A:
- Check the config file: did you specify the correct executor (e.g. SGE, SLURM, ...)?
- Check the Nextflow output to see which specific process failed. Try to increase the memory allocated to that process in the config file. For example, if the process
make_borda
failed, changememory = '10 GB'
tomemory = '20 GB'
for that process in the config file.
Q: The global Borda rank and the cluster-specific Borda rank are not following each other exactly. Is this a bug?
A: No, this is not a bug. There is not necessarily a monotonic relationship between the global Borda rank and the Borda rank of a subset. The example below shows this in more detail for a simple example of Borda ranking.