bugfix: UNT calculation

[kzm7k6r]

The ic loop should be inside. Pls confirm
!
!20181211 Kazuma Nakamura
!
!do ic=1,ncomp
do ik=1,NTK
do ib=1,N_BAND(ik)
do jb=1,n_occ
!ic=1
SUM_CMPX=0.0d0
do ic=1,ncomp !20181211 Kazuma Nakamura
do ig=1,NG0(ik)
SUM_CMPX
+ =SUM_CMPX
+ +conjg(C0(ig,ic,ib+N_BAND_BTM(ik),ik))*C0WN(ig,ic,jb,ik)
! + +conjg(C0(ig,ic,ib+N_BAND_BTM(ik),ik))*C0WN(ig,jb,ik)
enddo!ig
enddo!ic
UNT(ib,jb,ik)=SUM_CMPX
enddo!jb
enddo!ib
enddo!ik
!enddo!ic

[mnmpdadish]

I think that you are right about this bugfix. I included it in the version, but I need/want to test it to see if this is correct. This should have a significant effect on the result of chiqw. TBC.

[kzm7k6r]

When doing cRPA, we have to fix this bug.

[mnmpdadish]

ok I test a cRPA case then, and check the difference.