Please use this identifier to cite or link to this item: https://doi.org/10.1109/RSP.2010.5656347
Title: Performance-cost analyses software for H.264 forward/inverse integer transform
Authors: Do, T.T.T. 
Le, T.M. 
Nguyen, B.P. 
Ha, Y. 
Issue Date: 2010
Source: Do, T.T.T.,Le, T.M.,Nguyen, B.P.,Ha, Y. (2010). Performance-cost analyses software for H.264 forward/inverse integer transform. Proceedings of the International Workshop on Rapid System Prototyping : -. ScholarBank@NUS Repository. https://doi.org/10.1109/RSP.2010.5656347
Abstract: In the literature, Data Throughput rate per Unit Area (DTUA) has been used as the sole metric to evaluate the effectiveness of H.264 Forward/Inverse Integer Transform (FIT/IIT) designs. However, other than throughput and circuit area involved in DTUA, interconnection, power and delay are not considered. In this paper, we first summarize the Performance-Cost Metric (PCM) technique, where PCM is defined as the ratio of data throughput over the design cost including power, area, delay, and issues associated with interconnections in sub-micron designs. Compared to DTUA, PCM facilitates more comprehensive comparisons for VLSI designs, including FIT/IIT. The contribution of this paper is the software that helps facilitate the use of the PCM technique. When using this software, users are asked to enter some preliminary parameters of their design. Based on the given parameters and the reference designs managed using the software, it then analyzes and provides the possible boundaries of the users' design in order to have better PCMs compared to the reference designs. In addition, it can also export comparison results among different designs. The software is flexibly designed in order to facilitate the use of not only our PCM technique in FIT/IIT designs, but also different metrics in other architectures. © 2010 IEEE.
Source Title: Proceedings of the International Workshop on Rapid System Prototyping
URI: http://scholarbank.nus.edu.sg/handle/10635/71424
ISBN: 9781424470747
ISSN: 10746005
DOI: 10.1109/RSP.2010.5656347
Appears in Collections:Staff Publications

Show full item record
Files in This Item:
There are no files associated with this item.

Page view(s)

28
checked on Dec 16, 2017

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.