SC25 Home
Close

Session

Paper
: Scheduling, Tiling, and Parallelism
Session Chair
Zhiling Lan
University of Illinois Chicago
Argonne National Laboratory (ANL)
Event Type
Paper
TimeThursday, 20 November 202510:30am - 12:00pm CST
Location275
Tags
HPC for Machine Learning
State of the Practice
System Software and Cloud Computing
Recordings
Livestreamed
Recorded
Registration Categories
TP
Similar Sessions
Machine Learning: Training at Scale 2
Graph Neural Networks and Training
Machine Learning: Inference and Serving
Presentations
10:30am - 10:52am CSTSIREN: Software Identification and Recognition in HPC Systems
Authors
Thomas Jakobsche
Fredrik Robertsén
Jessica R. Jones
Utz-Uwe Haus
Florina M. Ciorba
State of the Practice
10:52am - 11:15am CSTHypertron: Efficiently Scaling Large Models by Exploring High-Dimensional Parallelization Space
Authors
Shigang Li
Jingkun Dong
Jihao Chen
Zhi Ma
Zhongzhe Hu
HPC for Machine Learning
11:15am - 11:38am CSTUltraAttn: Efficiently Parallelizing Attention Through Hierarchical Context-Tiling
Authors
Haoyu Yang
Zan Zong
Yuyang Jin
Kinman Lei
Jiaao He
Qigang Yang
Jidong Zhai
HPC for Machine Learning
11:38am - 12:00pm CSTTianheEngine: Hierarchy-Aware Adaptive Partitioning System for Trillion-Scale Graph Processing
Authors
Xinbiao Gan
Tiejun Li
Yiqi Wang
Qiang Zhang
Yongming Yi
Chunye Gong
Jie Liu
Kai Lu
System Software and Cloud Computing