Darfix memory consumption
@acampbel Reports darfix jobs running with 400 to 600GB per CPU on the cluster.
For SLURM 30GB/CPU is more or less the reasonable maximum.
This probably means refactoring the code to not keep an entire dataset in memory.
@acampbel Reports darfix jobs running with 400 to 600GB per CPU on the cluster.
For SLURM 30GB/CPU is more or less the reasonable maximum.
This probably means refactoring the code to not keep an entire dataset in memory.