GWA in vasp5.2.8

Message

sdwang · #1 Post by **sdwang** » Sat Nov 20, 2010 11:01 am

Hello,

I have been trying band-structure calculation using GWA on VASP 5.2.
As a first step, the calculation was executed by the following settings.

[INCAR]
general:
System = a
NBANDS = 120
ISMEAR = 0 ; SIGMA = 0.05
LORBIT = 11
LOPTICS = .TRUE.

As the next step, I added "ALGO = GW0 ; NOMEGA = 50" in the INCAR file, and performed GWA calculation.
But it appears 'ERROR in DETERMINE_SLOT_INTER: the frequency range is not sufficient' in the log file and it stoped.

I wonder if anyone can find what is wrong with my calculation.
Thank you very much.

tilak · #2 Post by **tilak** » Wed Nov 24, 2010 4:28 pm

"the frequency range is not sufficient" this massage tells, you need to increase the value of NOMEGA (may be double !). Even this sysytem is too heavy, I think you could check it.

PS: Do you have well conversed normal DFT calculation before going to GW0 calculation? Have you checked the EPSILON file to see Re and Im Epsilon ? How is it looks like ? What is the value of CSHIFT ? It should be same during GW0 calculation.

sdwang · #3 Post by **sdwang** » Thu Nov 25, 2010 7:32 am

CSHIFT =0.1 in all the calculations.I set NOMEGA=80, but it always appears the error. Is the biggest value of NOMEGA 100?If not it could be set more bigger than 100?

metosa · #4 Post by **metosa** » Thu Nov 25, 2010 10:23 am

Dear sdwang,

GW issues are really cumbersome to deal with. I prepared some tutorial-like README.txt in http://db.tt/eSTQbyx . Have a look at there and compare your INCARs with mine.

Regards.

sdwang · #5 Post by **sdwang** » Sun Nov 28, 2010 3:45 am

Thanks! But I can not open it.

metosa · #6 Post by **metosa** » Sun Nov 28, 2010 4:47 am

Then try this:

http://unam.bilkent.edu.tr/mt2/random/GW/