1

Assassyn: A Unified Abstraction for Architectural Simulation and Implementation
The continuous growth of on-chip transistors driven by technology scaling urges architecture developers to design and implement novel …
Jian Weng, Boyang Han, Derui Gao, Ruijie Gao, Wanning Zhang, Ceyu Xu, 辛继灏, Yangzhixin Luo, Lisa Wu Wills, Marco Canini
引用
Global QSGD: Allreduce-Compatible Quantization for Distributed Learning with Theoretical Guarantees
The increase of deep learning models and dataset sizes make training time-consuming, while heavy communication, primarily due to …
辛继灏, Marco Canini, Peter Richtárik, Samuel Horváth
PDF 引用 DOI
Immediate Communication for Distributed AI Tasks
Large AI models have necessitated efficient communication strategies across multi-GPU and multi-node infrastructures due to their …
辛继灏, Seongjong Bae, KyoungSoo Park, Marco Canini, Changho Hwang
PDF 引用
Kimad: Adaptive Gradient Compression with Bandwidth Awareness
In distributed training, communication often emerges as a bottleneck. In response, we introduce Kimad, a solution that offers adaptive …
辛继灏, Ivan Ilin, Shunkang Zhang, Marco Canini, Peter Richtárik
PDF 引用 DOI
OpenFunction for Software Defined IoT
The recent surge in the prosperity of the Internet of Things (IoT) has been attracting an increasing number of researchers and experts …
Nian Xue, Daojing Guo, Jie Zhang, 辛继灏, Zhen Li, Xin Huang
PDF 引用 DOI