These are chat archives for dereneaton/ipyrad

11th
Sep 2017
Isaac Overcast
@isaacovercast
Sep 11 2017 20:49
@tommydevitt Unfortunately step3 doesn't have very granular checkpointing at this time, so restarting a failed run will just restart from the beginning. It would be best to allocate more cpu/RAM or use MPI to obtain more CPUs across nodes. If this isn't possible you could try branching subsets of samples, running step 3, and then merging the subsets, but this seems pretty annoying.