article-journal

Communication-efficient ADMM-based distributed algorithms for sparse training
In large-scale distributed machine learning (DML), the synchronization efficiency of the distributed algorithm becomes a critical …
Guozheng Wang, Yongmei Lei, Yongwen Qiu, Lingfei Lou, Yixin Li
PDF Cite Code Dataset
Communication-efficient ADMM-based distributed algorithms for sparse training
2D-THA-ADMM: communication efficient distributed ADMM algorithm framework based on two-dimensional torus hierarchical AllReduce
Model synchronization refers to the communication process involved in large-scale distributed machine learning tasks. As the cluster …
Guozheng Wang, Yongmei Lei, Zeyu Zhang, Cunlu Peng
Cite Code Dataset
2D-THA-ADMM: communication efficient distributed ADMM algorithm framework based on two-dimensional torus hierarchical AllReduce
A Communication Efficient ADMM‑based Distributed Algorithm Using Two‑Dimensional Torus Grouping AllReduce
Large-scale distributed training mainly consists of sub-model parallel training and parameter synchronization. With the expansion of …
Guozheng Wang, Yongmei Lei, Zeyu Zhang, Cunlu Peng
PDF Cite Code Dataset
A Communication Efficient ADMM‑based Distributed Algorithm Using Two‑Dimensional Torus Grouping AllReduce
HSAC-ALADMM: an asynchronous lazy ADMM algorithm based on hierarchical sparse allreduce communication
The distributed alternating direction method of multipliers (ADMM) is an effective algorithm for solving large-scale optimization …
Dongxia Wang, Yongmei Lei, Jinyang Xie, Guozheng Wang
Cite Code Dataset
HSAC-ALADMM: an asynchronous lazy ADMM algorithm based on hierarchical sparse allreduce communication