Compatibility between IBRION=6 and LEPSILON

[jzwanzig]

Hi,

my understanding from the documentation is that IBRION=6 should be compatible with LEPSILON=.TRUE.; I'd like to use these features together as a simple way to compute the elasto-optic tensor by finite differences. However, when I combine them, the code runs and the elastic tensor is computed correctly but there is no LEPSILON-related output. I'm a bit new to VASP so not too sure at this point how to proceed. Any suggestions would be very welcome.

thanks,
Joe Zwanziger

[admin]

Thank you for reporting the incompatibility. It will be corrected in the documentation.
LEPSILON=.TRUE. combines with IBRION=8.
When using IBRION=6 you can initialize the calculation of the dielectric tensor
using LCALCEPS=.TRUE.

[jzwanzig]

Yes, LCALCEPS indeed gives a computation of the dielectric tensor for the original unstrained cell at the beginning of an IBRION=6 run, but does not appear to compute anything for each of the strained cells. Or am I missing something? Thanks--