From 0d09756430920a918e3df0c027a7822263c36f1f Mon Sep 17 00:00:00 2001
From: Frank <frank@dankware>
Date: Thu, 13 Aug 2026 22:46:55 -0700
Subject: [PATCH] dankclaw: revert MTP+ngram layering (slower); ubatch 8192

MTP+ngram benched 12.0-13.6 t/s (vs MTP alone 15.3-19.5) and is
parallel-safe, but slower - layering does not pay off. Back to MTP
alone. Test ubatch 8192 (== batch) using the VRAM headroom; watch for
OOM during load.
---
 scripts/frank-default-large.env | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/scripts/frank-default-large.env b/scripts/frank-default-large.env
index 6935cac..97adfce 100644
--- a/scripts/frank-default-large.env
+++ b/scripts/frank-default-large.env
@@ -14,7 +14,7 @@ PARALLEL=3
 #   draft-mtp     - MTP sidecar draft (a4lg/Qwen3.6-27B-MTP-ONLY-Q8_0.gguf)
 #   ngram-mod     - draftless ngram matching (no draft model, no shared ctx)
 # Benched 2026-08-14: see docs/frank-default-large-benchmark.md for the winner.
-SPEC_TYPE=draft-mtp,ngram-mod
+SPEC_TYPE=draft-mtp
 # Draft model file (only used when SPEC_TYPE contains 'draft-').
 # DFlash: Qwen3.6-27B-DFlash-Q8_0.gguf ; MTP: Qwen3.6-27B-MTP-ONLY-Q8_0.gguf
 DRAFT_FILE=Qwen3.6-27B-MTP-ONLY-Q8_0.gguf
@@ -39,7 +39,7 @@ CACHE_TYPE_V=q8_0
 # ubatch is the compute-graph driver; larger ubatch = more VRAM, faster prefill.
 # Re-verified 2026-08-14 during the spec swap (see benchmark doc).
 BATCH_SIZE=8192
-UBATCH_SIZE=4096
+UBATCH_SIZE=8192
 # Sampling settings — per HauhauCS Qwen3.6-27B model card.
 # Thinking-general: temp=1.0, top_p=0.95, top_k=20, min_p=0.
 # Clients may override per-request; these are the server-side defaults.