Running MD applications efficiently in HPC

The aim of the course is to provide MD users with a set of best practices to improve the performance of their simulations. This will be accomplished by tuning the appropriate SLURM switches and flags proper to the MD packages.

This course will be two half-days. It will have lectures interspersed with some hands-on sessions where you get to try out what you have just learned.

NOTE: The course will be online!

Topics to be covered are:

• Performance analysis of different versions of the packages (AMBER, NAMD, GROMACS, LAMMPS)
• Performance analysis on different architectures (CPU, GPU, KNL CPUs)
• General recommendations when using plugins.
• Setting up classical MD and QM/MM simulations with NAMD+VMD and GAUSSIAN

Instructor:

Pedro Ojeda

Helper:

Birgitte Brydsö