---
cited_by:
  - '[[thoughts/inference/kimi-k3#5-sharp-edges-each-fix-shows-an-invariant|kimik3 §5]]'
date: '2026-11-11'
kind: pr
local:
  - file:///Users/aarnphm/workspace/vllm/vllm/model_executor/layers/attention/mla_attention.py
merged: '2026-09-09'
number: 53664
source:
  - https://github.com/vllm-project/vllm/blob/d8d53f17c231bf477aa89f25a904f9b7c54fbcb9/vllm/model_executor/layers/attention/mla_attention.py
state: merged
tags:
  - vllm
  - vllm-pr
  - new-architectures
title: '#53664 [Rocm][Kimi-k3] Fix pipeline_parallel support for the kimik3 DCP mode'
topic: new-architectures
url: https://github.com/vllm-project/vllm/pull/53664
created: '2026-11-11'
modified: '2026-11-11'
published: '2026-11-11'
pageLayout: default
slug: hinterland/papers/maps/prs/53664
permalink: https://aarnphm.xyz/hinterland/papers/maps/prs/53664.md
generator:
  quartz: v4.6.0
  hostedProvider: Cloudflare
  baseUrl: aarnphm.xyz
full: https://aarnphm.xyz/llms-full.txt
---
Assert pipeline-parallel capability for both Kimi-K3 model architectures. · full dump: [[thoughts/inference/vllm-prs/53664|#53664]]

