Hi,
I have a number of h5part (hdf particle) files with particle data at a large number of timesteps. If I attempt to visulaize these files without an mpi server everything works as expected: the particles move according to their time time evolution. If I connect with a parallel server whether with auto mpi or by explicitly connecting, the particles disappear after the first time step, although it appears that a single particle continues to be tracked. This behavior is the same on my linux box where I have compiled 5.8.1 and the most recent release with both gnu and intel compilers and mpi. I have all the mpi falgs enabled. in the compilation. The same behavior also occurs on the sytem version (5.8.1) on the NCAR cheyenne casper visualization system.
Any ideas about what’s going on?
Thanks, John Lyon