Abstract | ||
---|---|---|
The communication efficiency plays a crucial role in achieving high system performance in many multimedia SoCs (System-On-Chips). Increased requirements on high bandwidth on-chip communication have led to the emergence of complicated communication architectures and algorithms. However, without thorough analysis of bandwidth requirements of applications, we tend to over-provision resources to avoid potential performance degradation. In this paper, we present a CBR-based traffic modeling technique along with a performance evaluation framework. By using this evaluation technique, we can estimate the sustained performance of on-chip communication infrastructure used in multimedia SoCs. The traffic behaviors of a multimedia application have been captured and analyzed to characterize individual operations which are modeled using traffic generators to replace hardware IPs in evaluation process. The simulation results show that this approach is also effective in discovering the peak performance of on-chip network and bandwidth allocation to IPs. |
Year | DOI | Venue |
---|---|---|
2009 | 10.1145/1644993.1645111 | ICHIT |
Keywords | Field | DocType |
complicated communication architecture,performance evaluation framework,communication efficiency,multimedia socs,high system performance,cbr traffic generator model,high bandwidth on-chip communication,potential performance degradation,sustained performance,on-chip communication infrastructure,peak performance,chip,system performance,system on chip,traffic generator,bandwidth allocation,simulation | Traffic generation model,Bandwidth allocation,Computer science,Computer network,Bandwidth (signal processing),Interconnection,Embedded system,High bandwidth | Conference |
Citations | PageRank | References |
0 | 0.34 | 1 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Kee Beom Kim | 1 | 0 | 0.68 |
Seong Min Jo | 2 | 0 | 0.68 |
Jin Woo Song | 3 | 5 | 1.89 |
Ki-seok Chung | 4 | 189 | 18.76 |
Yong Ho Song | 5 | 199 | 23.33 |