ABINIT Discussion Forums

In the release notes for the 7.0x series it says that one of the big changes is that Abinit now uses k-centered wavefunctions from _WFK files rather than the gamma-centered sets and hence the mpi-io bottleneck is gone. However, I can't see where this is documented - are we still supposed to generate...

It still happens in 6.12.3. As I've said in another thread, it's hard to show exactly what is going on because the hangs/crashes only happen on certain systems and might be separate phenomena. We have a patched version of 6.12.3 here that seems to avoid the problem but we're not sure why it works ye...

Ah, yeah, it's ONETEP. Great code! Why don't you just ask Nick to write an all-electron .cube output? ONETEP does PAW these days doesn't it?

By "slab", I just mean a cell that contains a thin section of your material and a vacuum gap - the same kind of geometry you use for studying surfaces in DFT. An excellent example (for CASTEP, not ABINIT) is given here: m As you can see in the graphic there is a thin section of material, i...

Hi, The way people normally do a surface bandstructure calculation is to do a normal bandstructure calculation on a surface slab geometry. You should use a 2D BZ path because there should be no dispersion perpendicular to the slab for any surface states. There are tutorials on how to calculate bands...

OK. Well, the best route might be to make a 'shadow' _DEN file - generate a simple Abinit input file for your model system (use the latest Avogadro GUI - it has a quick Abinit input file generator) and run a single SCF step. This will generate a _DEN file. Then write a fortran code to use the header...

Hi, Ok, so you want all-electron bader. You have a couple of options. The first (probably not very helpful) is that if you generate the density via an actual abinit calculation rather than just taking the .cube file, you can get an all-electron density. To do this you need to run a PAW calculation a...

Ah sorry I thought you wanted to use *bader* (m) to do bader analysis, not aim. As far as I know, there are no utilities to convert from a cube or dx to _DEN. It wouldn't be generally of much use to have such a utility apart from when you're trying to graft the output of one code to another without ...

Isn't it the other way around? You have a _DEN and want a .cube for bader? If so you can use cut3d - option 10 is OpenDx and option 14 gives you a cube, at least on the most recent public Abinit (6.12.3).

Hi Matthieu, We've made some progress in this recently but we have not yet isolated the problematic sequence of events. It appears the problem is that different MPI threads are blocking each others attempts at writing during the MPI_FILE_WRITE_ALL call. We don't know exactly how yet. It only happens...

Hi, This could be similar to the MPICH2 bug with the hvector type definition (although you're not using MPICH). Try setting: with_mpi_level="1" and reconfigure. Unlike the MPICH case you probably can't just patch your MPI libraries to fix the fundamental issue here so the with_mpi_level ch...

HI Elena, I'd first point out that while XCrysden is great, both VMD and VESTA are better, especially on a Mac. They don't suffer the same kind of slowdowns you get from XCrysden when loading a very large .xsf file. Both can read .xsf and VMD can also read the animated version (.axsf). VESTA is grea...

Hi all, We're getting a MPI-IO hang when MPI_FILE_WRITE_ALL is called on line 229 of the wffreadwrite_mpio.F90 routine in Abinit 6.12.3. The file to be written is on a GPFS file system and we only get the hang when we run across multiple nodes on our (infiniband-connected) cluster. It's been rather ...

Hi Mohamed, You are going about compilation the wrong way - look at the forum banner in red at the top of this page. If you want to compile with respect to MPI, you should use the mpif90, mpicc, mpicxx wrappers provided by your MPI distribution. The compilation step you posted above the error messag...

Hi all, I've been checking and making some changes to the optics_paw_core and conducti_paw_core routines and have been checking the outputs against a simple carbon-carbon dimer in a big box. The conducti utility now works as expected, but I've noticed that the KS orbitals are not aligned along the x...

You've specified that the two extra projectors for the s and p channels have the same energy (0.5 Ha for s and 1 Ha for p). Each projector should have a different and sensibly-chosen energy. Go and do the PAW tutorials for Abinit and pay close attention to the one which deals with constructing a PAW...

Ah it's in ./tests as well.

Hi Yann,

You'll need to apply that patch against tests/config/m4/util-fixes.m4 as well. It has the same problem.

Kane

Hi, Your config log is full of one particular kind of error that seems to be causing the Abinit configure script to choose the wrong kind of settings: conftest.c:1: error: bad value (native) for -march= switch Basically when configure tries to run a test on gcc it uses the flag -march=native which a...

I concur. How about you post your config.log for the misconfiguration? It might contain something helpful.

As a follow up to my own question, the methods are implemented (but undocumented) in the conducti utility, with options 4 and 5.

Hi, As noted in another thread ( http://forum.abinit.org/viewtopic.php?f=17&t=1609 ), there is a build bug in all recent versions of Abinit. The m4 macro AX_PROG_MKDIR_P from util_fixes in both the fallbacks subtree and the tests subtree incorrectly tries to assign MKDIR_P a prefix of "fall...

Hi, It seems your build environment is broken for some reason (could be a lot of things). Where did you get the fortran compiler you're using? Instead of using the command line options, copy the doc/config/build-config.ac file to the top level directory of your abinit code folder and rename it HOST....

The fallbacks LibXC compilation in Abinit is not great, just compile it separately and point to it in your build.ac file with --with-libxc-incs="-I/path/to/your/libxc/installation/include" --with-libxc-libs="-L/path/to/your/libxc/installation/lib -lxc". The errors you're seeing d...

Hi Elena, If you only need the WFK file from the final density (no other properties) you can just restart the calculation with iscf -3 and use tolwfr (instead of tolvrs/toldfe) with a suitably small number to make sure the KS orbitals are converged. Note that you may need to comment out other parts ...