From 941c42a04f336ecbce194b0c96248062f1b06d2a Mon Sep 17 00:00:00 2001
From: wanglin <2281216234@qq.com>
Date: Fri, 2 Oct 2026 19:53:25 +0800
Subject: [PATCH] =?UTF-8?q?v3:=20=E5=AF=B9=E9=BD=90=E6=88=90=E5=8A=9F?=
 =?UTF-8?q?=E9=85=8D=E7=BD=AE=20=E2=80=94=20=E8=A1=A5=20--enforce-eager=20?=
 =?UTF-8?q?=E4=B8=8E=20corex=20=E8=BF=90=E8=A1=8C=E5=BA=93=E7=8E=AF?=
 =?UTF-8?q?=E5=A2=83=E5=8F=98=E9=87=8F,=20max-model-len=20256K=20/=20gmu?=
 =?UTF-8?q?=200.95=20/=20seqs=204=20/=20bt=208192?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

v3: 对齐成功配置 — 补 --enforce-eager 与 corex 运行库环境变量, max-model-len 256K / gmu 0.95 / seqs 4 / bt 8192
---
 computility-run.yaml | 83 +++++++++++++++++++++++++++-----------------
 1 file changed, 52 insertions(+), 31 deletions(-)

diff --git a/computility-run.yaml b/computility-run.yaml
index b4f96c8..a929c2a 100644
--- a/computility-run.yaml
+++ b/computility-run.yaml
@@ -1,34 +1,55 @@
 concurrency: 1
 command:
-    - python3
-    - -m
-    - vllm.entrypoints.openai.api_server
-    - --model
-    - /model
-    - --served-model-name
-    - llm
-    - --max-model-len
-    - '100000'
-    - --gpu-memory-utilization
-    - '0.9'
-    - --trust-remote-code
-    - -tp
-    - '4'
-    - --max-num-seqs
-    - '8'
-    - --disable-log-requests
-    - --disable-frontend-multiprocessing
-    - --max-num-batched-tokens
-    - '16384'
-    - --enable-chunked-prefill
-    - --max-seq-len-to-capture
-    - '32768'
-    - --enable-auto-tool-choice
-    - --tool-call-parser
-    - qwen3_coder
-    - --reasoning-parser
-    - qwen3
-    - --enable-prefix-caching
+  - python3
+  - -m
+  - vllm.entrypoints.openai.api_server
+  - --model
+  - /model
+  - --served-model-name
+  - llm
+  - --max-model-len
+  - '256000'
+  - --gpu-memory-utilization
+  - '0.95'
+  - --trust-remote-code
+  - -tp
+  - '4'
+  - --max-num-seqs
+  - '4'
+  - --disable-log-requests
+  - --disable-frontend-multiprocessing
+  - --max-num-batched-tokens
+  - '8192'
+  - --enable-chunked-prefill
+  - --max-seq-len-to-capture
+  - '32768'
+  - --enforce-eager
+  - --enable-auto-tool-choice
+  - --tool-call-parser
+  - qwen3_coder
+  - --reasoning-parser
+  - qwen3
+  - --enable-prefix-caching
+  - --dtype
+  - half
 env:
-    - name: VLLM_ENGINE_ITERATION_TIMEOUT_S
-      value: 3600
+  - name: VLLM_ENGINE_ITERATION_TIMEOUT_S
+    value: 3600
+  - name: VLLM_ATTENTION_BACKEND
+    value: XFORMERS
+  - name: ENABLE_CUSTOM_IPC
+    value: 1
+  - name: PYTHONPATH
+    value: /usr/local/corex/lib/python3/dist-packages:/usr/local/corex/lib64/python3/dist-packages
+  - name: LD_LIBRARY_PATH
+    value: /usr/local/corex/lib64:/usr/local/openmpi/lib
+  - name: VLLM_COREX_FA2_LIBRARY
+    value: /usr/local/corex/lib64/libcorex_fa2.so
+  - name: VLLM_COREX_GDN_LIBRARY
+    value: /usr/local/corex/lib64/libcorex_gdn.so
+  - name: VLLM_COREX_MOE_LIBRARY
+    value: /usr/local/corex/lib64/libcorex_moe.so
+  - name: VLLM_REQUEST_METRICS_FILE
+    value: /tmp/vllm-request-metrics.jsonl
+  - name: VLLM_CACHE_BLOCK_SIZE
+    value: 16