MichaelPapka

1.0E55

I only cried for the first week. :)

Answer to Q3: We use a bunch of tools, mostly developed by ALCF, to monitor the state of the machines. The developers themselves use both homegrown and commercial tools to develop their codes to optimize the performance of their application.

Answer to Q1: We don't really use big data to perform research. Rather, we see resources like Mira as instruments combined with scientific codes (programs) as being the generators of big data. The big data is the output of the simulations that is mined by the scientists for insight into the scientific questions they are trying to answer.

Answer to Q2: As I oversee the Leadership Computing Facility here, we want to squeeze as much performance out of our system as possible to enable our scientific user community to make breakthroughs. What we are seeing in today's large systems is a large number of computer cores to get computational power needed to solve the most pressing problems. The challenge is for the scientific codes to be able to use all of the cores efficiently.