Challenges

GraphChallenge seeks input from diverse communities to develop graph challenges that take the best of what has been learned from groundbreaking efforts such as GraphAnalysis, Graph500, FireHose, MiniTri, and GraphBLAS to create a new set of challenges to move the community forward.

[NEW] Sparse Deep Neural Network Graph Challenge This challenge performs neural network inference on a variety of  sparse deep neural networks.

Static Graph Challenge: Subgraph Isomorphism This challenge seeks to identify a given sub-graph in a larger graph.

Streaming Graph Challenge: Stochastic Block Partition This challenge seeks to identify optimal blocks (or clusters) in a larger graph.

Note on static versus streaming graph challenges.  In static processing, given a large graph G the goal is to evaluate a function f(G).  In stateless streaming, given an additional smaller graph g, the goal is to evaluate the function f(g).  In stateful streaming, the goal is to evaluate a function f(G + g).  Stateful streaming is the focus of the streaming graph challenge.

Pre-Challenge: PageRank Pipeline This challenge is meant to test some of the specification and reference code approaches that might be used in subsequent challenges.  The community is encouraged to examine this specification and code and provide feedback as to how it can be improved for subsequent challenges.

Please use the following archival citations for the Graph Challenge when using the datasets and/or implementations made available via this Challenge:

Static Graph Challenge: Subgraph Isomorphism, Siddharth Samsi, Vijay Gadepally, Michael Hurley, Michael Jones, Edward Kao, Sanjeev Mohindra, Paul Monticciolo, Albert Reuther, Steven Smith, William Song, Diane Staheli, Jeremy Kepner, IEEE HPEC, 2017

Streaming Graph Challenge: Stochastic Block Partition, Edward Kao, Vijay Gadepally, Michael Hurley, Michael Jones, Jeremy Kepner, Sanjeev Mohindra, Paul Monticciolo, Albert Reuther, Siddharth Samsi, William Song, Diane Staheli, Steven Smith, IEEE HPEC, 2017

PageRank Pipeline Benchmark: Proposal for a Holistic System Benchmark for Big-Data Platforms, Patrick Dreher, Chansup Byun, Chris Hill, Vijay Gadepally, Bradley Kuszmaul, Jeremy Kepner, IEEE HPEC, 2016