The Bound-T WCET tool from Tidorum Ltd was entered in the WCET Challenge 2006 for the following target processors:
To introduce the report on this work, we briefly describe Bound-T and these target processors.
The WCET Challenge benchmarks were analysed separately and more or less independently by Tidorum itself and by Ms. Lili Tan from the University of Duisburg-Essen. Ms. Tan also analysed the benchmarks with the other tools entered in the Challenge. This report describes Tidorum's analysis of the benchmarks.
This report is structured as follows:
The table in section 4 reports the resulting WCET bounds for the benchmark programs for the chosen target processors and cross-compilers. Please note very carefully that readers should not, from these WCET bounds, draw any conclusions regarding the code-generation performance of the compilers, firstly because the numbers are only upper bounds and may be more precise (less overestimated) for one compiler than for another, and secondly because we used the same source-code for all compilers, while embedded software developers usually adapt the source-code to use the good features of their chosen compiler.