ABINIT Discussion Forums

Thank you for the hint. Understood something wrong.

Usually the acoustic modes in solids at gamma should be zero. Due to numerical inaccuracies the modes are near to zero. If omega is zero the partition function will have the value zero. sinh(0) = 0. But log(sinh(0)) is not defined. I do not quite understand how to deal with the acoustic frequencies ...

Hey, During the calculation of the phonon free entalphy with ANADDB I discovered differences for own values and ANADDB values. The code of ANADDB (src/77_ddb/thm9.F90, Var: gdos ) reveals that all phonon modes (also the acustic modes) are included in the calculation of thermodynamical properties. Si...

I just thinking about changing the mass to a high value should fix the atoms and stop the vibration. But I am not sure wether in RF case the atoms will displaced during calculation and so, the potential will change. Is this correct ?

Chris

Hello community I have problems concerning fixing atoms in response function calculation. I have tried to fix the atoms with natfix and iatfix in an RF calculations but it failed. Apparently, the natfix and iatfix options only fix atoms during structur optimization. My needs are to fix all atoms exc...

I find a solution.
I set occopt to 3 and tsmear to 0.01 to smear occupation.

Hello, I have some problems with the convergence of a GaAs supercell. The Input file is: ########################## Definition of cell ################################# acell 2.13678892749911E+01 2.13678892749911E+01 2.13678892749911E+01 bohr rprim 1.0000000000E+00 0.0000000000E+00 0.0000000000E+00 ...

Hi, I compiled Abinit 7.2.1 with MPI succesfully and wanted to make a benchmark test. I created a input file and ran it on the same machine with a different count of processors. I chose 16 and 64 processors. Now, the results are completely different. Input file: ########################## Definition...

Hi, again. It seems to be that the problem is not solved . After tests with much more pocessors the problem is back. I tried a lot of different configurations but get every time the same error. ---SELF-CONSISTENT-FIELD CONVERGENCE-------------------------------------------- getcut: wavevector= 0.000...

Hi, I have built Abinit 7.2.1 with IBM MPI and GNU compilers. Now, I encounter a problem with large scaling calculations. If I scale a calculation on more than 200 processors I will get some problems in the MPI communication. This problem only occurs when I do not set the parameter istwfk. If I set ...

I recompiled the code with the options enable_optim="standard" enable_mpi_inplace="no" But again no changes in convergence. Then I find out that IBM MPI use other compiler names as OpenMPI. In OpenMPI the compile names are: mpicc, mpiCC, mpifort In IBM MPI this compilers are with...

The log file

Hi Alain Thank you for your reply. Yes, the tests for serial calculation done successfully. Are there tests for parallel calculations also ? I ran the same input file with abinit compiled with OpenMPI 1.6.4 and it works fine. The result for the first iteration is (openMPI): === [ionmov= 2] Broyden-F...

Hi I have a problem while executing a parallel job. I compile Abinit 7.2.1 for a supercomputer with IBM MPI. The compilation is successful with no errors and serial calculation works fine. But if i try to calculate a parallel job the result of the scf loops oscillate and do not converge. The option ...