121 citations · 330 across the 20 of their papers we have counts for
7 papers · 1 filter
Fused3S: Fast Sparse Attention on Tensor Cores
Zitong Li, Aparna Chandramowlishwaran
Sparse attention is a core building block in many leading neural network models, from graph-structured learning to sparse sequence modeling. It can be decomposed into a sequence of…
Lessons Learned on MPI+Threads Communication
Rohit Zambre, Aparna Chandramowlishwaran
Hybrid MPI+threads programming is gaining prominence, but, in practice, applications perform slower with it compared to the MPI everywhere model. The most critical challenge to the…
Brief Announcement: On the Limits of Parallelizing Convolutional Neural Networks on GPUs
Behnam Pourghassemi, Chenghao Zhang, Joo Hwan Lee +1
GPUs are currently the platform of choice for training neural networks. However, training a deep neural network (DNN) is a time-consuming process even on GPUs because of the massiv…
How I Learned to Stop Worrying About User-Visible Endpoints and Love MPI
Rohit Zambre, Aparna Chandramowlishwaran, Pavan Balaji
MPI+threads is gaining prominence as an alternative to the traditional MPI everywhere model in order to better handle the disproportionate increase in the number of cores compared…
Breaking Band: A Breakdown of High-performance Communication
Rohit Zambre, Megan Grodowitz, Aparna Chandramowlishwaran +1
The critical path of internode communication on large-scale systems is composed of multiple components. When a supercomputing application initiates the transfer of a message using…
Parallel Performance-Energy Predictive Modeling of Browsers: Case Study of Servo
Rohit Zambre, Lars Bergstrom, Laleh Aghababaie Beni +1
Mozilla Research is developing Servo, a parallel web browser engine, to exploit the benefits of parallelism and concurrency in the web rendering pipeline. Parallelization results i…
Scalable Communication Endpoints for MPI+Threads Applications
Rohit Zambre, Aparna Chandramowlishwaran, Pavan Balaji
Hybrid MPI+threads programming is gaining prominence as an alternative to the traditional "MPI everywhere'" model to better handle the disproportionate increase in the number of co…