enter image description here

Total DoS for two different q-point grids, $100\times100\times100$ and $101\times101\times101$, are graphed above. The DoS for $63\times63\times63$, $65\times65\times65$, $71\times71\times71$, $72\times72\times72$, $81\times81\times81$, $85\times85\times85$, and $101\times101\times101$ are same (barely differences of $0.0000000001$). Even numbered grids have a similar DoS too but different from odd ones ($100$ vs. $101$ above). $72\times72\times72$ is the only odd (english) one that matches odd (math) grids.

Any reason for all of this? Which one should I use?

Find some data below (click to zoom in!):

enter image description here enter image description here enter image description here enter image description here

  • 3
    $\begingroup$ That color scheme is brutal to look at, yellow on white is hard to read. Might confuse people taking quick glances $\endgroup$ Oct 7 '20 at 1:17
  • $\begingroup$ @TristanMaxson, haha, that is true, but I always find green, blue, and red challenging to differentiate, whereas yellow somehow works for me. In final presentations, I go for the classic RGB. $\endgroup$ Oct 7 '20 at 12:49
  • 2
    $\begingroup$ Sadly, I know nothing of this problem though. Maybe it is a similar issue to k-points, does a shifted grid make sense in this context? That might be helpful to plot. $\endgroup$ Oct 7 '20 at 14:05
  • 1
    $\begingroup$ This question appears to be abandoned. It can be reopened with clarification from the OP or if another user would like to provide an answer. $\endgroup$
    – Tyberius
    Jan 23 at 13:12
  • 1
    $\begingroup$ You don't necessarily need to delete this question, but I think a new question about odd/even meshes more generally would be interesting. $\endgroup$
    – Tyberius
    Jan 23 at 19:31