Table of Contents |
Down-conversion: Code differences | Results: Up-conversion
The results described here are mainly concerned with the amount of time taken to convert the image. The timing results were generated by running the conversion loop (not the file IO) 1000 times and dividing the time taken to give an average execution time. The code was compiled using the GCC compiler with full optimisation settings. Code ran on a dual Athlon MP 2000+ processor machine with 1 GB RAM running SuSE Linux version 8.2. The code was not multi-threaded to take advantage of the dual processors, as this would be too platform specific.
In this section: