Wannier90 band structure calculation taking too long

Dear colleagues,

I am trying to have some experience on wannier90 tutorial here section 6 in wannier90 - abinit). For other tutorials (tw90_3,4,5) regarding Si and SiH4 it took few minutes, while I am running tw90_6.abi it seemingly takes forever. However, it says it should take couple of minutes in the documentation. I did not change any variable in the input file, I took it as it is.

About my system: I am using high power computing platform and here is the batch script.
#SBATCH --job-name=MPIJOB
#SBATCH --time=24:00:00
#SBATCH --mem-per-cpu=12000M
#SBATCH --nodes=4
#SBATCH --cpus-per-task=1
#SBATCH --ntasks-per-core=1
#SBATCH --ntasks-per-node=8
I have given 32 processors with 12GB memory for this calculation. I am using abinit 9.6.2.

As you can see my calculation will stop after 24 hrs and it did before finishing. Could someone please tell me what could be the problem?


Hi @rajeshatpsi! Could you share the abinit log files / output files that you get and any wannier90 output files if relevant. Thanks

Hi @ fgoudreault,

Please see the attached files.

tw90_6.abo (123.3 KB)
tw90_6o_DS3_w90.win.in (602 Bytes)
log.abi (481.7 KB)
tw90_6.abi (4.8 KB)

I will send the other wannier90 related files in next post as I can’t upload more than five.

Here is the other files:
tw90_6o_DS3_w90.nnkp.in (162.3 KB)
tw90_6o_DS3_w90.wout.in (30.1 KB)
tw90_6o_DS3_w90.amn.in (177.6 KB)
tw90_6o_DS3_w90.eig.in (67.5 KB)

Dear All,

Just to let others know if someone finds the similar issue. I have figured out that program actually looking for UNK files as I have asked for wannier plot as well. After setting “w90prtunk3 ==1” , it worked.


Dear Rajesh,

Thanks for highlighting this. Perhaps it would be worth adding a note to the user. I will look into it.