Department of Computer Science
The University of Texas at Austin
yishanlu AT utexas DOT edu
I am interested in making parallel computing more accessible to programmers in different application fields so that more people can enjoy the performance gain from parallel computing. This involves programming languages, compilers, computer architectures and understanding of applications.
I work with Prof. Keshav Pingali in Intelligent Software Systems research group.
I would like to come up with programming models suitable for constrained optimization/constraint satisfaction algorithms. I am exploring such models by using Galois framework to parallelize graph algorithms from the following application fields.
Electronic Design Automation (EDA): Hardware designers use EDA tools to synthesize circuit descriptions to final layouts for manufacturing. During the synthesis process, functionality should remain the same; timing constraints should be respected; and area, power consumption should be optimized. Since EDA algorithms work on circuits, naturally represented as graphs, and the parallelism is hard to know statically, how to parallelize EDA algorithms becomes challenging.
Graph analytics: Graph analytics are important to discover patterns in graph data. For example, k-truss decomposition can cluster nodes in a graph based on their shared links. Since graph analytics can be realized with various combinations of programming models and implementations, the tradeoff between programming productivity and implementation efficiency needs to be addressed.
Graph pattern matching: Certain properties emerge from groups but are not observed from constituent individuals. To enable such analyses on graph data, we can locate interesting regions using graph pattern matching, i.e. matching for a query graph to subgraphs in a data graph. By working directly with (sparse) graphs, we may be able to match graph patterns more efficiently than with matrix/tensor forms. Besides, this approach allows the matching to be fuzzy to various degrees. Graph pattern matching can be applied to network intrusion detection, bioinformatics and verification in VLSI, etc.
Chad Voegele, Yi-Shan Lu, Sreepathi Pai, Keshav Pingali. Parallel Triangle Counting and k-Truss Identification using Graph-centric Methods. In IEEE/DARPA/Amazon Graph Challenge 2017 at IEEE High Performance Extreme Computing Conference (HPEC), Waltham, MA, September 12-14, 2017. (Graph Challenge Champion) [pdf][slides][DOI]
Lu Y.-S., Pingali K. (2018) Can parallel programming revolutionize EDA tools?. In Reis A., Drechsler R. (eds) Advanced Logic Synthesis. Springer, Cham. [DOI]