I noticed that during the first round you do the refining step on just one thread. I was just wondering if you think it is worth parallelising this step, as far as I can see from the output files each family refinement is ran independently, it would be a matter of just adding an additional step to produce one or two additional file per family and gathering them in the end. Or do you think the main bottleneck here is the I/O?
Hello,
I noticed that during the first round you do the refining step on just one thread. I was just wondering if you think it is worth parallelising this step, as far as I can see from the output files each family refinement is ran independently, it would be a matter of just adding an additional step to produce one or two additional file per family and gathering them in the end. Or do you think the main bottleneck here is the I/O?
Cheers!