This is the authors' abstract. We don't add key points for this paper.
This paper presents the system architecture, modeling, and design constraints of a wireless chip-to-chip-communication transceiver as a low-power alternative to wireline links, such as PCI-Express. On top of the potential power savings, the wireless link provides lower latency times, better flexibility, lower complexity, and easier heat diffusion. The proposed transceiver uses impulse-radio ultra-wideband communication at 10 GHz. It supports a 2.5-Gb/s data rate with pulse position modulation over short distances (~10 cm). The hardware complexity is reduced by using a modified non-coherent receiver with a rectifying RF front-end and a relative-compare analog baseband. The system performance is quantified and tradeoffs are explored with the main motivation of reducing power. A discussion of the modulation choice, the performance specifications of the receiver blocks and the clock generation principle is presented. We show that energy efficiency better than 6 pJ/bit could be reached with the proposed architecture.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Gimeno et al. (2017) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: