Hi, Huy,
I am not that good at neither;)
I stored the output of the nodes separately. To get output from 3nodes, I generated 3 files. Each file, you need to give it an ID. In the above example case, the file ID is 123, name is TEMP1_89506.data, and at each time step, you want to open it and write the results into it, so we need to open it, and write the results. If you only need 3 point, it is easier to only have 3 files, so you can loop the codes. In this way, you can only ask the precessor who has your point t write out the result, so your rest processors will no generate files.This can save you some space.
Best regards
Yuan