From 391866785e44a6eae719dcdc7aad29b980a81e6a Mon Sep 17 00:00:00 2001
From: project6 <dev@project6.com>
Date: Fri, 7 Aug 2026 08:48:20 +0000
Subject: [PATCH] perf(config): match competitor Sub168's proven engine params
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

From competitor docker log analysis:
- max_model_len: 100000 → 256000 (competitor proven, 19259 GPU blocks)
- gpu_memory_utilization: 0.90 → 0.95 (competitor proven)
- max_num_batched_tokens: None → 4096 (competitor proven)
- enable_chunked_prefill: off → on (competitor proven, critical for 256K context)
- max_num_seqs stays at 2 (matches competitor)

Competitor Sub168 scored 60194 with these exact params before OOM at replay tail.
Our code has OOM-surviving advantages they lack (n>1 clamp, max_completion_tokens).

Docker log evidence: competitor's vLLM started with 19259 GPU blocks at 0.95 util,
ran for ~1h18m before OOM in layernorm.py x.float() at 31.72 GiB GPU.
---
 computility-run.yaml | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/computility-run.yaml b/computility-run.yaml
index e5da63a7..0be4999a 100644
--- a/computility-run.yaml
+++ b/computility-run.yaml
@@ -8,14 +8,16 @@ command:
     - --served-model-name
     - llm
     - --max-model-len
-    - '100000'
+    - '256000'
     - --gpu-memory-utilization
-    - '0.90'
+    - '0.95'
     - --trust-remote-code
     - -tp
     - '4'
     - --max-num-seqs
     - '2'
+    - --max-num-batched-tokens
+    - '4096'
     - --disable-log-requests
     - --disable-frontend-multiprocessing
     - --enforce-eager
@@ -25,6 +27,7 @@ command:
     - --reasoning-parser
     - qwen3
     - --enable-prefix-caching
+    - --enable-chunked-prefill
     - --dtype
     - half
 env: