Blink: Fast and Generic Collectives for Distributed ML

no code implementations • 11 Oct 2019 • Guanhua Wang, Shivaram Venkataraman, Amar Phanishayee, Jorgen Thelin, Nikhil Devanur, Ion Stoica

Model parameter synchronization across GPUs introduces high overheads for data-parallel training at scale.

Image Classification

PipeDream: Fast and Efficient Pipeline Parallel DNN Training

1 code implementation • 8 Jun 2018 • Aaron Harlap, Deepak Narayanan, Amar Phanishayee, Vivek Seshadri, Nikhil Devanur, Greg Ganger, Phil Gibbons

PipeDream is a Deep Neural Network(DNN) training system for GPUs that parallelizes computation by pipelining execution across multiple machines.

Distributed, Parallel, and Cluster Computing

Truth and Regret in Online Scheduling

no code implementations • 1 Mar 2017 • Shuchi Chawla, Nikhil Devanur, Janardhan Kulkarni, Rad Niazadeh

The service provider's goal is to implement a truthful online mechanism for scheduling jobs so as to maximize the social welfare of the schedule.

Scheduling

Budget Constraints in Prediction Markets

no code implementations • 7 Oct 2015 • Nikhil Devanur, Miroslav Dudík, Zhiyi Huang, David M. Pennock

We give a detailed characterization of optimal trades under budget constraints in a prediction market with a cost-function-based automated market maker.