ABINIT Discussion Forums

Dear all

I tried to run through the tutorial on Abinit+wannier90 interface, tw90_4.in, with abinit-6.12.3.(I am about to switch from abinit-6.6.2). I encountered the following warning and error and the run was terminated.

--------------------------------
mkdenpos : WARNING -
Density went too small (lower than xc_denpos) at 88 points
and was set to xc_denpos= 1.00E-14. Lowest was -0.15+296.
Likely due to too low boxcut or too low ecut for pseudopotential core charge.
scfcv : after setvtr, energies%e_hartree= 0.000000000000000E+000
vtorho : nnsclo_now=200, note that nnsclo,dbl_nnsclo,istep= 0 0 1
fxphas : BUG -
fxphas : BUG -
The eigenvector number 1 has zero norm.
The eigenvector number 1 has zero norm.
Action : contact ABINIT group.
Action : contact ABINIT group.

.Delivered 9 WARNINGs and 5 COMMENTs to log file.

leave_new : decision taken to exit ...
---------------------------------------

The above error seemed to occur during the non-self consistent calculation in dataset2. However, the run can be completed if the input variable "symmorphi 0" is commented out. Anyway, I checked the reference output file tw90_4.out provided within the package, and found that the run of tw90_4.in was completed with the "symmorphi 0" imposed. May I note that the reference output file was generated with Abinit-6.11.2 (MPI version, prepared for a x86_64_linux_gnu4.4 computer). So I feel there may be something not right with my Abinit. Could this be related to the compilation?

My system is Abinit-6.12.3 (MPI version, x86_64_linux_intel12.0, on centOS5.0(Rocks clusters 5.4). The full log file is attached.

All comments and suggestions will be greatly appreciated.

Kind regards,
Thanusit

Dear All

The above errors and some other odd behaviours don't occur at all after I compiled Abinit with gnu4.4. Well, I have learnt the lesson.

kind regards,
Thanusit

Hi Thanusit,

for info, the use of version intel 12.1.4 under CentOS 6.2 works with abinit....

regards

jmb