| 注册
首页|期刊导航|计算机科学技术学报(英文版)|SwFormer:Enabling Faster Foundation Models on New Sunway Supercomputer via Holistic Kernel Tiling and Scheduling

SwFormer:Enabling Faster Foundation Models on New Sunway Supercomputer via Holistic Kernel Tiling and Scheduling

吴若晗 朱先语 陈俊仕 安虹

计算机科学技术学报(英文版)2025,Vol.40Issue(6):1512-1529,18.
计算机科学技术学报(英文版)2025,Vol.40Issue(6):1512-1529,18.DOI:10.1007/s11390-025-4761-0

SwFormer:Enabling Faster Foundation Models on New Sunway Supercomputer via Holistic Kernel Tiling and Scheduling

SwFormer:Enabling Faster Foundation Models on New Sunway Supercomputer via Holistic Kernel Tiling and Scheduling

吴若晗 1朱先语 1陈俊仕 1安虹1

作者信息

  • 折叠

摘要

关键词

deep learning/foundation model/Sunway architecture/fine-grained tiling/operator scheduling

Key words

deep learning/foundation model/Sunway architecture/fine-grained tiling/operator scheduling

引用本文复制引用

吴若晗,朱先语,陈俊仕,安虹..SwFormer:Enabling Faster Foundation Models on New Sunway Supercomputer via Holistic Kernel Tiling and Scheduling[J].计算机科学技术学报(英文版),2025,40(6):1512-1529,18.

基金项目

The work was partly supported by the Strategic Priority Research Program of the Chinese Academy of Sciences under Grant No.XDB0500102.Computing resources of this work are partly financially supported by Laoshan Laboratory under Grant No.LSKJ202300305. ()

计算机科学技术学报(英文版)

1000-9000

访问量1
|
下载量0
段落导航相关论文