With autormt I get the same error
Total charge density incorrect for s.c. loop 15
Calculated : 12.00012744
Required : 12.00000000
If I set autormt to false, I get no errors, but again, 12 hours later the calculation is still running.
Running the serial calculation does seem to run OK. All mpi parallel calculations appear to stay on task 320 indefinitely. I have tried excitingmpi and excitingsmp, although I must admit I am not too familiar with smp jobs. The smp job actually finishes, but there is no performance gain seen at all. I do export the number of threads=processors.
So it would appear the problem exists in the parallelizing of task 320. The ground state seems to parallelize fine in excitingmpi.
I am using mvapich2 for mpi
mpiexec -npernode pn excitingmpi
where pn is the number of processors per node.
I did not have this problem with the lithium release. This is new and possibly related to the beryllium release. I will run this same job in the lithium release and see if any problem exists.
I have not tried the species file you suggested. What is the reason for changes in this species file?