PROJECT TITLE :

Using Transmission Lines for Global On-Chip Communication

ABSTRACT:

The growing range of cores in chip multiprocessors increases the importance of interconnection for overall system performance and energy potency. Compared to ancient distributed shared-memory architectures, chip-multiprocessors (CMPs) offer a totally different set of style constraints and opportunities. As a result, a conventional packet-relay multiprocessor interconnect architecture is a valid, however not essentially optimal, design point. Worsening wire delays, energy-inefficient routers, and therefore the decreased importance of in-field scalability, build the standard packet-switched network-on-chip a less attractive choice. Another solution uses well-built transmission lines as communication links. These transmission lines, along with straightforward, practical circuits using trendy complementary metal–oxide–semiconductor technology, will give low latency, low energy, high throughput channels that will be used as a shared-medium point-to-purpose link. The look of the transmission lines and transceiver circuits has vital architectural impact. This paper includes a initial-step style effort for these elements, significantly when used for a globally shared-medium bus. For medium-scale CMPs, this interconnect backbone will eliminate the need for packet switching and provide energy, along with performance advantages in comparison to a conventional mesh interconnect. We will provide a style of such a system from the bottom up, together with design of the transmission lines, transceiver circuits, and a simple, nonetheless effective, architectural design for a shared-medium interconnect, and show that such a style can be a compelling different to packet-switched networks for CMPs.