ABINIT Discussion Forums

As far as I remember, there is a small bug in the FFT initialization. If you have a GPU version of abinit, you should not make any FFT on CPU since it is far more efficient on GPU therefore you should not use fftw3. If you compile abinit without fftw3 and with cuda it should work. Try to set fftalg ...

MPI is a standard not compiler related library. Which means that there is a standard. All MPI implementation (openmpi,mpich,intelmpi) proide wrapper for mpif90 mpicc mpic++ When the MPI library is compile, it "remembers" with which compiler it's been compiled so when you compile with ifort...

Sorry we cannot help more. the problem is that this issue is very specific to one abinit executale. If you give us your starting point and input file and we run it, we probably won't be able to reproduce the bug. Neverthelesse, if you work out to produce a minimal input file (taht can be run on a fe...

Hi, When you just run abinit you have to type on your keyboard a few things : the path to input file, output file, root name for input files, root name for output files, root name for log files and all the path to the pseudo-potential you want to use (this is the iofn2 message you get "please e...

Hi,

No, you cannot use hybrid XC with the current abinit. Maybe soon ...

Jordan

Hi I already encountered the problem but I could not resolve it. The best thing to do is to restart form the previous geometry. In you PAW example, I wonder if it doesn't come just from the default value of nstep which is 20 ? Do you set nstep to something higher than 20 in your input file ? Cheers ...

Hello You should always keep the ecutsm value from the relaxation in any additional run. It is very important to keep the pressure/max force residue the same as during the relaxation. You should always use the same pseudo for all your calculations on 1 material. Don't switch from NC to PAW. In NC, o...

To answer the question, yes you need to have the same ecutsm in relaxation and gs calculation. Check also that the FFT grid the same (ngfft)

Why no one helping out?? what is the use of this forum if I post many questions and barely get answers???? This is being rude to me or at least not very kind. We are not a paid technical support. Don't expect us to be one and to answer all the questions in this forum. The implementation of a featur...

Abinit is an open source software developed and maintained by many people in many countries for their activities. We generously share our work with the community. If nobody answers your questions this might be because nobody reading the forum has the answer or maybe because you are being rude. You m...

compile with libx

before make, configure with --with-dft-flavor="libxc"

Hi

band2eps is simply a tool to plot phonon band structure.
You have all the data you need in a file ended by _PHFREQ or .freqs to plot the band structure with any tool of you choice like gnuplot matplotlib xmgrace ....

Cheers

Intel16 compiler is still not fully supported, which might be you issue. Can you try to compile abinit with ./configure --prefix=/opt/para3/abinit-7.10.5/build --enable-mpi="yes" --with-mpi-prefix="/opt/apps/intel16/cray_mpich/7.3.0/" FCFLAGS_OPTIM="-O2 -no-vec" Also, s...

Hi, The CPU you have on the login node does not seem to be the same as the one you have on the computation nodes. Don't compile abinit with -xHost (which is I think a default option when using basic optimization and intel compiler) add FCFLAGS_OPTIM="-O2" when you configure abinit Cheers

For point 2) You can move the atoms keeping the same cell (optcell 0 and ionmov 2) or optimize both the positions and the cell parameter (optcell 1 ionmov 2) 3) ratsph is a variable of abinit for integrating the density inside a sphere around an atom. In the cas of PAW, ratsph is set to the PAW radi...

Hi The problem you have with PAW is that the PAW spheres are overlapping because you geometry is not optimized. If you perform a relaxation (read about optcell ionmov, ecutsm, dilatmx, ntimes, tolmxf toldff input parameters), forces between atoms will decreases and the distances between atoms should...

Yes you still have to define the cell parameter since the xyz file format does not provide it. So in your input file you should have xyzfile Diamond.xyz rprim .......... acell ......... nband ..... ecut......... nstep ....... ........ The xyz file contains cartesian positions which will be used as t...

Try to provide a input file with the basic variables like natom ntypat typat znucl acell rprim xred or xcart or xangst Or if you have a xyz file to be reuse ? You acell guesses for 1D and 2D are correct in the idea Once you have that, you can work on other input parameter which drive the calculation...

Abinit takes into account the magnetic symmetries. So maybe you AFM breaks some symmetries what results in a lower space group. Try without magnetism, you should recover the correct space group. Usually when constructing a AFM cell, you have to multiply the cell by at least 2 so yes your cell withou...

It hard to answer the question but as a starting reflexion, for general purpose converge the density, then if you are interested in the energy only, consider using toldfe. In the case of structural optimization, you just want the forces and streses to be as minimal as possible so use toldff or tolrd...

Doing the symmetry analysis with any symmetry tool (findsym, ....) gives I4mmm.
You can try abinit with changing tolsym to something larger than the default value which is quite small.
Their must be one digit that is wrong when running abinit that breaks symmetries.

Cheers

Dear all, Have you tried to configure with the --enable-static option ? and then compile. The problem you have is that you have a dynamic executable which needs to load the dynamic libatompaw library. Enabling static library should switch atompaw to a static executable with a static library. Cheers.

I won't be able to help you with all your question but i'll give a try 1) Yes you are right, DFT is not able, in theory to produce the band structure of the system. Indeed, we just know that the density and only the density matches between the real system and the fake KS system. But, surprisingly, i...

I don't know if it is normal or not but you can download some - maybe different - other paw atomic data here

Cheers

1) It is usually not recommanded to have "." in the path since you may have conflict with other search directories. 2) You use the old fashion test farm which is not upgraded anymore but migrated to the new test farm. So if you want to avoid the problem you mentioned, use the new test farm...