dAElite: A TDM NoC Supporting QoS, Multicast, and Fast Connection Set-Up

Abstract

Networks-on-Chip are seen as promising interconnect solutions, offering the advantages of scalability and high frequency operation which the traditional bus interconnects lack. Several NoC implementations have been presented in the literature, some of them having mature tool-flows. The main differentiating factor between the various implementations are the services and communication patters they offer to the end-user. In this paper we present dAElite, a TDM Network-on-Chip that offers a unique combinations of features, namely guaranteed bandwidth and latency per connection, built-in support for multicast, and a short connection setup time. While our NoC was designed from the ground up, we leverage on existing tools for network dimensioning, analysis and instantiation. We have implemented and tested our proposal in hardware and we compared it to AEthereal, a state-of-theart NoC with similar features, but no multicast. We find that the connection setup time is reduced by a factor of 10 and the network traversal latency is decreased by 33%. Moreover, considering realistic values of the network parameters dAElite has a lower hardware area when synthesized in 65 nm technology.

A Minimal Network Interface for a Simple Network-on-Chip

“…For a 2x2 NoC, the S4NOC uses 1183 4-input LUTs and 1110 flip-flops. In contrast, aelite uses 7665 6-input LUTs and 15444 flip-flops [16].…”

A real-time on-chip network architecture for mixed criticality aerospace systems

“…The concept of TDM is used in Refs. (21) and (22)(23)(24), where the resource is allocated to channels based on time slots as an alternative to circuit switching. In Ref.…”

Conflict-Free Networks on Chip for Real Time Systems

“…149 A. 14 Gradually moving the added delays by adding two pipeline stages (a). By adding six pipelines stages every hop have the same length (b).…”

“…However, TDM NoCs have difficulties in finding the optimal schedules. TDM schedules can be statically [12], [13], [14], [52], [1], [53], [47], [54] or dynamically computed [55] and may be placed locally at each switch [55] for distributed routing or globally in the network interfaces (NIs) for source routing [12], [14].…”

“…The NoC implemented in real-time MP-SoCs is, therefore, impacted by this requirement and needs to be designed accordingly. To address the communication contention problem, Time Division Multiplexing (TDM) [2,[12][13][14][15] is the most popular way to ensure time predictability in NoCs. Relying on TDM is one of the classical methods of providing configurable bandwidth reservation, guaranteed latency and throughput.…”