2025-12-04_aws-trainium3-deep-dive-a-potential::TRN3-0017
① SA Source
- Source: 開啟完整 SA 文章
- Section:
Trainium3 Server Types and Specifications Overview - Line hint:
111
Context Before
The scale-up bandwidth per Trainium3 chip is doubled vs Trn2 by moving to PCIe Gen 6 which offers 64Gbps per lane (uni-directional) vs the 32Gbps per lane offered by PCIe Gen 5. Trainium3 uses 144 active lanes of PCIe for scale up, which on Gen6 means each Trainium3 supports 1.2 TB/s (uni-directional) of scale-up bandwidth per chip.
Scale-out bandwidth support is doubled to a maximum of 400 Gb/s, but most Trainium3 racks produced will stick with the 200Gb/s per XPU scale-out speed that was used for Trn2.
Evidence
Context After
Trainium3 Rack Architecture
Zooming out to the rack solution level, AWS announced at re:Invent the Trainium3 (Gen1) UltraServer and the Trainium3 (Gen2) UltraServer, which correspond to the Trainium3 NL32x2 Switched and Trainium3 NL72x2 Switched names respectively. The key difference between the Trainium3 NL32x2 Switched and the Trainium3 NL72x2 Switched is in the scale-up networking topology and rack architecture – this section will cover how topology and architecture differs between the two SKUs and discuss the AI workloads that each architecture is the most suitable and optimized for.
② Atomic Claim
Trainium4 計畫使用 8 stacks HBM4，相較 Trainium3 達到 4 倍 memory bandwidth 與 2 倍 capacity。
- Epistemic Mode:
EXPECTED - Mapping Status:
PARTIAL
③ Semantic Frame
{
"attribute": "trainium4_memory_configuration",
"context_nodes": [
{
"id": "04_knowledge_base/HBM4",
"label": "HBM4"
},
{
"id": "04_knowledge_base/Trainium",
"label": "Trainium3"
}
],
"entity": {
"id": "02_companies/Amazon",
"label": "Amazon"
},
"frame_type": "ATTRIBUTE",
"qualifiers": {
"condition_text": null,
"numeric_mentions": [
"8",
"4x",
"2x"
],
"temporal_mentions": []
},
"value": {
"numeric_mentions": [
"8",
"4x",
"2x"
],
"value_text": "8 HBM4 stacks; 4x bandwidth and 2x capacity vs Trainium3"
}
}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。