[Pw_forum] about projwfc.x parallelization.

Mutlu COLAKOGULLARI mutlucolakogullari at trakya.edu.tr
Fri Jul 4 16:54:35 CEST 2014


Hello,

I was playing with PP/example02 and I realized a difference (I am using 
the latest svn version)...I given eye to projwfc.f90 but I couldn't find 
the solution...I think I am missing something but what?

PROBLEM:

If I use the "mpirun -n 8 projwfc.x <ni.kpdos.in " then I got the charge 
just on bands

Lowdin Charges:

      Atom #   1: total charge =   4.0634, s =  0.5671, p =  1.9492, d =  
1.5393, f =  0.0078
      Atom #   2: total charge =   4.0572, s =  0.5624, p =  1.8897, d =  
1.5971, f =  0.0080
      Atom #   3: total charge =   4.0269, s =  0.5868, p =  1.8322, d =  
1.6000, f =  0.0080
      Atom #   4: total charge =   4.1149, s =  0.5961, p =  1.8426, d =  
1.6684, f =  0.0076
      Spilling Parameter:  -0.0164

BUT If I use the "mpirun -n 8 projwfc.x  -nk 1 -nd 1 -nb 1 -nt 1 
<ni.kpdos.in " then I got the charge on bands and sub-bands either.

Lowdin Charges:

      Atom #   1: total charge =   4.0634, s =  0.5671,
      Atom #   1: total charge =   4.0634, p =  1.9492, pz=  0.5950, px=  
0.6337, py=  0.7206,
      Atom #   1: total charge =   4.0634, d =  1.5393, dz2=  0.4008, 
dxz=  0.2394, dyz=  0.2183, dx2-y2=  0.4784, dxy=  0.2023,
      Atom #   1: total charge =   4.0634, f =  0.0078, fz3=  0.0015, 
fxz2=  0.0008, fyz2=  0.0012, fzx2-zy2=  0.0011, fxyz=  0.0004, 
fx3-3xy2=  0.0009, f3yx2-y3=  0.0019,
      Atom #   2: total charge =   4.0572, s =  0.5624,
      Atom #   2: total charge =   4.0572, p =  1.8897, pz=  0.6015, px=  
0.6348, py=  0.6534,
      Atom #   2: total charge =   4.0572, d =  1.5971, dz2=  0.3933, 
dxz=  0.2446, dyz=  0.2945, dx2-y2=  0.4629, dxy=  0.2019,
      Atom #   2: total charge =   4.0572, f =  0.0080, fz3=  0.0015, 
fxz2=  0.0008, fyz2=  0.0010, fzx2-zy2=  0.0015, fxyz=  0.0004, 
fx3-3xy2=  0.0009, f3yx2-y3=  0.0019,
      Atom #   3: total charge =   4.0269, s =  0.5868,
      Atom #   3: total charge =   4.0269, p =  1.8322, pz=  0.5863, px=  
0.5796, py=  0.6662,
      Atom #   3: total charge =   4.0269, d =  1.6000, dz2=  0.3741, 
dxz=  0.3236, dyz=  0.2714, dx2-y2=  0.4552, dxy=  0.1756,
      Atom #   3: total charge =   4.0269, f =  0.0080, fz3=  0.0018, 
fxz2=  0.0008, fyz2=  0.0011, fzx2-zy2=  0.0016, fxyz=  0.0005, 
fx3-3xy2=  0.0006, f3yx2-y3=  0.0015,
      Atom #   4: total charge =   4.1149, s =  0.5961,
      Atom #   4: total charge =   4.1149, p =  1.8426, pz=  0.5829, px=  
0.5688, py=  0.6909,
      Atom #   4: total charge =   4.1149, d =  1.6684, dz2=  0.4034, 
dxz=  0.3344, dyz=  0.2373, dx2-y2=  0.4827, dxy=  0.2108,
      Atom #   4: total charge =   4.1149, f =  0.0076, fz3=  0.0016, 
fxz2=  0.0007, fyz2=  0.0012, fzx2-zy2=  0.0013, fxyz=  0.0003, 
fx3-3xy2=  0.0007, f3yx2-y3=  0.0018,
      Spilling Parameter:  -0.0164


QUESTION: Why this difference is happening while the exact 
parallelization is setted up as "-nk 1 -nd 1 -nb 1 -nt 1"?

Best wishes,

      Mutlu.

-- 
Dr. Mutlu COLAKOGULLARI
Trakya Universitesi
Fen Fakultesi
Fizik Bölumu



More information about the users mailing list