The competition includes two parts:
Part 1 – Artificial Intelligence – DLRM (Deep Learning Recommendation Model)
DLRM (Deep Learning Recommendation Model) is a neural network-based recommendation tool. It is a state-of-the-art deep learning recommendation model which Facebook developed in both PyTorch and Caffe2 frameworks. Competition teams should utilize the model parallelism while exploiting data parallelism to scale-out compute from the fully-connected layers
Benchmark download: https://github.com/facebookresearch/dlrmBenchmark Guidelines
Part 2 – High-Performance Computing – GROMACS (GROningen MAchine for Chemical Simulations)
GROMACS (GROningen MAchine for Chemical Simulations), a popular molecular dynamics simulation tool designed for biochemical molecules like proteins and nucleic acids, is chosen to continue the global fight against COVID-19. Competition teams should optimize the GROMACS task on current baseline.
Benchmark download: http://www.gromacs.org/Benchmark Guidelines