clm_87dcde8f24b048cd913c9f8cdcc83fd1_r1

IX2-0389

① SA Source

• Source: 開啟完整 SA 文章
• Section: WideEP
• Line hint: 632

Context Before

WideEP

By now, most frontier AI labs employ Mixture of Experts (MoE) model architectures as opposed to dense. In MoE architectures, only a subset of “experts” are activated for each token. For instance, DeepSeek R1 has 671B total parameters, but only 37B active parameters. Specifically, DeepSeek R1 has 256 routed experts (and 1 shared expert) with each token being routed to 8 distinct experts. This architecture lends itself naturally to expert parallelism (EP), which evenly distributes expert weights across some number of GPUs.

Evidence

At 671B parameters, some form of parallelism is required to fit the model across available HBM

Context After

Expert parallelism takes a more well-suited approach, assigning whole experts to individual GPUs. With EP=8, we divide the 256 experts per layer across 8 GPUs for a total of 32 experts/layer/GPU. Each GPU holds approximately 1/8th of the expert weights plus a full replica of the non-expert weights (attention projections, embeddings, normalization, and the shared expert). Since roughly 90%+ of DeepSeek R1’s parameters are routed expert weights, EP captures most of the memory savings, and replicating the remaining less than 30B non-expert parameters across all 8 GPUs is affordable.

The forward pass proceeds in two phases per layer. During attention, each GPU acts as an independent data-parallel rank, processing its own subset of requests using its replicated non-expert weights, no inter-GPU communication is needed. During the MoE phase, a lightweight router determines which experts each token requires, and tokens are dispatched to the appropriate GPUs via all-to-all communication. Each GPU executes its local experts on only the tokens routed to it, and results are returned via a second all-to-all.

② Atomic Claim

模型有 671B parameters，因此必須使用某種 parallelism，才能把模型放進可用 HBM。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "attribute": "LIFECYCLE_STATUS",
  "context_nodes": [],
  "entity": {
    "id": "04_knowledge_base/HBM",
    "label": "HBM"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "671B"
    ],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [
      "671B"
    ],
    "value_text": "模型有 671B parameters，因此必須使用某種 parallelism，才能把模型放進可用 HBM。"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	HBM	HBM

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。