UWB Logo

Computer Science and Software Engineering Capstone Presentations

Summer Quarter

August 21, 2020

Elias Alabsie

"MASS C++ Enhancement: Porting parallel I/O and Graph"

(UWB CSS Faculty Research)

Faculty Advisor: Dr. Munehiro Fukuda

Abstract

I worked as a developer for Multi-Agent Spatial Simulation (MASS) C++ library in the Summer of 2020, at University of Washington, DSLAB, with Professor Fukuda, Munehiro. MASS is a highly scalable, multi-agent based, parallel-computing library for spatial simulation over a cluster of computing nodes. MASS composes a user application of distributed arrays and multi-agents, each representing an individual simulation place or an active entity.

The Scalability and efficiency of MASS is attributed to its capability to distributed workload over a cluster of computing nodes and executing the work in parallel. The library uses machine independent indices to address each place over the simulation space.

As of now, the library offers implementations with three programming languages: Java, C++, and CUDA.

The MASS java library has more features than the MASS C++ library. To narrow this gap, I was assigned to port the Parallel I/O and Graph features of the MASS java library into the MASS C++ library. Porting these features will eventually increase the productivity of MASS C++ library users by using these additional functionalities for their application development. Making the MASS C++ library rich in features will help the library users to create feature rich application with less time and effort in return.

To implement these features, first I studied the documentation and implementation details of the MASS Java library's parallel I/O and graph components. Then I designed, optimized, and implemented a C++ language solution to port these functionalities into MASS C++ library.

As a result of this work, now MASS C++ has to offer additional APIs for its users.

Updated August 20, 2020, 00:44