Low cost loosely coupled parallel processing system using standard processor
The experiment is to use low cost processors and integrate them to work in parallel. Along the way to achieve the objective, many fields of knowledge need to be studied. As in this project, knowledge of matrix operation and how it can be broken down for the purpose of parallel processing are studied...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2009
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/18395/1/ZahariAwangAhmadMFKE2009.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The experiment is to use low cost processors and integrate them to work in parallel. Along the way to achieve the objective, many fields of knowledge need to be studied. As in this project, knowledge of matrix operation and how it can be broken down for the purpose of parallel processing are studied. Special task is identified to focus in this project that is matrix operation for image processing. Eventually, the ultimate objective is to see how much speed of processing that can be improved when parallel processing is implemented. The speedup is the keywords. There are three main parts in this development. Firstly, the parallel processing system requires a hardware integration whereby two ARM processors are integrated. Secondly, the system requires firmware to run on both processors. The firmware is to develop based on C ARM compiler. This firmware will handle the image processing, data transfer between processors and data transfers to Laptop PC. Thirdly, the whole system requires a program on the laptop PC to display data on LCD monitor and capture the processing time. Upon all the system in operation, processing time of single processor and dual parallel processor can be compared and analyzed. A significant improvement in term of speedup is observed and there seems to be many ways to further develop the system to benefit wide range of applications. |
|---|