Robust SPMD-like structure

I'm interested in running some subroutine on different sets of data in parallel. Sometimes, a dataset will cause errors in the subroutine, but most of the time it does not. Is there a way to salvage the valid results from all of the datasets without errors? SPMD documentation says that it will stop all workers upon encountering an error in one.

Answer 1

[Walter Roberson]

Put in your own try/catch

You might also want to look at parfeval()