"matrixuhu" in wannier is taking very long time
I have two queries regarding wannier as implimented in fleur package.
"matrixuhu" option in wannier class in inp.xml is consuming lot of time to finish the job. I did not see any improvement in speed with mpirun command also.
In Fe tutorial. sp3d5 projections are considered. So there are 9 states with spin degeneracy it would be 18. May I know why you have chosen 2*18+(9 LOs) bands for maxSpinUP?
For example, I have a Heusler alloy Co2MnSi and I have a following projections:
projgen_inp file: Co 2 2 0 Co 2 3 0 Co 2 5 0 Co -5 0 0 Mn 2 0 0 Si -3 0 0
Here I have used the projections as follows: first three are dxy, dyz, dxz on Co. Next sp3d2 for -5 0 0 on Co. All 5 d orbitals on Mn, sp3 hybrid orbitals on Si. So, total I have 9*2+5+4=27 orbitals. Due to spin degeneracy it would be 27x2. So, do I need to consider 2x54+LOs bands for maxSpinUP?
I am using spin orbit coupling in system.
Thanking you sir.
With best regards Dr. Y. Venkateswara, Department of Physics, IIT Kanpur, INDIA.