vllm-project/vllm-ascend

[Community Test] Qwen3-32B-W4A4 Memcache 池化后端验证

Open

#11,745 opened on Jul 9, 2026

View on GitHub
 (0 comments) (0 reactions) (0 assignees)C++ (1,318 forks)github user discovery
help wanted

Repository metrics

Stars
 (2,180 stars)
PR merge metrics
 (Avg merge 4d 5h) (559 merged PRs in 30d)

Description

Source: 推理众测任务书-vllm.xlsx row 10. This issue tracks one community testing task for vLLM-Ascend.

Task Information

  • 分类: vllm
  • 任务名称: vLLM-Ascend支持Memcache池化后端
  • 任务类别: 资料验证和部署测试
  • 蓝区镜像: quay.io/ascend/vllm-ascend:v0.18.0

https://docs.vllm.ai/projects/ascend/en/latest/user_guide/feature_guide/kv_pool.html

任务概述

在A2环境上部署Qwen3-32B量化模型，开启Memcache池化，并跑通精度测试

核心要求

任务范围：部署vllm、vllm-ascend，拉起Memcache池化+VLLM服务，使用aisbench进行模型精度验证。 精度与基线误差1%以内，精度达标。

验收标准 / 交付件

交付件 测试报告 交付标准 完成模型功能验证，精度验证

测试方法

（1）进入容器 （2）安装版本配套 （3）安装memcache_hybrid （4）启动MetaService 服务 （5）拉起vLLM-Ascend服务 （6）配置aisbench （7）aisbench发起精度测试

对接信息

  • 技术对接人: 应宇轩  郑志崇  崔青
  • PAE/小巧灵: 邹长江 00888932 宋洋 00835984
  • 工作量: 暂无
  • 任务书路径: 任务书\vllm任务书\修改任务书\任务书-xyq1.docx

任务问题和需求

memcache_hybrid无单独安装包，并且依赖cann，当前cann版本8.5.1不包含memcache_hybird

Contributor guide

Research direction
Tech stack
Domain
Issue type
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Prerequisites
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.