The problem I have is that some of the parallel channels in my application are not doing any work and some are overloaded. I know that the data is skewed towards some values of the parallel attribute and I’m trying to find a solution to make the parallel split more even, for example convert the string to an integer (?). The string values are actually IPV4 addresses so converting to integer would be easy.
Could someone clarify how the values of an rstring attribute are split when specified as the only column in an @parallel annotation? I know from other posts that there’s a hash+mod algorithm, for an rstring does this use all of the value or just a subset?