PROJECT TITLE :

Swizzle-Switch Networks for Many-Core Systems

ABSTRACT:

This work revisits the look of crossbar and high-radix interconnects in lightweight of advances in circuit and layout techniques that improve crossbar scalability, obviating the necessity for deep multi-stage networks. We tend to use a replacement building block, the Swizzle-Switch—an energy- and area-economical switching part that may readily scale to radix 64—that has recently been validated via silicon check chips in forty five nm technology. We evaluate the Swizzle-Switch as both the high-radix building block of a Flattened Butterfly and as a single-stage interconnect, the Swizzle-Switch Network. In the process we address the architectural and layout challenges associated with centralized crossbar systems. Compared to a typical Mesh, the Flattened Butterfly provides a 15percent performance improvement with a 2.five$times$ reduction in the standard deviation of on-chip access times. The Swizzle-Switch Network achieves further gains, providing a 21percent improvement in performance, a three$times$ reduction in on-chip access variability, a 33% reduction in interconnect power, and a 25p.c reduction in total system energy whereas solely increasing chip space by 7percent. Finally, this paper details a three-D integrated version of the Swizzle-Switch Network, showing up to a thirtyp.c gain in performance over the 2-D Swizzle-Switch Network for benchmarks sensitive to interconnect latency. One major concern with 3-D styles is thermal dissipation. We show through detailed thermal analysis that with the highly energy-economical Swizzle-Switch Network design that the thermal budget is well inside that of passive cooling solutions.