A lightweight, effective and easy-to-extend inference runtime
Project description
Xlite轻量化推理运行时
介绍
xlite (GVirt前端):轻量级Transformer模型运行时,支持多样性算力协同,当前支持在昇腾硬件上高效运行。 xlite公开了Transformer运行所需的模型构图以及算子,所有算子基于昇腾AscendC/CCE开发。
背景与动机
在大模型推理场景中,传统的单流串行执行模式存在以下问题:
- 核间负载不均:不同AICORE之间的任务分配不均衡,部分核心闲置
- 资源浪费多:计算资源和传输资源利用率低,存在显著浪费
- 执行时间长:Host CPU下发算子开销大,造成严重的host bond问题
xlite通过以下技术手段解决上述问题:
- 多流并行:充分利用卡内资源,将单流串行改为多流并行执行
- 核间负载均衡:核间负载均衡,提升资源利用率
- CPU NPU协同:C++侧完全消除Python的GC、线程等干扰,简化Host tiling计算,去除小块内存申请释放及拷贝,消除Host bond
性能效果显著:
在GLM-4.7双机推理场景(40K输入、1K输出、prefix cache命中率约90%)下:
- TPOT时延降低17%~30%
- 吞吐提升13%~41%
详细性能数据参考 PR #7935。
软件架构
xlite已适配vllm_ascend,可通过xlite_graph_config配置快速使能xlite加速效果,使用方法参考官方指导文档。
xlite支持模型见模型列表。
快速开始
安装
# 安装vllm_ascend, 可参考https://github.com/vllm-project/vllm-ascend/blob/main/README.md
# 安装xlite
pip install xlite --extra-index https://download.pytorch.org/whl/cpu/
离线推理示例
import os
from vllm import LLM
# xlite默认支持decode-only模式, 可通过设置 "full_mode": True 使能full模式
model = LLM(model="path/to/Qwen3-32B", tensor_parallel_size=8, additional_config={"xlite_graph_config": {"enabled": True, "full_mode": True}})
outputs = model.generate("Hello, how are you?")
在线服务示例
vllm serve path/to/Qwen3-32B --tensor-parallel-size 8 --additional-config='{"xlite_graph_config": {"enabled": true, "full_mode": true}}'
开发指南
编译构建、容器镜像、源码安装等开发相关内容,请参考 开发指南
环境变量
xlite使用的环境变量及其配置方法,请参考 环境变量
目录结构
.
├── csrc/ - 轻量化运行时的核心代码
├── xlite/ - python代码,包括tools等
├── doc/ - 相关文档介绍
├── docker/ - 容器镜像Dockerfile
└── tests/ - 测试用例
致谢
本项目的核心算子初始版本和思路来自于华为终端小艺AI Infra团队的贡献,相关优化实现可参考论文:
《XY-Serve: End-to-End Versatile Production Serving for Dynamic LLM Workloads》 [ASPLOS 2026]
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distributions
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file xlite-0.1.0rc12.tar.gz.
File metadata
- Download URL: xlite-0.1.0rc12.tar.gz
- Upload date:
- Size: 205.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.11
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
15cc2725dbe821bf7c728300b6b3b5dfa8c1ae691af32a62c16d983dd2c48bd9
|
|
| MD5 |
dcf9a3106f053f50e54e22cb80b65e34
|
|
| BLAKE2b-256 |
e86cc4acf9680ee2c8647694e4ce8682dc7064f508ac641177396a3a18aebda5
|
File details
Details for the file xlite-0.1.0rc12-cp312-cp312-manylinux2014_aarch64.whl.
File metadata
- Download URL: xlite-0.1.0rc12-cp312-cp312-manylinux2014_aarch64.whl
- Upload date:
- Size: 838.7 kB
- Tags: CPython 3.12
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.11
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
b816d37257b132a8c0604885bbe99184d65ce1aaa91571d5166d97ee5bef4546
|
|
| MD5 |
11d4fe3c212f9de74a7ed891a7227bcf
|
|
| BLAKE2b-256 |
9a14c5265815981e61aa3b8ab508a1e300e7fba847f2fc20db630e56b6023b2d
|
File details
Details for the file xlite-0.1.0rc12-cp311-cp311-manylinux2014_aarch64.whl.
File metadata
- Download URL: xlite-0.1.0rc12-cp311-cp311-manylinux2014_aarch64.whl
- Upload date:
- Size: 821.6 kB
- Tags: CPython 3.11
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.11
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
cccb74688f6acb9cc219290c3a04b6005b81dba941b9d63c79bd52d02854fc8a
|
|
| MD5 |
9664f4c84060e558acacec359bd86200
|
|
| BLAKE2b-256 |
823b920f7309f5963655e14a685fe9f3d88e014e24024bd2de03167ea7012199
|