More then one attempt.The number of attempts can strongly vary depending on many factors. Anyway, computational materials science could allow to do speed up the search and screening of the desired materials reducing the waste of resources. In a way, there’s still a usage and waste of resources, time and electricity, which are needed to run the calculation but they’re of a lesser extent then research done by “trial and error”. I say that computational physics still waste a lot of resources because algorithms might not be yet that efficient and still a lot of power and electricity and time are required to run the calculations. Nowadays, the progress made in the field of machine learning are being implemented in the field of materials science to allow for more efficient and faster research.