As part of the GAMESS ECP project, we are working on porting parts of our CUDA code to OpenMP offload and DPC++. To do this, we would like access to V100s so that we can 1) get baselines of the CUDA code and 2) try DPC++ and OpenMP on the V100s to see how “performance portable” the choice of programming model is across Intel and Nvidia hardware. Being able to explore and compare CUDA to OpenMP to DPC++ will help the project make a choice about a programming model to use for the future.