Different NGsBlkXp in input and report

[AzulineSky]

Hello,

I noticed that during a GW calculation the report file shows a different input value for NGsBlkXp than the actual input file. I attached the input and report for a fast test calculation for diamond. In the input file I set NGsBlkXp = 4 Ry. The end of the report states NGsBlkXp = 6 Ry. Why does this happen?

Best regards,
Max

[Daniele Varsano]

Dear Max,

this is probably due to the fact that Yambo includes extra G vectors to close the shells. In your case, I can see that you are using a very small number of Gvectors. Here, 4 Ry (2Ha) is in between the third and fourth shell of Gvectors. Yambo include all the gvectors to close the fourth shells (27 gvectors). These can be seen in the report file.
To have a better control you can assign the matrix size in terms of RL (number of reciprocal lattice vectors) or in mHa, mRy.

Code: Select all

G-vecs. in first 23 shells:  [ Number ]
     1    9   15   27   51   59   65   89  113                            
    137   169   181   229   259   283   307   331   339
    387   411   459   531   537
  ...                
  Shell energy in first 23 shells:  [ mHa ]
    0.00000      1307.89      1743.85      3487.71      4795.60      5231.56      6975.42      8283.31      8719.27
    10463.1      11771.0      13950.8      15258.7      15694.7      17438.5      18746.4      19182.4      20926.3
    22234.1      22670.1      24414.0      25721.8      27901.7

Best,

Daniele