From 41305bf62cbfcb3660e57d3da579832e2da0f7f2 Mon Sep 17 00:00:00 2001
From: Vuong Hoang <vh@phasefinal.com>
Date: Wed, 8 Jul 2026 15:41:02 -0700
Subject: [PATCH] config(char-rp-gguf): Deckard reasoning seat to full 256K
 (GDN-hybrid KV cheap)

Deckard (Qwen3.5-27B) native ctx = 262144; GDN-hybrid arch (16 KV-caching layers) makes KV
cheap (8.7G q8_0 @ 256K vs Magidonia 10.9G @ 128K/40 dense layers). Bumped 40960 -> 262144;
GPU0 ~4G free (static -> stable). Canonical RP-reasoning samplers pending dvalin (card has
only generic base-Qwen3.5 profiles); empirical temp1.0/top_p0.95/top_k40/min_p0.05+DRY0.8 live.
---
 persistent-memory.md             | 7 +++++--
 stacks/char-rp-gguf/.env.example | 5 +++--
 stacks/char-rp-gguf/compose.yaml | 4 ++--
 3 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/persistent-memory.md b/persistent-memory.md
index 590902d..e761131 100644
--- a/persistent-memory.md
+++ b/persistent-memory.md
@@ -155,8 +155,11 @@ _As of 2026-07-08 — OFF-THE-SHELF INFERENCE STACK is the active work (home-tra
   field — beats even Magidonia char-rp 1.774 — BUT 7/30 explicit refusals → rejected: clean-but-refuses loses to
   clean-and-uncensored). **MEMORY-LAYOUT note:** Deckard (Qwen3.5-27B) is GDN-hybrid → only **16 KV-caching layers** →
   KV is TINY (1.4G @ 40K vs Magidonia's 10.9G @ 128K, 40 dense layers); Deckard native ctx = 256K, so the reasoning
-  seat's ctx scales nearly free (40K→128K ≈ +3G, →256K ≈ +7G). GPU0 layout: gen ~35G (256K/16-seq/util0.37), char-rp
-  Magidonia ~30G (128K), char-rp-reasoning Deckard ~21G (40K), ~11.5G free.
+  seat's ctx scales nearly free. **Deckard BUMPED to full 256K (262144) 2026-07-08** (KV 8.7G q8_0). GPU0 layout now:
+  gen ~35G (256K/16-seq/util0.37), char-rp Magidonia ~30G (128K/4-slot), char-rp-reasoning Deckard ~28G (256K/4-slot),
+  **~4G free** (all static allocations → stable, no OOM). **Canonical Deckard RP-reasoning samplers PENDING dvalin**
+  (card has only generic base-Qwen3.5 profiles, no RP/DRY canon); running empirical temp1.0/top_p0.95/top_k40/min_p0.05
+  + DRY0.8 (A/B-proven: 0 loops/0 refusals) until dvalin confirms — thread 01KX1Y7P.
   **MAX CONTEXT (2026-07-08):** char-rp **128K** (Magidonia FULL native 131072), char-rp-reasoning **40K** (QwQ
   native 40960, YaRN-free max), **q8_0 KV cache both** (near-lossless, ~2× ctx/GB; verified coherent, no Qwen
   gibberish). **Funded by gen util 0.40→0.37** (freed ~2.9G of gen's IDLE KV headroom — gen KV usage runs 0-2%,
diff --git a/stacks/char-rp-gguf/.env.example b/stacks/char-rp-gguf/.env.example
index acb12ee..dbb35cd 100644
--- a/stacks/char-rp-gguf/.env.example
+++ b/stacks/char-rp-gguf/.env.example
@@ -40,8 +40,9 @@ CHARRP_REASONING_PORT=8018
 # 0/30 refusals, clean slop; beat RpR-v4 (1/30 loop, no-DRY), Pantheon (7/30 refusals),
 # Snowdrop + Gembrain (template-incompatible with llama.cpp managed reasoning).
 CHARRP_REASONING_MODEL=rp/Qwen3.5-27B-Deckard-PKD-Heretic-Uncensored-Thinking.i1-Q5_K_M.gguf
-# Qwen3.5-27B native ctx is large; 40960 = a sane reasoning-seat cap. ~5G VRAM @ q8_0 KV.
-CHARRP_REASONING_CTX=40960
+# Deckard = Qwen3.5-27B GDN-hybrid (only 16 KV-caching layers → KV is CHEAP). Native ctx 262144
+# (256K, no YaRN). Full 256K @ q8_0 KV ≈ 8.5G VRAM (~0.033G/1K-tok, vs Magidonia ~0.085). Fits GPU0.
+CHARRP_REASONING_CTX=262144
 # KV cache dtype: q8_0 verified coherent; f16 if a future model gibbers.
 CHARRP_REASONING_KV_TYPE=q8_0
 # Thinking-token cap (concise scene-plan before the response). 300-500 is a good band.
diff --git a/stacks/char-rp-gguf/compose.yaml b/stacks/char-rp-gguf/compose.yaml
index 39443f6..46e4a3f 100644
--- a/stacks/char-rp-gguf/compose.yaml
+++ b/stacks/char-rp-gguf/compose.yaml
@@ -124,8 +124,8 @@ services:
       - "${CHARRP_REASONING_CTX:-40960}"
       - --flash-attn
       - on
-      # Deckard = Qwen3.5-27B (native ctx large); 40960 is a sane reasoning-seat cap. q8_0 KV
-      # verified coherent; flip to f16 in .env if a future model shows gibberish.
+      # Deckard = Qwen3.5-27B GDN-hybrid (16 KV-caching layers → KV cheap); native ctx 262144
+      # (256K, no YaRN). Full 256K @ q8_0 KV ≈ 8.5G. q8_0 verified coherent; f16 in .env if gibberish.
       - --cache-type-k
       - ${CHARRP_REASONING_KV_TYPE:-q8_0}
       - --cache-type-v