Totem

From NetSysLab

Jump to: navigation, search

The goal of this project is to understand the challenges in supporting graph algorithms on commodity, hybrid platforms; platforms that consist of processors optimized for sequential processing and accelerators optimized for massively-parallel processing.

This will fill the gap between current graph processing platforms that are either expensive (e.g., supercomputers) or inefficient (e.g., commodity clusters). Our hypothesis is that hybrid platforms (e.g., GPU-supported large-memory nodes and GPU supported clusters) can bridge the performance-cost chasm, and offer an attractive graph-processing solution for many graph-based applications such as social networks and web analysis.

Contents

People

Abdullah Gharaibeh
Scott Sallinen
Tahsin Reza
Tanuj kr Aasawat
Elizeu Santos-Neato
Lauro Beltrão Costa
Matei Ripeanu

Download

Git repository
User guide v0.1

Talks

[2] The Energy Case for Graph Processing on Hybrid CPU and GPU Systems, GPU Technology Conference (GTC), San Jose, CA, March 2014, talk
[1] TOTEM: Accelerating Graph Processing on Hybrid CPU+GPU Systems, GPU Technology Conference (GTC), San Jose, CA, March 2013 talk slides

Publications

[7] Efficient Large-Scale Graph Processing on Hybrid CPU and GPU Systems, Abdullah Gharaibeh, Tahsin Reza, Elizeu Santos-Neto, Lauro Beltrão Costa, Scott Sallinen, Matei Ripeanu. Submitted to ACM Transactions on Parallel Computing (ToPC) on December 2013. Revised December 2014.TechnicalReport
[6] Accelerating Direction-Optimized Breadth First Search on Hybrid Architectures, Scott Sallinen, Abdullah Gharaibeh, Matei Ripeanu, 13th International Workshop on Algorithms, Models and Tools for Parallel Computing on Heterogeneous Platforms (HeteroPar’15), August 2015 (acceptance rate: 31%) pdf slides
[5] Exploring Hybrid Hardware and Data Placement Strategies for the Graph 500 Challenge, Scott Sallinen, Daniel Borges, Abdullah Gharaibeh, Matei Ripeanu, poster at SC’14, New Orleans, LA, November 2014 pdf poster
[4] Parallel Clustering Coefficient Computation using GPUs, Tahsin Reza, Tanuj Kr Aasawat, Matei Ripeanu, poster at SC’14, New Orleans, LA, November 2014 pdf poster
[3] The Energy Case for Graph Processing on Hybrid CPU and GPU Systems, Abdullah Gharaibeh, Elizeu Santos-Neto, Lauro Beltrão Costa, Matei Ripeanu. Workshop on Irregular Applications: Architectures & Algorithms (IA3) in conjunction with SC13, Denver, Colorado USA, November 2013 pdf
[2] On Graphs, GPUs, and Blind Dating: A Workload to Processor Matchmaking Quest, Abdullah Gharaibeh, Lauro Beltrão Costa, Elizeu Santos-Neto, Matei Ripeanu, IEEE International Parallel & Distributed Processing Systems (IPDPS), Boston, MA, May 2013 (acceptance rate: 21%) pdf slides
[1] A Yoke of Oxen and a Thousand Chickens for Heavy Lifting Graph Processing, Abdullah Gharaibeh, Lauro Beltrão Costa, Elizeu Santos-Neto, Matei Ripeanu, IEEE/ACM International Conference on Parallel Architectures and Compilation Techniques (PACT 2012). Minneapolis, MN September 2012. (acceptance rate 19%) pdf slides

Related Projects

STINGER