Encoding serial data for energy-delay-product and energy minimization

Serial or parallel buses are widely used to communicate information in most electronic devices. The energy consumed by bus interconnects may comprise a significant portion of the overall energy consumption of the device. Hence, techniques to reduce the energy consumption of bus interconnects have bec...

Full description

Bibliographic Details
Main Author: Ekambavanan, Sasidharan
Other Authors: Khatri, Sunil P.
Format: Others
Language:en_US
Published: 2010
Subjects:
EDP
LCD
Online Access:http://hdl.handle.net/1969.1/ETD-TAMU-1490
http://hdl.handle.net/1969.1/ETD-TAMU-1490
Description
Summary:Serial or parallel buses are widely used to communicate information in most electronic devices. The energy consumed by bus interconnects may comprise a significant portion of the overall energy consumption of the device. Hence, techniques to reduce the energy consumption of bus interconnects have become an important area of research. One common method adopted to reduce energy is Bus Encoding, where redundant bits are added to the original data stream either in time or space to reduce the energy consumption. In this thesis, a novel bus encoding technique, called the Multiple Codebook Approach (MCA), is presented to reduce the Energy-Delay-Product (EDP) and Energy for data transmission over serial buses. For any symbol that is to be transmitted on the bus, the best code is selected (from an EDP or energy minimization standpoint) from among a set of codebooks. In particular, the implementation utilizes 3 codebooks. To minimize EDP, the codeword for each symbol is selected based on the product of the number of transitions resulting from its transmission in a serial manner, and the codeword length. To minimize energy, the codeword for each symbol is selected based on the number of transitions alone. The MCA is compared with other reported techniques in the literature, and the results are quite promising. The MCA achieves a 11% improvement in EDP and 3% improvement iv in Energy over the best approach known for serial data transmission, which was reported by Macii and others.