Automated regression testing and code coverage analysis of the CP 2 K application Testing the big super - computing

ou have probably heard the story of St. George and the Dragon. According to legend, the Dragon caused many problems for the ancient city of Lasia, somewhere in the Middle East. Dragons are mighty, firebreathing creatures, but even in those auld times, there were brave men like St. George who slayed the Dragon, saved the city, and became a legend. In modern days we do not believe in those mythical creatures, but rather use them as a metaphor of something really huge and powerful. Nowadays, supercomputers are somehow like dragons – they are mighty and powerful, consuming vast amounts of energy and producing heat, but still vulnerable due to the errors in software. Like the dragon Smaug from J. R. R. Tolkien's ‘The Hobbit’ who had a weak spot, an ‘Achilles' heel’ in his armour that eventually led to his defeat. Supercomputers will not disappear, that is certain, but rather continue to evolve, leading to diversity in High Performance Computing ecosystem, and thus posing problems to application developers. Instead of dealing with one sort of “supercomputing beast”, they have to deal with many in order to run their application on different platforms. Figuratively speaking, it makes more sense to think of a parallel, large scale application as the dragon with more than one head, or, if you like, as more powerful mythical creature – the Lernaean Hydra. As you remember, Hydra had many heads – just like the supercomputers developers use to build, test and execute their parallel applications. It is one of the reasons why testing of large scale applications is becoming increasingly important.