Enough kinks are out of the SGE queuing for I-TASSER that it finally completed a parallel run on our cluster.
One thing I notice is that all 14 simulations ended at the time limit of 5 hours. Is there an endpoint condition for each simulation that would cause it to run for less time, or will they always run for the full set time?
How was the combination of 14 runs x 5 hours selected? Would 28 x 2.5 be equivalent? What would be lost in going to 14 x 1, or 4 x 5?
Thanks.