IX2-0511

① SA Source

Context Before

image

Source: SemiAnalysis InferenceX

Evidence

In comparing DeepSeek R1 FP8 perf between H200 (SGLang) and MI325X (SGLang), not much has changed since our initial release of InferenceXv1 last October

Context After

One thing we note is the considerably smaller interactivity range for MI325X than H200, with H200 ranging from 30-90 tok/sec/user whereas MI325X ranges from only 13-35 tok/sec/user. This is problematic for providers who would like to serve users at a broader range of interactivity.

image

② Atomic Claim

比較 DeepSeek R1 FP8H200SGLang)與 MI325XSGLang)上的效能後,自去年 10 月 InferenceXv1 首次發布以來,整體變化不大。

  • Epistemic Mode: ASSERTED
  • Mapping Status: COMPLETE

③ Semantic Frame

{
  "comparison_expression": "比較 DeepSeek R1 FP8 在 H200(SGLang)與 MI325X(SGLang)上的效能後,自去年 10 月 InferenceXv1 首次發布以來,整體變化不大。",
  "entities": [
    {
      "id": "02_companies/DeepSeek",
      "label": "DeepSeek"
    },
    {
      "id": "04_knowledge_base/FP8",
      "label": "FP8"
    },
    {
      "id": "04_knowledge_base/H200",
      "label": "H200"
    },
    {
      "id": "04_knowledge_base/SGLang",
      "label": "SGLang"
    },
    {
      "id": "04_knowledge_base/MI325X",
      "label": "MI325X"
    }
  ],
  "frame_type": "COMPARISON",
  "metric": "COMPUTE_PERFORMANCE",
  "operator": "UNSPECIFIED_COMPARISON",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "10"
    ],
    "temporal_mentions": []
  }
}

④ Canonical Entity Mapping

RoleSurface LabelCanonical Target
comparison_entity_0DeepSeekDeepSeek
comparison_entity_1FP8FP8
comparison_entity_2H200H200
comparison_entity_3SGLangSGLang
comparison_entity_4MI325XMI325X

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。