a scale-out bandwidth of 1.6T per GPU while the latter only has a scale-out bandwidth of 800G per GPU.
② Atomic Claim
……與透過 copper cables 連接的 scale-up switches;相較之下,VR HGX system 的 servers 由 8 顆 Rubin GPUs 與 4 顆 NVLink Switch chips 組成。
- Epistemic Mode:
ASSERTED - Mapping Status:
COMPLETE
③ Semantic Frame
{
"frame_type": "NARY_RELATION",
"participants": [
{
"node": {
"id": "04_knowledge_base/HGX",
"label": "HGX"
},
"role": "server_platform"
},
{
"node": {
"id": "04_knowledge_base/Vera Rubin",
"label": "Rubin"
},
"role": "processor"
},
{
"node": {
"id": "04_knowledge_base/GPU",
"label": "GPU"
},
"role": "processor_type"
},
{
"node": {
"id": "04_knowledge_base/NVLink",
"label": "NVLink"
},
"role": "switch_interconnect"
}
],
"qualifiers": {
"condition_text": "continuation of VR2R-0567; copper-connected scale-up switches are plain context",
"numeric_mentions": [
"8",
"4"
],
"temporal_mentions": []
},
"relation_type": "VR_HGX_SERVER_COMPONENT_CONFIGURATION"
}④ Canonical Entity Mapping
| Role | Surface Label | Canonical Target |
|---|---|---|
| server_platform | HGX | HGX |
| processor | Rubin | 04_knowledge_base/Vera Rubin |
| processor_type | GPU | GPU |
| switch_interconnect | NVLink | NVLink |
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。