VR2-0136
① SA Source
- Source: 開啟完整 SA 文章
- Section:
BlueField-4 - Line hint:
173
Context Before
Source: Nvidia VR NVL72 Component BoM and Power Budget Model ↗
Evidence
Instead of doing a bespoke tapeout with compute and networking, NVIDIA simply reuses their large Grace CPU die, co-packaged with a ConnectX-9 die to make an 800G DPU with massive compute capabilities
Context After
Spectrum-6
② Atomic Claim
NVIDIA 不再為 compute 與 networking 做專用 tapeout，而是直接重用大型 Grace CPU die，並與 ConnectX-9 die co-package，形成具備大量 compute 能力的 800G DPU。
- Epistemic Mode:
ASSERTED - Mapping Status:
COMPLETE
③ Semantic Frame
{
"frame_type": "NARY_RELATION",
"participants": [
{
"node": {
"id": "02_companies/NVDA",
"label": "NVIDIA"
},
"role": "endpoint_or_system"
},
{
"node": {
"id": "04_knowledge_base/Tapeout",
"label": "tapeout"
},
"role": "connected_entity"
},
{
"node": {
"id": "04_knowledge_base/NVIDIA Grace",
"label": "Grace CPU"
},
"role": "connected_entity"
},
{
"node": {
"id": "04_knowledge_base/ConnectX-9",
"label": "ConnectX-9"
},
"role": "connected_entity"
},
{
"node": {
"id": "04_knowledge_base/800G",
"label": "800G"
},
"role": "connected_entity"
}
],
"qualifiers": {
"condition_text": null,
"numeric_mentions": [
"9",
"800"
],
"temporal_mentions": []
},
"relation_type": "CONNECTS_TO"
}④ Canonical Entity Mapping
| Role | Surface Label | Canonical Target |
|---|---|---|
| endpoint_or_system | NVIDIA | NVDA |
| connected_entity | tapeout | Tapeout |
| connected_entity | Grace CPU | 04_knowledge_base/NVIDIA Grace |
| connected_entity | ConnectX-9 | ConnectX-9 |
| connected_entity | 800G | 800G |
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。