MRMCD 2018 Vortragende: Michael F. Herbst
In my professional life I am a researcher in theoretical chemistry, a branch of chemistry, which tries to describe the chemical and physical properties of matter from simulations rather than experiments. More specifically my work deals with developing new algorithms and codes for such simulations.
More details on my research can be found on michael-herbst.com.
Pitfalls for performance: Typical latencies to keep in mind
"Less computation implies faster code": Sounds reasonable at a first glance, but a closer inspection considering memory or disk latency numbers of modern computer architectures quickly reveals this to be a fallacy. In this talk I will discuss some ballpark latency numbers every programmer should know and hint at implications for designing algorithms in the context of high-performance computing as well as scientific simulations.