---
cited_by:
  - '[[thoughts/inference/replayssm|replayssm]]'
date: '2026-11-11'
kind: pr
local:
  - file:///Users/aarnphm/workspace/vllm/vllm/v1/attention/backends/mamba_attn.py
  - file:///Users/aarnphm/workspace/vllm/vllm/model_executor/layers/mamba/mamba_utils.py
  - file:///Users/aarnphm/workspace/vllm/vllm/model_executor/layers/mamba/mamba_mixer2.py
number: 47576
source:
  - https://github.com/vllm-project/vllm/blob/99dcd7600195c3a7ffaa23eb782591d04df1dc65/vllm/model_executor/layers/mamba/ops/selective_state_update_replayssm_spec.py
  - https://github.com/vllm-project/vllm/blob/99dcd7600195c3a7ffaa23eb782591d04df1dc65/vllm/model_executor/layers/fla/ops/gdn_replayssm_spec_decode.py
  - https://github.com/vllm-project/vllm/blob/99dcd7600195c3a7ffaa23eb782591d04df1dc65/vllm/model_executor/layers/mamba/ops/selective_state_update_replayssm_output_only.py
  - https://github.com/vllm-project/vllm/blob/99dcd7600195c3a7ffaa23eb782591d04df1dc65/vllm/model_executor/layers/fla/ops/fused_recurrent_replayssm.py
  - https://github.com/vllm-project/vllm/blob/99dcd7600195c3a7ffaa23eb782591d04df1dc65/vllm/model_executor/layers/mamba/ops/selective_state_update_replayssm_state_and_output.py
  - https://github.com/vllm-project/vllm/blob/99dcd7600195c3a7ffaa23eb782591d04df1dc65/vllm/v1/attention/backends/mamba_attn.py
  - https://github.com/vllm-project/vllm/blob/99dcd7600195c3a7ffaa23eb782591d04df1dc65/vllm/model_executor/layers/mamba/mamba_utils.py
  - https://github.com/vllm-project/vllm/blob/99dcd7600195c3a7ffaa23eb782591d04df1dc65/vllm/model_executor/layers/mamba/mamba_mixer2.py
state: closed
tags:
  - vllm
  - vllm-pr
  - new-architectures
title: '#47576 [Kernel] ReplaySSM: cache SSM inputs instead of state for faster standard and speculative decode (Mamba2 + GDN)'
topic: new-architectures
url: https://github.com/vllm-project/vllm/pull/47576
created: '2026-11-11'
modified: '2026-11-11'
published: '2026-11-11'
pageLayout: default
slug: hinterland/papers/maps/prs/47576
permalink: https://aarnphm.xyz/hinterland/papers/maps/prs/47576.md
generator:
  quartz: v4.6.0
  hostedProvider: Cloudflare
  baseUrl: aarnphm.xyz
full: https://aarnphm.xyz/llms-full.txt
---
Config flags (all default off ⇒ baseline path byte-for-byte unchanged): · full dump: [[thoughts/inference/vllm-prs/47576|#47576]]

