Pipeline Scheduling with Input Port Constraints for an FPGA-Based Biochemical Simulator

Authors:
Tomoya Ishimori;Hideki Yamada;Yuichiro Shibata;Yasunori Osana;Masato Yoshimi;Yuri Nishikawa;Hideharu Amano;Akira Funahashi;Noriko Hiroi;Kiyoshi Oguri
Affiliations:
Department of Computer and Information Sciences, Nagasaki University, Email: bio@pca.cis.nagasaki-u.ac.jp,;Department of Computer and Information Sciences, Nagasaki University, Email: bio@pca.cis.nagasaki-u.ac.jp,;Department of Computer and Information Sciences, Nagasaki University, Email: bio@pca.cis.nagasaki-u.ac.jp,;Department of Computer and Information Science, Seikei University,;Department of Information and Computer Science, Keio University,;Department of Information and Computer Science, Keio University,;Department of Information and Computer Science, Keio University,;Department of Biosciences and Informatics, Keio University,;EMBL-EBI,;Department of Computer and Information Sciences, Nagasaki University, Email: bio@pca.cis.nagasaki-u.ac.jp,
Venue:
ARC '09 Proceedings of the 5th International Workshop on Reconfigurable Computing: Architectures, Tools and Applications
Year:
2009

Citing 3
Cited 1

Algorithms for High-Level Synthesis

IEEE Design & Test
Cone Based Clustering for List Scheduling Algorithms

EDTC '97 Proceedings of the 1997 European conference on Design and Test
An Efficient List-Based Scheduling Algorithm for High-Level Synthesis

DSD '02 Proceedings of the Euromicro Symposium on Digital Systems Design

A heuristic scheduler for port-constrained floating-point pipelines

International Journal of Reconfigurable Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper discusses design methodology of high-throughput arithmetic pipeline modules for an FPGA-based biochemical simulator. Since limitation of data-input bandwidth caused by port constraints often has a negative impact on pipeline scheduling results, we propose a priority assignment method of input data which enables efficient arithmetic pipeline scheduling under given input port constraints. Evaluation results with frequently used rate-law functions in biochemical models revealed that the proposed method achieved shorter latency compared to ASAP and ALAP scheduling with random input orders, reducing hardware costs by 17.57 % and by 27.43 % on average, respectively.