---
cited_by:
  - '[[thoughts/inference/kv-transfer#2-what-each-layout-costs-the-transfer|reading-map §2]]'
  - '[[thoughts/inference/kv-transfer#5-reading-order|reading-map §5]]'
date: '2026-11-11'
kind: pr
local:
  - file:///Users/aarnphm/workspace/vllm/vllm/distributed/kv_transfer/kv_connector/v1/hf3fs/utils/gather_scatter_helper.py
  - file:///Users/aarnphm/workspace/vllm/vllm/distributed/kv_transfer/kv_connector/v1/offloading/worker.py
  - file:///Users/aarnphm/workspace/vllm/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.py
  - file:///Users/aarnphm/workspace/vllm/vllm/distributed/kv_transfer/kv_connector/utils.py
  - file:///Users/aarnphm/workspace/vllm/vllm/distributed/kv_transfer/kv_connector/v1/hf3fs/hf3fs_connector.py
merged: '2026-08-22'
number: 51718
source:
  - https://github.com/vllm-project/vllm/blob/8bdc70ec7b379279ec0152343239c2d50aced687/vllm/distributed/kv_transfer/kv_connector/v1/nixl/base_worker.py
  - https://github.com/vllm-project/vllm/blob/8bdc70ec7b379279ec0152343239c2d50aced687/vllm/distributed/kv_transfer/kv_connector/v1/hf3fs/utils/gather_scatter_helper.py
  - https://github.com/vllm-project/vllm/blob/8bdc70ec7b379279ec0152343239c2d50aced687/vllm/distributed/kv_transfer/kv_connector/v1/offloading/worker.py
  - https://github.com/vllm-project/vllm/blob/8bdc70ec7b379279ec0152343239c2d50aced687/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.py
  - https://github.com/vllm-project/vllm/blob/8bdc70ec7b379279ec0152343239c2d50aced687/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/store/worker.py
  - https://github.com/vllm-project/vllm/blob/8bdc70ec7b379279ec0152343239c2d50aced687/vllm/distributed/kv_transfer/kv_connector/utils.py
  - https://github.com/vllm-project/vllm/blob/8bdc70ec7b379279ec0152343239c2d50aced687/vllm/distributed/kv_transfer/kv_connector/v1/hf3fs/hf3fs_connector.py
  - https://github.com/vllm-project/vllm/blob/8bdc70ec7b379279ec0152343239c2d50aced687/vllm/distributed/kv_transfer/kv_connector/v1/moriio/moriio_layout.py
state: merged
tags:
  - vllm
  - vllm-pr
  - hetero-tp-kv-layout
title: '#51718 [6/N][KV-Cache Layout Refactor] Standardize KV cache layout'
topic: hetero-tp-kv-layout
url: https://github.com/vllm-project/vllm/pull/51718
created: '2026-11-11'
modified: '2026-11-11'
published: '2026-11-11'
pageLayout: default
slug: hinterland/papers/maps/prs/51718
permalink: https://aarnphm.xyz/hinterland/papers/maps/prs/51718.md
generator:
  quartz: v4.6.0
  hostedProvider: Cloudflare
  baseUrl: aarnphm.xyz
full: https://aarnphm.xyz/llms-full.txt
---
The core of the KV-cache layout standardization series (RFC [#42082](https://github.com/aarnphm/aarnphm.github.io/issues/42082)). · full dump: [[thoughts/inference/vllm-prs/51718|#51718]]

