From 5767f548c3438de17b548a47500f345498edc178 Mon Sep 17 00:00:00 2001
From: ArthurErlich <a.erlich@live.de>
Date: Sat, 5 Sep 2026 11:32:17 +0200
Subject: [PATCH] perf(llama-server): enable flash attention
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Cuts prefill time with no accuracy cost, per markaicode.com's
llama.cpp timeout writeup — directly helps the prefill-vs-stream-idle
contention issue fixed in the prior two commits.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
---
 docker-compose.yml | 1 +
 1 file changed, 1 insertion(+)

diff --git a/docker-compose.yml b/docker-compose.yml
index a925ad4..51f82f4 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -20,6 +20,7 @@ services:
       --n-gpu-layers ${LLAMA_GPU_LAYERS:-999}
       --ctx-size ${LLAMA_CTX_SIZE:-131072}
       --parallel ${LLAMA_PARALLEL:-2}
+      --flash-attn on
       --jinja
     # No published host port: llama-server is reached only via the omniroute
     # gateway on the ai-stack docker network now — see issue #15. Its