원문정보
초록
영어
The high speed CPU based routers currently in use could not handle the massive data required for real-time multimedia communication. Graphics processing units (GPUs) offer an appreciable alternative due to high computation power which results from their parallel execution units. This paper presents the implementation of the Dijkstra’s link state IP routing algorithm using GPU. Experimental results show that the proposed GPU-based approach outperforms the same sequential CPU-based implementation in terms of execution time for the same dense graph. In addition, the proposed GPU-based approach reduces about 99% energy consumption over the CPU-based implementation.
목차
1. Introduction
2. Background Information
2.1. GPU and CUDA Interaction
2.2. Graph Mapping on CUDA
2.3. The Dijkstra’s Algorithm
3. Implementation of Parallel Dijkstra’s Algorithm Using CUDA
4. Performance Metrics
4.1. Throughput
4.2. Speedup
4.3. Power and Energy Consumption
5. Experimental Results and Analysis
5.1. Performance of the GPU-based Approach
5.2. Average Power and Energy Consumption of GPU over CPU
6. Conclusions
Acknowledgements
References