The K80 CUDA node on Rostam was recently re-added to the cluster. Since our tests will run on all free CUDA nodes since #395 (and we don't support the K80 for the Kokkos kernels) this caused failing Jenkins tests in case this node was picked. Hence, we will now only build on the 4 CUDA machines on the V100 and A100 partitions.
The K80 CUDA node on Rostam was recently re-added to the cluster. Since our tests will run on all free CUDA nodes since #395 (and we don't support the K80 for the Kokkos kernels) this caused failing Jenkins tests in case this node was picked. Hence, we will now only build on the 4 CUDA machines on the V100 and A100 partitions.