> #PortegiesZwart thinks that instead of teaching #physicists more computing, perhaps physics research institutes should employ more computer experts. “We are great at physics, but a computer scientist spent all the time we learned about physics learning how to communicate with a computer,” he says. “There is no doubt that [they] will be better at #programming.”
@julm That’s a misunderstanding. Computer scientists learn how to create maintainable programs. That’s not really what high performance physics computing needs. With mpi, openmp, and high performance math libraries, scientific computing runs circles around code from most computer experts. But it has horrible abstractions (if at all) and is terrible to adapt to new situations.
@ArneBab @julm I don't completely agree. Most people working in what we call the "computer science" field had learned how to create maintainable programs. Some of us, computer scientists, had learned how to write efficient (in speed and energy) programs (or even not programs). Actually, where do you think MPI, OpenMP, maths libraries and stuff come from ? 😉
Review of #GreenAlgorithms' paper on the @BE4FOSS 's mailing-list:
Where I see CS but *not* physicists is stuff like https://www.pypy.org
¹: I prefer the "and", but there’s also the inefficient-maximum-parallelism and the serial-but-energy-saving subtype.
Le réseau social de l'avenir : pas de publicité, pas de surveillance institutionnelle, conception éthique et décentralisation ! Gardez le contrôle de vos données avec Mastodon !