SC19 Proceedings

The International Conference for High Performance Computing, Networking, Storage, and Analysis

HPC Graph Toolkits and the GraphBLAS Forum


Authors: Antonino Tumeo (Pacific Northwest National Laboratory (PNNL)), José Moreira (IBM Corporation), John Feo (Pacific Northwest National Laboratory (PNNL)), Mahantesh Halappanavar (Pacific Northwest National Laboratory (PNNL)), Aydin Buluç (Lawrence Berkeley National Laboratory), Timothy Mattson (Intel Corporation)

Abstract: Government agencies, industry, and academia are demanding a new generation of tools to efficiently solve large scale analytics problems in a variety of business, scientific and national security applications. This BoF aims at gathering the community of people interested in frameworks and workflows for large scale graph analytics, surveying the current approaches, identifying new challenges and opportunities, and laying a path toward future interoperable infrastructures. As in previous editions, we will invite the GraphBLAS community to participate in a discussion of the current state and evolution of GraphBLAS, with the goal of developing requirements and recommendations for future tools.

Long Description: Graph analytics represents a growing area of activity in government, industry and academia. Large scale graph problems require ever growing compute power and impose significant requirements on modern supercomputing architectures. Graph analytics are typically done using graph toolkits and libraries and a BOF on this topic falls right in the area of interested of the SC conference.

This BOF was held before at SC17 and SC18 and we had approximately 100 attendees. Audience participating was intense, with good response to the short presentations and vibrant discussions. We expect the same in SC19.

This BOF aims at gathering the community of people interested in frameworks and workflows for large-scale graph analytics, with the objective to discuss the current situation, identifying new challenges and opportunities, and laying the path towards future and interoperable infrastructures.

We particularly want to address the new and upcoming challenges in large scale graph analytics applications: the support for streaming graphs, the ability to deal with attributed graphs (that couple graphs with dense tables of attributes), the need to integrate the graph methods within broader machine learning frameworks, and the need to better support irregular data structures and graph methods in scientific simulation frameworks. Current and future graph toolkits will have to evolve to handle these new requirements and domains.

This BOF also has the specific focus to discuss GraphBLAS as one of the key components of emerging graph toolkits, and bring together the GraphBLAS community with the wider graph algorithms community. In particular, there will be space to discuss the existing implementations of GraphBLAS, identifying pros and cons, and suggesting potential changes for GraphBLAS 2.0. Additionally, the BOF will seek input from the current and potential GraphBLAS user community for key design patterns and requested features.

Our lineup of speakers will touch key themes such as applications, use cases, programming models, application programming interfaces and libraries, data structures and algorithms, and integration of tools, including common data structures, data storages, and data frames. The discussions and panels will also specifically delve into the dynamic runtime technologies needed to make graph toolkits and/or sparse linear algebra approaches execute efficiently. While remaining vendor agnostic, we expect to touch also architectural requirements and architectural support for such runtime technologies and workloads.

The main outcomes of this integrated BOF will be to effectively gather the community around graph algorithms and graph toolkits, including GraphBLAS, providing an initial identification of abstractions and potential use cases of graph toolkits, and the identification of common layers and interfaces (e.g., algorithms, graph primitives, runtime functionalities) across components and tools. In the context of GraphBLAS, the BOF will continue to provide information on the essential design patterns and requirements for GraphBLAS 2.0.


URL: https://hpc.pnl.gov/BOF/


Back to Birds of a Feather Archive Listing