So, we have:
So, we likely will want to go forwards with the Future approach. We are however not fully aware how well this works across different computers (laptops, servers, HPC, etc.) since the behaviour seems to be computer-dependent. @janaobsteter is testing this ... so we will wait to hear back and then decide how to move onwards.
So, we have:
https://github.com/HighlanderLab/SIMplyBee/tree/Parallelisation was the first attempt to use parallelisation using
doParallelthat by default usesforking, but this proved challenging for use on HPChttps://github.com/janaobsteter/SIMplyBee_fork/tree/onePSOCK was then the second attempt (by creating a branch from the
Parallelisationbranch) to address theforkingissue by usingPSOCKapproachhttps://github.com/janaobsteter/SIMplyBee_fork/tree/Future was then the third attempt (by creating a branch from the
onePSOCKbranch, likely) since theFuturepackage enables much simpler code than thePSOCKapproach - also, we can simply plan if a script is run in serial/sequential or parallel mode, so looks like a no brainer.So, we likely will want to go forwards with the
Futureapproach. We are however not fully aware how well this works across different computers (laptops, servers, HPC, etc.) since the behaviour seems to be computer-dependent. @janaobsteter is testing this ... so we will wait to hear back and then decide how to move onwards.