store a single HDF5 file on different machines

Hello,

We are working on a project that uses a Linux cluster to visualize large
scale data and use a display wall to show the rendered image. The size of
data generated by the nodes is very large, so it is not appropriate to
transfer the distributed data to the master node to assemble a single HDF
file. I know HDF5 has MPI support , but I wonder can a single HDF5 file be
stored on different nodes of the cluster.

Ning