IX2-0077
① SA Source
- Source: 開啟完整 SA 文章
- Section:
Disaggregated Prefill - Line hint:
109
Context Before
When prefill and decode performed on the same engine, prefill constantly disrupts decode batches leading to worse overall performance.
Disaggregated Prefill
Evidence
Disaggregated prefill (aka PD disaggregation or simply “disagg”) is the practice of separating the prefill and decode phases across separate pools of GPUs or clusters
Context After
Tensor Parallel, Expert Parallel, Data Parallel (TP, EP, DP)
TP allows for maximize interactivity at small batch sizes, but it must carry out an all-reduce at every layer. EP shards experts, exploiting MoE sparsity, with the drawback being an all-to-all collective (which is more costly than simpler collectives like all-reduce) is carried out for MoE layers and can be imbalanced at small batches. DP replicates the entire model (or just parts of a model, like attention) on multiple groups of GPUs (ranks) and then load balances requests among ranks. It is the simplest to scale, but repeats weight loading which can be wasteful at scale.
② Atomic Claim
Disaggregated prefill（也稱 PD disaggregation 或簡稱 disagg）是把 prefill 與 decode 階段分離到不同的 GPUs pools 或 clusters 的做法。
- Epistemic Mode:
ASSERTED - Mapping Status:
PARTIAL
③ Semantic Frame
{
"attribute": "UNSPECIFIED_ATTRIBUTE",
"context_nodes": [
{
"id": "04_knowledge_base/Prefill",
"label": "prefill"
},
{
"id": "04_knowledge_base/Decode",
"label": "decode"
},
{
"id": "04_knowledge_base/GPU",
"label": "GPUs"
}
],
"entity": {
"id": "04_knowledge_base/Disaggregated prefill",
"label": "Disaggregated prefill"
},
"frame_type": "ATTRIBUTE",
"qualifiers": {
"condition_text": null,
"numeric_mentions": [],
"temporal_mentions": []
},
"value": {
"numeric_mentions": [],
"value_text": "Disaggregated prefill（也稱 PD disaggregation 或簡稱 disagg）是把 prefill 與 decode 階段分離到不同的 GPUs pools 或 clusters 的做法。"
}
}④ Canonical Entity Mapping
| Role | Surface Label | Canonical Target |
|---|---|---|
| entity | Disaggregated prefill | 04_knowledge_base/Disaggregated prefill |
| context_0 | prefill | Prefill |
| context_1 | decode | Decode |
| context_2 | GPUs | GPU |
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。