Skip to content

Why does the time to run all-by-other comparisons vary so much among checkpoint re-runs? #309

@svedwards

Description

@svedwards

Hello. I have had to re-run RM2 a few times due to various (fixable) issues on my end. Pointing it to the previously made directory is otherwise working well. I have re-run round-5 several times with both 16 and 32 cores, and the time for conducting the all-by-other comparisons varies widely each time. The first time with 16 cores it said about 250 hours (this is for a 1G bird genome - should not have many repeats). Then with 32 cores it said over 1300 hours. The third time with 32 cores it said about 250 hours again. And now running again with 32 cores it is back up to 675 hours. This is all with Version 2.0.7, and using basically the same set of cores each time. Thanks for any insight.

My command is:

singularity exec docker://dfam/tetools:latest RepeatModeler -engine ncbi -database Rhea_db3 -threads 32 -recoverDir RM_559040.ThuApr20232162026

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions