clm_d73aab9539444f7290c1d511d03ab235_r1

IX2-0384

① SA Source

• Source: 開啟完整 SA 文章
• Section: WideEP
• Line hint: 630

Context Before

In this section, we will go deeper on expert parallelism and go on to explain what _wide _expert parallelism is. We will then explain the idea of Disaggregated Prefill, how it is different from WideEP, and how WideEP and Disaggregated Prefill are used in unison to achieve SOTA performance.

WideEP

Evidence

In MoE architectures, only a subset of “experts” are activated for each token

Context After

Consider serving DeepSeek R1 on a single 8-GPU server. At 671B parameters, some form of parallelism is required to fit the model across available HBM. The naive approach is tensor parallelism (TP), which shards every weight matrix across all GPUs. This works well for dense models but ignores the sparse activation pattern of MoE. With TP=8, each expert’s weights are sharded across all 8 GPUs, meaning every expert activation requires an all-reduce across all GPUs & the reduction dims of the GEMM is smaller leading to lower arithmetic intensity, even though only 8 of 256 experts activate per token. TP treats each expert like a dense layer, paying full cross-GPU communication cost while the model’s sparsity goes unexploited.

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.

② Atomic Claim

在 MoE architecture 中，每個 token 只會啟動部分「experts」。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "attribute": "COUNT",
  "context_nodes": [],
  "entity": {
    "id": "04_knowledge_base/Mixture of Experts",
    "label": "MoE"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [],
    "value_text": "在 MoE architecture 中，每個 token 只會啟動部分「experts」。"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	MoE	04_knowledge_base/Mixture of Experts

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。