Hi, i failed compile the test case with Chrono
details are shown below
and the system are:
Openmpi mpi-3.1 openmpi-4.0
And also, i want to ask can i run the code by N processes on single node, and each processes control M devices? For example, DGX-2, 12 (maybe, can’t remember the exact number) V100 in a machine, but the code only support 6 GPUs per node, so in this situation, how can i maximum the compute power.