IX2-0034
① SA Source
- Source: 開啟完整 SA 文章
- Section:
Key Observations and Results to Highlight - Line hint:
49
Context Before
When it comes to the latest inference techniques that are used by the most prominent frontier large-scale inference services (such as disagg prefill+wideEP+FP4), Nvidia absolutely frame mogs with the B200, B300 and ASU frat leader, rack scale GB200/GB300 NVL72 across both SGLang and TRTLLM. Nvidia GPUs also dominate when it comes to energy efficiency, with much lower all-in provisioned picoJoules of energy per token across all workloads.
Turning to AMD, we find that the biggest issue with inference on their systems and using their software is composability ↗. That is, many of AMDs inference optimization implementations work well in isolation, but when combined with other optimizations, the result is not as competitive as one would expect. Specifically, the composability of disagg prefill, wideEP and FP4 inference optimizations needs significant improvement.
Evidence
This will happen after Chinese New Year as most of their disagg prefill+wideEP 10x inference engineers are based in China
Context After
Nvidia’s GB300 NVL72 doesn’t disappoint. It achieves up to 100x on FP8 vs FP4 compared to even a strong H100 disagg+wideEP+MTP baseline and 65x on FP8 vs FP8. On H100 vs GB200 NVL72, we see up to 55x realized performance difference at 75 tok/s/user. Rack scale Blackwell NVL72 is framemogging hopper and makes hopper looks like it is jestermaxxing. As Jensen said at GTC 2025, he is chief revenue destroyer. ↗
At GTC 2024, Jensen claimed that Blackwell will deliver up to 30x perf on inference compared to H100, Jensen under promised & overdelivered on Blackwell inference performance. This should curtail the instances of analysts cracking “Jensen Math” jokes for some time.
② Atomic Claim
這項工作預計在農曆新年後展開，因為 AMD 多數負責 disagg prefill+wideEP 的核心 inference engineers 位於中國。
- Epistemic Mode:
ASSERTED - Mapping Status:
PARTIAL
③ Semantic Frame
{
"attribute": "UNSPECIFIED_ATTRIBUTE",
"context_nodes": [
{
"id": "04_knowledge_base/Prefill",
"label": "prefill"
}
],
"entity": {
"id": "02_companies/AMD",
"label": "AMD"
},
"frame_type": "ATTRIBUTE",
"qualifiers": {
"condition_text": null,
"numeric_mentions": [],
"temporal_mentions": []
},
"value": {
"numeric_mentions": [],
"value_text": "這項工作預計在農曆新年後展開，因為 AMD 多數負責 disagg prefill+wideEP 的核心 inference engineers 位於中國。"
}
}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。