Ypp -r

[sdwang]

Dear developers,
I am now trying to use E-RIM method to converge BSE of 1 and 2 dimensional sysytem, but I found ypp -r generates random k-poins for 1D system like,
# c1 c2 c3 wt
0.3870484 0.5192293 -0.7772745 0.0125000
-0.0272696 0.5241297 -0.2687416 0.0125000
-0.7236645 0.6694298 0.3861077 0.0125000
....

What I understood is for 1D sysytem, the k-points should be (0,0,n) where n is the periodic direction, but ypp -r are not like this.
Did I make a misunderstood? Or ypp -r is only for 3D system?
Thanks in advance!

[Daniele Varsano]

Dear Shudong,
I'm not expert on that, anyway it looks me that as you argued ypp -r generates 3D grids.
I think anyway that you can provide to the obtained k-point list to pw considering only the sampling in the directions you need (ie 0,0,kpoint).
I could be wrong, but the 1D list should be still uniform, next you go on with the procedure.
If it does not work, please post it again, and people who coded it will answer to you.

Best,
Daniele

[sdwang]

Hi,
1). I think the generated k-points even only along one direction are also not the same as the generated by pw itself. Does this random k-poins have been generated in the whole box like unitcell instead of considering its dimension?(may be I am wrong).
2). I also used wannier interpolation to calculte BSE. When I use ypp to get the interpolated k-poins of 1D system, it complaints as:
....
---> [03] K-point grid
<---> :: Q-points (IBZ): 24
<---> :: X K-points (IBZ): 24
<---> [04] CORE Variables Setup
<---> [04.01] Unit cells
<---> [04.02] Symmetries
<---> [04.03] RL shells
<---> [04.04] K-grid lattice
<---> [04.05] Energies [ev] & Occupations
<---> [05] Read wannier interpolation file
<21s> :: Wannier-interpolated K-points (BZ) : 92 92 92
<21s> :: Internal (coarse) K-point grid (BZ): 46 1 1
<21s> :: Range of Wannier-interpolated bands: 31 55
<21s> :: Wannier-interpolated bands: 25
<32s> [06] Employ double grid method
<32s> :: Perform polynomial interpolation of order : 2
<32s> Not enough knots to interpolate with order 2
<32s> :: ERROR while preparing double grid. Error: 4
<32s> [07] Game Over

my system is 1D along Z direction, it seems the interpolated k-points also not considering the system dimension. For 3D system, it is working well.
I have checked the source code in ypp/double_k_gid.F, and it relates to
...
!
if(order_ipol + 1 .gt. minval(nkpoint_c))then
call msg('s','Not enough knots to interpolate with order', order_ipol)
err=4
...

Could any one explain this ?
Thanks!

[Daniele Varsano]

Dear Shudong,
the output of the generated k-points is controlled by the cooOut variable. Please check the documentation and set the variable in the frame you need. I'm not expert at all of the interpolation part, try to fix point 1 and hopefully point 2 it is related. If not successful post here again the problem and you will receive help form someone expert on that.

Best,
Daniele