Webinar: Fuzzball (HPC 2.0): Workflows & Solutions for HPC Use-Cases Linux is Now Fully Supported on GCP
In this webinar, we will discuss workflows and solutions to complex HPC use cases:
LAMMPS and GROMACS:
LAMMPS and GROMACS are two molecular dynamics (MD) suites commonly used in HPC that support research across a variety of domains, including biology, pharmaceutical discovery, materials science, and engineering. Both can utilize GPUs, if available, to accelerate computation, and can also take advantage of MPI software stacks for parallelization across multiple discrete compute nodes at once. In this demo, we will watch LAMMPS and GROMACS each run an application performance benchmarking test across multiple GPU-enabled compute nodes.
TensorFlow workflow lifecycle:
TensorFlow is one of the most popular frameworks for machine learning, offering a number of advanced capabilities for AI training and inference – how do we set up a Fuzzball workflow that can utilize it? In this demo, we will look at a simple Fuzzball workflow lifecycle process: defining what our workflow should do, developing or discovering a container with the necessary applications, creating the workflow definition (with a sneak peek at the GUI-based Fuzzball workflow editor!), running this workflow on public cloud resources orchestrated with Fuzzball, and then connecting to the Jupyter Notebook interface the workflow will be built to provide in order to run some simple TensorFlow code with a GPU.