Henrik Bengtsson

The `future_lapply()` polls each future on a regular basis to check whether it is resolved or not - that's the `resolved()` call - and this will keep repeating until the...

@mllg, not sure if you suggested that in your comment, and not sure if it makes a difference for all schedulers, but do you think the overall load on the...

Forgot to say, including what the working directory is also helpful in cases where it is set to, say, a local temp folder that is not available on the compute...

Actually, even more informative would be: ```r error("Working directory (%s) for the batchtools registry does not exist on host %s", sQuote(jc$work.dir), sQuote(Sys.info()[["nodename"]])) ``` Annotating other errors that occur when trying...

The future.batchtools package exports [each global object using `batchtools::batchExport()`](https://github.com/HenrikBengtsson/future.batchtools/blob/1895d859e666c04ce18df6038cda418e768a11fd/R/BatchtoolsFuture-class.R#L372-L375): ```r ## (iii) Export globals? if (length(future$globals) > 0) { batchExport(export = future$globals, reg = reg) } ``` So, if there...

Agree with @mllg. FWIW, using `saveRDS(..., compress=FALSE)` is significantly faster than the default `compress=TRUE)` - at least on my local SSD drive. @kendonB, can you run some benchmarking on your...

Do they specify whether they use the default `compress=TRUE` or the `compress=FALSE`? Because that's a low-hanging fruit that I can imagine @mllg could implement as an option in batchtools without...

Just to clarify, I'm _not_ trying to argue against `qs`, which looks great (and it's not my decision). I just like to see some real-world benchmark, especially on your 20...

> @HenrikBengtsson can you explain how the exports work in `future.batchtools`? Or is the foreach connector the problem in example 1? Without having gone through the details in the above...

Thanks for adding this. Will it be possible to set the default of these `compress` argument(s)? That way the end-user would be able to change this too even if they...