Abstract:The 2D mesh topology-based Network-on-Chip (NoC) is a prevalent structure in System-on-Chip (SoC) designs, offering implementation and fabrication benefits. However, increased NoC scale leads to longer communication paths, more hops, and higher end-to-end latency and energy consumption. To mitigate these issues, Wireless NoC (WiNoC) integrates wireless communication, enhancing data rates, energy efficiency, and routing flexibility. Despite several mapping algorithms for NoCs, optimal techniques for hybrid WiNoCs are underexplored. This study proposes two novel application mapping methods for 2D mesh topology-based hybrid WiNoCs, using quadratic programming (QP) and simulated annealing (SA). Our goal is to minimize communication-related energy consumption. We evaluated these methods across various wireless router configurations, benchmarks, and custom application graphs. The QP-based method excels in smaller problems, while the SA-based approach yields optimal or near-optimal results for larger sizes within practical runtimes.
computer science, theory & methods,engineering, electrical & electronic, hardware & architecture
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to optimize the application mapping method in the wireless network - on - chip (WiNoC) based on 2D grid topology in the hybrid wireless network - on - chip (WiNoC) to reduce communication - related energy consumption. Specifically, as the scale of NoC increases, it leads to longer communication paths, more hops, and higher end - to - end latency and energy consumption. To alleviate these problems, the research proposes two new application mapping methods, based on quadratic programming (QP) and simulated annealing (SA) respectively, aiming to minimize the communication energy consumption of applications on the hybrid WiNoC.
### Main Contributions
1. **Model Based on Quadratic Programming (QP)**: A model for application mapping in 2D - grid hybrid WiNoC is proposed with the goal of minimizing energy consumption. Although this model performs optimally when dealing with smaller problems, it is not suitable for problems with a large number of nodes due to its long execution time. The results of this model serve as a benchmark for evaluating meta - heuristic methods.
2. **Method Based on Simulated Annealing (SA)**: A meta - heuristic method is proposed, which can efficiently solve problems of various scales and provide optimal or near - optimal solutions. This method shows significant improvement in terms of running - time performance.
3. **Comparison under Different Wireless Router Configurations**: The QP and SA methods are compared under different wireless router configurations, and the influence of the number of routers and the number of tasks on communication cost and system design is analyzed. The results show that the SA method is superior to the slower QP model in terms of running time and effectiveness.
4. **Comparison with Existing Heuristic Methods**: The SA method is compared with the existing heuristic mapping methods in the literature, showing that it can achieve better performance in all benchmark tests.
### Method Overview
- **Pre - processing Algorithm**:
- **Weight Calculation**: Calculate the communication cost weights between each connected node, including the costs of wired and wireless connections.
- **Minimum Communication Cost Calculation**: Use the Floyd - Warshall algorithm to determine the shortest path between all pairs of nodes in a given topological graph.
- **QP - based Method**:
- Define binary variables \(\mu_{i,j}\), which indicate whether task \(i\) is mapped to node \(j\).
- Each task must be mapped to one and only one node (formula (5)).
- Each node can be assigned at most one task (formula (6)).
- Define binary variables \(\delta_{i,j,r,s}\), which indicate that task \(i\) is mapped to node \(r\) and task \(j\) is mapped to node \(s\) (formula (7) and (8)).
- Calculate \(\delta_{i,j,r,s}\) through the logical AND operation \(\mu_{i,r} \times \mu_{j,s}\) (formula (9)).
- The total communication energy consumption \(E_{\text{Total}}^{r,s}\) is calculated by formula (11).
- The total system communication energy consumption \(E_{\text{WiNoC}}\) is calculated by formula (12).
- The objective function is to minimize the total system communication energy consumption.
- **SA - based Method**:
- Determination of Initial State: Use a heuristic method to generate an initial mapping solution, avoiding randomly generated initial solutions.
- Initial Configuration and Cooling Plan: Automatically determine the initial temperature and the final temperature to ensure an initial acceptance probability of 98% and a final improvement rate of 0%. In the experiment, the maximum number of attempts per trial is set to 12.
### Conclusion
This research effectively solves the application mapping problem in the hybrid WiNoC by proposing two new methods based on quadratic programming and simulated annealing, significantly reduces communication energy consumption, and performs well in problems of different scales.