I created my own reader plugins which can read raw data from files and convert data into vtk data successfully.
However, in my situation, there will be a large number of files, so i want to use MPI to read data, then all the data can be convert on each separate cpu on server side, finally show them together on the client side screen.
My general ideas is:
- change reader to receive a directory
- find all the files under this directory
- distribute files according to its processID
My questions are: what’s the best practice for my situation?
Thanks and Best Regards,