Jeff Rasley

发表

HyperDrive: exploring hyperparameters with POP scheduling

Olatunji Ruwase, Yuxiong He, Rodrigo Fonseca, 2017, Middleware.

Low-latency Network Monitoring via Oversubscribed Port Mirroring

Rodrigo Fonseca, John B. Carter, Kanak Agarwal, 2014, ONS.

Efficient queue management for cluster scheduling

Srikanth Kandula, Rodrigo Fonseca, Konstantinos Karanasos, 2016, EuroSys.

Detecting latent cross-platform API violations

Yuriy Brun, Justin Cappos, Shriram Krishnamurthi, 2015, 2015 IEEE 26th International Symposium on Software Reliability Engineering (ISSRE).

DeepSpeed: System Optimizations Enable Training Deep Learning Models with Over 100 Billion Parameters

Olatunji Ruwase, Yuxiong He, Jeff Rasley, 2020, KDD.

Crowdsourcing from Scratch: A Pragmatic Experiment in Data Collection by Novice Requesters

Guan Wang, Connor Gramazio, Jeff Huang, 2015, HCOMP.

ZeRO: Memory Optimization Towards Training A Trillion Parameter Models pdf

Olatunji Ruwase, Yuxiong He, Samyam Rajbhandari, 2019, SC.

Rodrigo Fonseca, John B. Carter, Kanak Agarwal, 2014, SIGCOMM.

Planck: millisecond-scale monitoring and control for commodity networks

Rodrigo Fonseca, John B. Carter, Kanak B. Agarwal, 2015, SIGCOMM 2015.

Retaining sandbox containment despite bugs in privileged memory-safe code

Ivan Beschastnikh, Justin Cappos, Justin Samuel, 2010, CCS '10.

Accelerating Large Scale Deep Learning Inference through DeepCPU at Microsoft

Olatunji Ruwase, Minjia Zhang, Yuxiong He, 2019, OpML.

ZeRO-Infinity: Breaking the GPU Memory Wall for Extreme Scale Deep learning

Olatunji Ruwase, Samyam Rajbhandari, Jeff Rasley, 2021, SC21: International Conference for High Performance Computing, Networking, Storage and Analysis.

DeepSpeed-MoE: Advancing Mixture-of-Experts Inference and Training to Power Next-Generation AI Scale

Ammar Ahmad Awan, Samyam Rajbhandari, Jeff Rasley, 2022 .

ZeRO: Memory optimizations Toward Training Trillion Parameter Models

Olatunji Ruwase, Samyam Rajbhandari, Jeff Rasley, 2020, SC20: International Conference for High Performance Computing, Networking, Storage and Analysis.