@dereneaton@isaacovercast What’s more, how to speed up running on step3? Although, I run ipyrad on a HPC with more than 500G RAM and assigned 30 cores to an average 5G compressed fastq dataset. It always spent more than one day doing this step. My data are pairddrad.
@ChaoShenzjs It sounds like you have allocated sufficient resources. 1 day for step 3 is well within the normal range of duration, especially for PE data, which can take longer to cluster because of the longer reads. If you really want to try to increase performance you might manipulate the -t parameter, which assigns more threads per clustering process. I think 2 is the default. You might get a performance bump by upping this to 4 or 8, but I wouldn't expect it to make a huge difference.