Abstract:Network on Chip (NoC) has become a promising solution for the communication paradigm of the next-generation multiprocessor system-on-chip (MPSoC). As communication has become an integral part of on-chip computing, researchers are paying more attention to its implementation and optimization. Traditional techniques that model inter-processor communication inaccurately will lead to unexpected runtime performance, which is on average 90.8% worse than the predicted results based on an observation. In this paper, we present an application mapping and scheduling technique for NoC-based MPSoCs that integrates fine-grain optimization on inter-processor communications with the objective of minimizing the schedule length. A communication model is proposed to address properly the latency of inter-processor communication with network contention. Performance evaluation results show that solutions obtained by the proposed technique can generate realistic performance that is on average 34.7% higher than traditional techniques, and the Integer-Linear Programming (ILP) based approach can outperform the state-of-the-art heuristic algorithms by 31.1%. A case study on H.264 HDTV decoder shows that our approach achieves 22.8% improvement in prediction accuracy, 20.9% improvement in performance and 40% reduction in the number of network contentions.

Noc Self-Similar Traffic Modelling Based on Systolic Architecture Unit

SoC performance modeling methodology and implementation basedon transaction dataflow

Systolic Traffic Modelling in Network on Chip

Modeling and Simulation of Self-Similar Traffic Based on FBM Model

A Systematic Network-On-Chip Traffic Modeling and Generation Methodology

Affect Analysis of Fft Algorithm Length on Traffic Self-Similarity in Noc

Design and Implementation of Self-similar Network Traffic Generator

Self-similar traffic study of on-chip interconnection networks

Modeling Self-Similar Traffic for Network Simulation

Analysis of Network Performance for a New Self-similar Traffic Model

On Study and Implementation of Self-Similar Network Traffic

A NoC Traffic Suite Based on Real Applications

Research on Network on Chip Performance Evaluation Technology Supporting Various Traffic Distribution

Performance Evaluation of Noc-Based Multicore Systems: from Traffic Analysis to Noc Latency Modeling

The Implementation of a Distributed System Based on a Parallel Algorithm for Self-Similar Network Traffic Simulation

Self-similar Traffic: a New Model for Next Generation ATM Switch

A Novel Synthetic Traffic Pattern for Power/Performance Analysis of Network-on-Chips Using Negative Exponential Distribution

Traffic behavior analysis with Poisson sampling on high-speed network

The Study of Algorithms Generating Self-similar Traffic

Traffic-Aware Application Mapping for Network-on-Chip Based Multiprocessor System-on-Chip

A New Self-similar Teletraffic Generation Arithmetic and Simulink