From 41e918929f2fb581531b5aa6c42dd955f307ac82 Mon Sep 17 00:00:00 2001 From: Frank Date: Mon, 28 Sep 2026 12:03:56 -0700 Subject: [PATCH] frank-primary: enable prefix caching (E8 qualified: 9x-63x reuse TTFT, zero VRAM cost, cold path unchanged) --- scripts/frank-primary.env | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/scripts/frank-primary.env b/scripts/frank-primary.env index 0b598b5..6412479 100644 --- a/scripts/frank-primary.env +++ b/scripts/frank-primary.env @@ -44,4 +44,9 @@ R9V_FULL_MUTABLE_CACHE=0 # Braces in values get the same single-quote law as parentheses (source # parses assignment values). R9V_CHAT_TEMPLATE_FILE=/home/dank/serverAppConfigs/scripts/qwen-sharp-v22.4.0.jinja -R9V_CHAT_TEMPLATE_KWARGS='{"preserve_thinking":true}' \ No newline at end of file +R9V_CHAT_TEMPLATE_KWARGS='{"preserve_thinking":true}' +# Prefix caching (bench matrix E8, 2026-09-28): qualified on the q8 profile +# at 131072, zero VRAM cost (reuses freed KV blocks), cold path unchanged. +# Reuse-heavy traffic (harness retries/resends) hits from the third round: +# 32k TTFT 44.2 s -> 3.6 s, 131k 439 s -> 6.9 s. Reversible: flip to 0. +R9V_ENABLE_PREFIX_CACHING=1 \ No newline at end of file