clm_0e9cc67503414ec59589cc94b0abe5b4_r1

2025-06-13_amd-advancing-ai-mi350x-and-mi400-ualoe72-mi500-ual256::AMD25-0014

① SA Source

• Source: 開啟完整 SA 文章
• Section: MI350X and MI355X Specs
• Line hint: 79

Context Before

MI350X and MI355X Specs

There are two versions of the CDNA4 chips in this series – namely the MI350X and the MI355X. The MI350X is the 1,000W version that is air cooled while the MI355X is a 1,400W version that supports both air cooling and DLC liquid cooling. Even though the MI355X uses 1.4x more power, the on-paper specs show that it is less than 10% faster than the MI350X in terms of TFLOPS throughput. However, we expect realized performance to be greater than 10% better for the MI355X because published specs are often never achieved due to power limitations. These published specs assume that the peak clock speed can be held in real workloads, but that’s simply not the case on both AMD and Nvidia systems.

Evidence

This means that MI355X FP6 is 2.2x faster than B200 FP6.

Context After

SemiAnalysis benchmarking has shown ↗ that even though the MI300X and the H100 each show the same on-paper TFLOP/s for FP16 as for BF16 (i.e. Nvidia’s FP16 TF = BF16 = 989 TFLOP/s, AMD’s FP16 = BF16 = 1307 TFLOP/s) in practice – each card delivers different realized TFLOPs when running FP16 vs BF16. We will be publishing an article in the near future running microbenchmarks to figure out a realistic TFLOP/s for MI355X FP6 versus FP4.

② Atomic Claim

依 paper specs，MI355X FP6 peak throughput 約為 B200 FP6 的 2.2 倍。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "comparison_expression": "MI355X FP6 2.2x B200 FP6",
  "entities": [
    {
      "id": "04_knowledge_base/MI355X",
      "label": "MI355X"
    },
    {
      "id": "04_knowledge_base/NVIDIA B200",
      "label": "B200"
    },
    {
      "id": "04_knowledge_base/FP6",
      "label": "FP6"
    }
  ],
  "frame_type": "COMPARISON",
  "metric": "peak_fp6_throughput",
  "operator": "2_2X",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "2.2x"
    ],
    "temporal_mentions": []
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
comparison_entity_0	MI355X	MI355X
comparison_entity_1	B200	04_knowledge_base/NVIDIA B200
comparison_entity_2	FP6	FP6

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。