Currently each process has to read the ICs that have been cached. This is because we can't pickle data more than 4GB. We should figure out a workaround, because when all the processes try to read the same file at the same time, it makes the processes wait for the file to be unlocked.
Currently each process has to read the ICs that have been cached. This is because we can't pickle data more than 4GB. We should figure out a workaround, because when all the processes try to read the same file at the same time, it makes the processes wait for the file to be unlocked.