VR2-0059
① SA Source
- Source: 開啟完整 SA 文章
- Section:
Rubin - Line hint:
111
Context Before
Increasing clock speed 25% from 1.90GHz to 2.38GHz
Evidence
Additionally, Nvidia claims up to an effective 50 PFLOPS of FP4 performance can be achieved with an updated 3rd generation Transformer Engine that replaces 2:4 structured sparsity from prior generations
Context After
Notably, the Tensor core width doubling only applies to FP4 and FP8, with BF16 and TF32 remaining the same as Blackwell, resulting in performance scaling only 1.6x of Blackwell. This architectural decision reflects NVIDIA’s belief that most training and inference workloads will move away from TF32 and BF16 and onto FP8 and FP4.
On the memory front, the move to HBM4 means double the bus width per stack, running at 10.8 GT/s for 22TB/s total bandwidth or 2.75x Blackwell at the same 288GB capacity as GB300. Memory bandwidth has been upgraded significantly from the original 13TB/s advertised at GTC 2025. In order to catch up to AMD MI450’s memory bandwidth, Nvidia requested much higher HBM4 pin speeds from the DRAM suppliers - well above the speeds that was in the JEDEC specification for HBM4.
② Atomic Claim
此外，Nvidia 宣稱，更新後的第三代 Transformer Engine 取代過去世代的 2:4 structured sparsity 後，可達最高有效 50 PFLOPS 的 FP4 效能。
- Epistemic Mode:
ATTRIBUTED - Mapping Status:
PARTIAL
③ Semantic Frame
{
"frame_type": "NARY_RELATION",
"participants": [
{
"node": {
"id": "02_companies/NVDA",
"label": "Nvidia"
},
"role": "subject"
},
{
"node": {
"id": "04_knowledge_base/Transformer Engine",
"label": "Transformer Engine"
},
"role": "participant"
},
{
"node": {
"id": "04_knowledge_base/2-of-4 Sparsity",
"label": "2:4 structured sparsity"
},
"role": "participant"
},
{
"node": {
"id": "04_knowledge_base/FP4",
"label": "FP4"
},
"role": "participant"
}
],
"qualifiers": {
"condition_text": null,
"numeric_mentions": [
"2",
"4 s",
"50 PFLOPS"
],
"temporal_mentions": []
},
"relation_type": "UPDATES"
}④ Canonical Entity Mapping
| Role | Surface Label | Canonical Target |
|---|---|---|
| subject | Nvidia | NVDA |
| participant | Transformer Engine | 04_knowledge_base/Transformer Engine |
| participant | 2:4 structured sparsity | 04_knowledge_base/2-of-4 Sparsity |
| participant | FP4 | FP4 |
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。