Achieving efficient realization of Kalman Filter on CGRA through algorithm-architecture co-design

In this paper, we present efficient realization of Kalman Filter (KF) that can achieve up to 65% of the theoretical peak performance of underlying architecture platform. KF is realized using Modified Faddeeva Algorithm (MFA) as a basic building block due to its versatility and REDEFINE Coarse Graine...

Full description

Saved in:
Bibliographic Details
Main Authors: Merchant, Farhad, Vatwani, Tarun, Chattopadhyay, Anupam, Raha, Soumyendu, Nandy, Soumitra Kumar, Narayan, Ranjani
Other Authors: School of Computer Science and Engineering
Format: Conference or Workshop Item
Language:English
Published: 2020
Subjects:
Online Access:https://hdl.handle.net/10356/142291
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-142291
record_format dspace
spelling sg-ntu-dr.10356-1422912020-06-18T07:05:33Z Achieving efficient realization of Kalman Filter on CGRA through algorithm-architecture co-design Merchant, Farhad Vatwani, Tarun Chattopadhyay, Anupam Raha, Soumyendu Nandy, Soumitra Kumar Narayan, Ranjani School of Computer Science and Engineering 14th International Symposium, Applied Reconfigurable Computing 2018 Hardware and Embedded Systems Lab Engineering::Computer science and engineering Kalman Filter Reconfigurable Architectures Computation In this paper, we present efficient realization of Kalman Filter (KF) that can achieve up to 65% of the theoretical peak performance of underlying architecture platform. KF is realized using Modified Faddeeva Algorithm (MFA) as a basic building block due to its versatility and REDEFINE Coarse Grained Reconfigurable Architecture (CGRA) is used as a platform for experiments since REDEFINE is capable of supporting realization of a set algorithmic compute structures at run-time on a Reconfigurable Data-path (RDP). We perform several hardware and software based optimizations in the realization of KF to achieve 116% improvement in terms of Gflops over the first realization of KF. Overall, with the presented approach for KF, 4-105x performance improvement in terms of Gflops/watt over several academically and commercially available realizations of KF is attained. In REDEFINE, we show that our implementation is scalable and the performance attained is commensurate with the underlying hardware resources. 2020-06-18T07:05:33Z 2020-06-18T07:05:33Z 2018 Conference Paper Merchant, F., Vatwani, T., Chattopadhyay, A., Raha, S., Nandy, S. K., & Narayan, R. (2018). Achieving efficient realization of Kalman Filter on CGRA through algorithm-architecture co-design. Proceedings of the 14th International Symposium, Applied Reconfigurable Computing 2018, 119-131. doi:10.1007/978-3-319-78890-6_10 978-3-319-78889-0 https://hdl.handle.net/10356/142291 10.1007/978-3-319-78890-6_10 2-s2.0-85046284941 119 131 en © 2018 Springer International Publishing AG, part of Springer Nature. All rights reserved.
institution Nanyang Technological University
building NTU Library
country Singapore
collection DR-NTU
language English
topic Engineering::Computer science and engineering
Kalman Filter
Reconfigurable Architectures Computation
spellingShingle Engineering::Computer science and engineering
Kalman Filter
Reconfigurable Architectures Computation
Merchant, Farhad
Vatwani, Tarun
Chattopadhyay, Anupam
Raha, Soumyendu
Nandy, Soumitra Kumar
Narayan, Ranjani
Achieving efficient realization of Kalman Filter on CGRA through algorithm-architecture co-design
description In this paper, we present efficient realization of Kalman Filter (KF) that can achieve up to 65% of the theoretical peak performance of underlying architecture platform. KF is realized using Modified Faddeeva Algorithm (MFA) as a basic building block due to its versatility and REDEFINE Coarse Grained Reconfigurable Architecture (CGRA) is used as a platform for experiments since REDEFINE is capable of supporting realization of a set algorithmic compute structures at run-time on a Reconfigurable Data-path (RDP). We perform several hardware and software based optimizations in the realization of KF to achieve 116% improvement in terms of Gflops over the first realization of KF. Overall, with the presented approach for KF, 4-105x performance improvement in terms of Gflops/watt over several academically and commercially available realizations of KF is attained. In REDEFINE, we show that our implementation is scalable and the performance attained is commensurate with the underlying hardware resources.
author2 School of Computer Science and Engineering
author_facet School of Computer Science and Engineering
Merchant, Farhad
Vatwani, Tarun
Chattopadhyay, Anupam
Raha, Soumyendu
Nandy, Soumitra Kumar
Narayan, Ranjani
format Conference or Workshop Item
author Merchant, Farhad
Vatwani, Tarun
Chattopadhyay, Anupam
Raha, Soumyendu
Nandy, Soumitra Kumar
Narayan, Ranjani
author_sort Merchant, Farhad
title Achieving efficient realization of Kalman Filter on CGRA through algorithm-architecture co-design
title_short Achieving efficient realization of Kalman Filter on CGRA through algorithm-architecture co-design
title_full Achieving efficient realization of Kalman Filter on CGRA through algorithm-architecture co-design
title_fullStr Achieving efficient realization of Kalman Filter on CGRA through algorithm-architecture co-design
title_full_unstemmed Achieving efficient realization of Kalman Filter on CGRA through algorithm-architecture co-design
title_sort achieving efficient realization of kalman filter on cgra through algorithm-architecture co-design
publishDate 2020
url https://hdl.handle.net/10356/142291
_version_ 1681056042694213632