DDK eigenvalues: bug in EIG output?

jzwanzig · Post by **jzwanzig** » Tue May 05, 2015 8:08 pm

Hi,

when I run a ddk calculation (rfelfd 2 or rfddk 1, same thing) in the manner of test 5 of section 5 (v5, t05.in) in order to compare finite difference eigenvalues with ddk eigenvalues, I find a huge difference between the output from the .out file and the output in the EIG file.

For example, running v5[05], I find in the output file:

Expectation of eigenvalue derivatives (hartree) for nkpt= 32 k points:
(in case of degenerate eigenvalues, averaged derivative)
kpt# 1, nband= 4, wtk= 0.03125, kpt= -0.2500 0.5000 0.0000 (reduced coord)
-0.12504 -0.04341 0.21919 0.17883

While in the t05_1WF1_EIG file, I find:
Expectation of eigenvalue derivatives (hartree) for nkpt= 32 k points:
(in case of degenerate eigenvalues, averaged derivative)
kpt# 1, nband= 4, wtk= 0.03125, kpt= -0.2500 0.5000 0.0000 (reduced coord)
0.19049 -0.15096 -0.21281 -0.06461

The output file is consistent with the finite difference derivative computed from an NSCF run at close by k points, while the EIG output is totally off from this. It appears that the .outfile content is "correct" while the EIG file is not. But maybe I'm misunderstanding the EIG file, and it is reporting something different?

Can someone comment on this?

Joe