IEEE Transactions on Services Computing最新文献_第10页

TPI-LLM: Serving 70B-Scale LLMs Efficiently on Low-Resource Mobile Devices TPI-LLM：在低资源移动设备上高效地服务70b级llm

IF 5.8 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-08-08 DOI: 10.1109/TSC.2025.3596892

Zonghang Li;Wenjiao Feng;Mohsen Guizani;Hongfang Yu

{"title":"TPI-LLM: Serving 70B-Scale LLMs Efficiently on Low-Resource Mobile Devices","authors":"Zonghang Li;Wenjiao Feng;Mohsen Guizani;Hongfang Yu","doi":"10.1109/TSC.2025.3596892","DOIUrl":"https://doi.org/10.1109/TSC.2025.3596892","url":null,"abstract":"LLM serving is shifting from cloud to edge due to privacy concerns over user interaction data. However, mobile devices struggle with very limited computing power and memory, requiring collaboration among multiple devices to run LLM apps. The mainstream solution, pipeline parallelism, is inefficient for such cases because mobile devices typically run only one inference task at a time. This article argues that tensor parallelism, despite its high communication cost, can better fit such scenarios. We introduce TPI-LLM, a compute and memory-efficient tensor parallel inference system designed to run 70B-scale LLMs on low-resource mobile devices. It keeps sensitive raw data local on users’ devices and employs a sliding window memory scheduler to dynamically manage layer weights. It overlaps disk I/O with computation and communication, enabling efficient operation of large models on memory-limited devices. Extensive experiments show that TPI-LLM reduces token latency by 80%–90% compared to Transformers, Accelerate, and Galaxy. It also cuts the peak memory footprint by 90%, requiring just 3.1 GiB of memory for 70B-scale models.","PeriodicalId":13255,"journal":{"name":"IEEE Transactions on Services Computing","volume":"18 5","pages":"3321-3333"},"PeriodicalIF":5.8,"publicationDate":"2025-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145248073","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

IRS-Assisted Hyperspectral Image Processing in Satellite Edge Computing Services 卫星边缘计算服务中irs辅助高光谱图像处理

IF 8.1 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-08-04 DOI: 10.1109/tsc.2025.3595166

Xiaoteng Yang, Jie Feng, Lei Liu, Qingqi Pei, Shahid Mumtaz, Keqin Li, Schahram Dustdar

引用次数: 0

E-Log: Fine-Grained Elastic Log-Based Anomaly Detection and Diagnosis for Databases E-Log：基于细粒度弹性日志的数据库异常检测与诊断

IF 8.1 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-08-01 DOI: 10.1109/tsc.2025.3594870

Lingzhe Zhang, Tong Jia, Xinyu Tan, Xiangdong Huang, Mengxi Jia, Hongyi Liu, Zhonghai Wu, Ying Li

引用次数: 0

Multi-Objective Unlearning in Recommender Systems via Preference Guided Pareto Exploration 基于偏好引导Pareto探索的推荐系统多目标遗忘

IF 8.1 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-07-31 DOI: 10.1109/tsc.2025.3593906

Yuyuan Li, Yizhao Zhang, Weiming Liu, Xiaohua Feng, Zhongxuan Han, Chaochao Chen, Chenggang Yan

引用次数: 0

Data-Locality-Aware Task Assignment and Scheduling for Distributed Job Executions 分布式作业执行的数据位置感知任务分配和调度

IF 8.1 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-07-31 DOI: 10.1109/tsc.2025.3594158

Hailiang Zhao, Xueyan Tang, Peng Chen, Jianwei Yin, Shuiguang Deng

引用次数: 0

FrigateBird: Decoupling Metadata/Data Services for Continuously Fast Object Storage 为持续快速对象存储解耦元数据/数据服务

IF 8.1 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-07-31 DOI: 10.1109/tsc.2025.3594340

Yiming Zhang, Kekun Hu, Gang Dong, Hailiang Chen, Rengang Li

引用次数: 0

ARMM: Sandwich-Attack Resilient Automated Market Maker ARMM：三明治攻击弹性自动做市商

IF 8.1 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-07-30 DOI: 10.1109/tsc.2024.3376195

Shunrong Jiang, Jingwei Chen, Jinpeng Li, Zhihui Liu, Yonggang Li

引用次数: 0

An Effective Multi-Scale Contrastive Learning System for Online Group Recommendation Services in Event-Based Social Networks 基于事件的社交网络中在线群组推荐服务的多尺度对比学习系统

IF 8.1 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-07-28 DOI: 10.1109/tsc.2025.3593346

Xiaomei Huang, Zhiheng Zhou, Jianjun Li, Neal N. Xiong, Yugen Yi, Jin Liu, Guoqiong Liao

引用次数: 0

Deep Learning to Hash for Time-aware QoS Prediction Based on VQ-VAE 基于VQ-VAE的时间感知QoS预测的深度学习哈希算法

IF 8.1 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-07-25 DOI: 10.1109/tsc.2025.3592409

Lingzhen Kong, Xiyuan Hu, Lianyong Qi, Xiaolong Xu, Yiwen Zhang, Lina Yao, Xuyun Zhang

引用次数: 0

CATScaler: A Convolution-Augmented Transformer Scaling Framework for Cloud-Native Applications CATScaler：一个用于云原生应用的卷积增强变压器缩放框架

IF 8.1 2区计算机科学

IEEE Transactions on Services Computing Pub Date : 2025-07-24 DOI: 10.1109/tsc.2025.3592383

Fan'an Meng, Hongjun Dai, Guoqing Cong, Bo Zhu, Hailiang Zhao

引用次数: 0