Skip to main content

A lightweight, effective and easy-to-extend inference runtime

Project description

xlite logo

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

image

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

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


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

xlite-0.1.0rc9.dev210-cp312-cp312-manylinux2014_x86_64.whl (784.4 kB view details)

Uploaded CPython 3.12

xlite-0.1.0rc9.dev210-cp312-cp312-manylinux2014_aarch64.whl (744.2 kB view details)

Uploaded CPython 3.12

xlite-0.1.0rc9.dev210-cp311-cp311-manylinux2014_x86_64.whl (780.0 kB view details)

Uploaded CPython 3.11

xlite-0.1.0rc9.dev210-cp311-cp311-manylinux2014_aarch64.whl (743.6 kB view details)

Uploaded CPython 3.11

File details

Details for the file xlite-0.1.0rc9.dev210-cp312-cp312-manylinux2014_x86_64.whl.

File metadata

File hashes

Hashes for xlite-0.1.0rc9.dev210-cp312-cp312-manylinux2014_x86_64.whl
Algorithm Hash digest
SHA256 f77debc259fb706a9d1e253996e8ce2b5c0943b3c626cbe57638c3e2346e7986
MD5 31b9107a355d331aab0a94e331db5741
BLAKE2b-256 0eec6f099419dee996d2e2b00f007426c7ff8974539be32992f5afde53745c7c

See more details on using hashes here.

File details

Details for the file xlite-0.1.0rc9.dev210-cp312-cp312-manylinux2014_aarch64.whl.

File metadata

File hashes

Hashes for xlite-0.1.0rc9.dev210-cp312-cp312-manylinux2014_aarch64.whl
Algorithm Hash digest
SHA256 26b15763232541018b042f9077e8d7ae0079de05695031b01895eb605dbc5f07
MD5 c4e8987078ffa9fdd1c4d7ae1413fffb
BLAKE2b-256 0a3eaf4f44745d8a3c7657a1fb4a2ee463ce89557d1ea6e8fa9b85a4d91b1ad8

See more details on using hashes here.

File details

Details for the file xlite-0.1.0rc9.dev210-cp311-cp311-manylinux2014_x86_64.whl.

File metadata

File hashes

Hashes for xlite-0.1.0rc9.dev210-cp311-cp311-manylinux2014_x86_64.whl
Algorithm Hash digest
SHA256 26caa721bc85f020987d929633f5ee58ee0043126d21c2b327acbbbf942d9222
MD5 ba26c3d68928f86be81bd704b1e00c64
BLAKE2b-256 779f84b345b4fd21d4add49e3255308c7eb44bf0d8b462f92c8813027960d65a

See more details on using hashes here.

File details

Details for the file xlite-0.1.0rc9.dev210-cp311-cp311-manylinux2014_aarch64.whl.

File metadata

File hashes

Hashes for xlite-0.1.0rc9.dev210-cp311-cp311-manylinux2014_aarch64.whl
Algorithm Hash digest
SHA256 b228633a76d411bbb7a3a0a00b91a0af5bd0ff29261aa4eb4a38c08093188f69
MD5 6b28654546c98d5ee396ab44657d4b7b
BLAKE2b-256 87da8f4bd750bffeb5a650855a8bc567295fd3b55672820d092dc7175c2982bd

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page