The way how parallel local execution should be modeled is to have a LocalJobManager with a limited size thread-pool and a job queue. With the current model instead we have a runParallel() method in the very-very-very client code, leaking job execution model into the the bioinformatic code.
The way how parallel local execution should be modeled is to have a LocalJobManager with a limited size thread-pool and a job queue. With the current model instead we have a
runParallel()method in the very-very-very client code, leaking job execution model into the the bioinformatic code.Originally posted by @vinjana in https://github.com/TheRoddyWMS/BatchEuphoria/pull/129/files