An Improved Connection Method for Multi-Core SoC

An Improved Connection Method for Multi-Core SoC

Chunping Zhang
DOI: 10.4018/japuc.2012010105
OnDemand:
(Individual Articles)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

With rapid development of multi-core processor, the communication becomes a bottleneck increasingly. Traditional electronic connections face a serious complexity restriction. The network on chip has been a prevalent solution in the present. However using the recent technology (ultra-wide band interconnection) technology, the over-all performance including delay and throughput can achieve a new level in multi-core SOC. This paper will introduce a new solution designed for UWB-I based SOC. This solution includes multi-channeling, topology design and routing etc.
Article Preview
Top

1. Introduction

Multi-processor chips (CMP) became a prevalent solution to stride the huge gap between single processor throughput and overall performance requirement in recent years. However, the performance of shared bus connection including throughput, delay and energy consumption restricts the overall performance. The shared bus architecture use the connection resource in a preemption way, that means communication of a lower priority component is frequently interrupted by a higher priority component. In all, this is not an efficient way.

Network-On-Chip (NOC) is a solution to replace the conventional shared-bus architecture to provide scalable communication for CMP. In NOC, each processor is directly connected with a router; routers in the plane are connected together. The communication could be flexible. Packets in different router pairs are not interference with each other. This architecture gives a scalable, parallel structure to CMP. Therefore, the performance is improved significantly.

The typical NOC connection structure is based on a hard-wire based mesh, which is easy for routing packets but not efficient for hardware cost. Radio Frequency/Wireless interconnect technology has emerged recently to address future communication needs and surpass the fundamental limitation of hard-wired electronic interconnects.

In these wireless solutions, the Ultra-Wide-Band interconnection (UWB-I) is the most promising to provide a high-bandwidth, low latency communication. Moreover, the high flexibility and free-of-wiring make UWB-I an attractive solution for the on-chip inter-core communication.

Ultra-wideband is a radio technology which may be used at a very low energy level for short-range, high-bandwidth communications using a large portion of the radio spectrum.

Similar to spread spectrum, UWB communications transmit in a way which does not interfere with conventional narrowband and carrier wave uses in the same frequency band. Unlike spread spectrum, however, ultra-wideband does not employ frequency-hopping (FHSS).

Ultra-wideband is a technology for transmitting information spread over a large bandwidth (>500 MHz); this should, in theory and under the right circumstances, be able to share spectrum with other users. Regulatory settings by the Federal Communications Commission (FCC) in the United States intend to provide an efficient use of radio bandwidth while enabling high-data-rate personal area network (PAN) wireless connectivity; longer-range, low-data-rate applications; and radar and imaging systems.

Ultra wideband was formerly known as “pulse radio”, but the FCC and the International Telecommunication Union Radiocommunication Sector (ITU-R) currently define UWB in terms of a transmission from an antenna for which the emitted signal bandwidth exceeds the lesser of 500 MHz or 20% of the center frequency. Thus, pulse-based systems—where each transmitted pulse occupies the UWB bandwidth (or an aggregate of at least 500 MHz of narrow-band carrier; for example, orthogonal frequency-division multiplexing (OFDM)—can gain access to the UWB spectrum under the rules. Pulse repetition rates may be either low or very high. Pulse-based UWB radars and imaging systems tend to use low repetition rates (typically in the range of 1 to 100 megapulses per second). On the other hand, communications systems favor high repetition rates (typically in the range of one to two gigapulses per second), thus enabling short-range gigabit-per-second communications systems. Each pulse in a pulse-based UWB system occupies the entire UWB bandwidth (thus reaping the benefits of relative immunity to multipath fading, but not intersymbol interference), unlike carrier-based systems which are subject to deep fading and intersymbol interference.

A valuable aspect of UWB technology is the ability for a UWB radio system to determine the “time of flight” of the transmission at various frequencies. This helps overcome multipath propagation, as at least some of the frequencies have a line-of-sight trajectory. With a cooperative symmetric two-way metering technique, distances can be measured to high resolution and accuracy by compensating for local clock drift and stochastic inaccuracy.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 11: 4 Issues (2019)
Volume 10: 4 Issues (2018)
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing