The following git diff shows that just a few lines of code suffice to add useful progress reporting to mclapply: https://github.com/dcnorris/precautionary/commit/3e561cc8bea7305a03faf000cd21da19d285842f#diff-328c889e85e15d64ac52741999d1afe338d9f299076fe4a7448625301317d887
I've done this in package https://dcnorris.github.io/precautionary/ to support the 'odometer' that you can appreciate by loading the following app: https://precision-methodologies.shinyapps.io/EscRisk/ ... and clicking the "CRM" radio button about halfway down on the left. 1. In this application, I report *absolute* progress because it's not knowable in advance what the final total will be. But the more customary type of "% complete" reporting would easily be managed by a 'progreport' function that knows the denominator in advance. 2. The 'proginit' argument is needed for my application because the initial single-threaded unrolling (to generate partial computations which the parallel workers finish) may yield some completed computations as well. Would there be interest in discussing modifications to parallel::mclapply along these lines? Kind regards, David Norris ______________________________________________ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
