Parallelism and Scalability in an Image Processing Application

Morten S. Rasmussen, Matthias B. Stuart and Sven Karlsson

International Workshop on OpenMP (IWOMP 2008)
Purdue University, West Lafayette, IN, USA, May 12th - May 14th 2008

---

Summary

The recent trends in processor architecture show that parallel processing is moving into new areas of computing in the form of many- core desktop processors and multi-processor system-on-chip. This means that parallel processing is required in application areas that traditionally have not used parallel programs. This paper investigates parallelism and scalability of an embedded image processing application. The major chal- lenges faced when parallelizing the application were to extract enough parallelism from the application and to reduce load imbalance. Experi- mental results show that the immediately available parallelism is limited and small data sets cause a high parallelization overhead. Load imbalance is also an issue. The non-uniform memory latency of the test platform further increases load imbalance. Even so, we show that with some tuning relative speed ups in excess of 9 on a 16 CPU system can be reached.

---

  

START Conference Manager (V2.54.5)