Picture for Gregory R. Ganger

Gregory R. Ganger

GraphPipe: Improving Performance and Scalability of DNN Training with Graph Pipeline Parallelism

Add code
Jun 24, 2024
Viaarxiv icon

Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep Learning

Add code
Aug 27, 2020
Figure 1 for Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep Learning
Figure 2 for Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep Learning
Figure 3 for Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep Learning
Figure 4 for Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep Learning
Viaarxiv icon

Accelerating Deep Learning by Focusing on the Biggest Losers

Add code
Oct 02, 2019
Figure 1 for Accelerating Deep Learning by Focusing on the Biggest Losers
Figure 2 for Accelerating Deep Learning by Focusing on the Biggest Losers
Figure 3 for Accelerating Deep Learning by Focusing on the Biggest Losers
Figure 4 for Accelerating Deep Learning by Focusing on the Biggest Losers
Viaarxiv icon

SysML: The New Frontier of Machine Learning Systems

Add code
May 01, 2019
Viaarxiv icon

MLtuner: System Support for Automatic Machine Learning Tuning

Add code
Mar 20, 2018
Figure 1 for MLtuner: System Support for Automatic Machine Learning Tuning
Figure 2 for MLtuner: System Support for Automatic Machine Learning Tuning
Figure 3 for MLtuner: System Support for Automatic Machine Learning Tuning
Figure 4 for MLtuner: System Support for Automatic Machine Learning Tuning
Viaarxiv icon