mirror of https://github.com/abinit/abinit.git
72 lines
2.7 KiB
Plaintext
72 lines
2.7 KiB
Plaintext
|
|
This directory contains the file needed for the build-in tests of the
|
|
abinit code (sequential version).
|
|
|
|
Please, see the README file in the tests directory (above this one) to get help.
|
|
_____________________________________________________________________
|
|
|
|
A set of seven build-in tests exist in the abinit code.
|
|
These short tests should be used as frequently as possible
|
|
in the course of development or modification of the existing code.
|
|
|
|
1) Usage
|
|
|
|
In the main directory, type the command `make` with one of the following keywords:
|
|
|
|
- test_fast
|
|
- test_v1
|
|
- test_v5
|
|
- test_libxc
|
|
- test_bigdft
|
|
- test_wannier90
|
|
|
|
like:
|
|
|
|
$ make test_fast
|
|
|
|
The code (sequential version) will run automatically, and print on screen
|
|
different status files during the execution, followed by an automatic check
|
|
of the exactness of the results obtained (total energy, eigenvalues,
|
|
nuclei positions, forces and stresses). If everything is OK, the final output
|
|
on screen is rather short:
|
|
|
|
--------------------------------------------
|
|
Status file, reporting on built-in test fast
|
|
|
|
==> The run finished cleanly.
|
|
Moreover, comparison of the total energy, and other (few) relevant quantities with reference values has been successful.
|
|
This does not mean that no problem is present, however.
|
|
Please run the complete set of ABINIT tests to gain a better confidence in your installation.
|
|
|
|
If there is a problem, the final output on screen will describe briefly
|
|
the error.
|
|
--------------------------------------------
|
|
|
|
The computed values are converged up to some level of convergence.
|
|
The code takes into account predefined levels of convergence,
|
|
in such a way that this suite of test should still be valid if
|
|
the algorithm in the code is modified, even profoundly.
|
|
Indeed, only the final result is checked, not the intermediate results.
|
|
Unlike the basic algorithm, the parameters of the calculation (ecut, ngfft, nkpt ...)
|
|
should not change for this suite of test to be valid.
|
|
|
|
2) Description of the tests
|
|
|
|
test_fast: H2 molecule, determination of the H-H distance by the Broyden algorithm.
|
|
test_v1: Yb cristal, self-consistent
|
|
test_v5: HBe molecule, spin-polarized, determination of forces
|
|
test_bigdft: H atom within BigDFT
|
|
test_libxc: Bi atom with PAW, GGA PBE from LibXC
|
|
test_wannier90: GaAs crystal with few bands.
|
|
|
|
|
|
The physical content is weak, because the cut-off energies are unreasonably low.
|
|
|
|
|
|
3) Related information
|
|
|
|
The files that define the suite of tests are contained in the built-in directory.
|
|
These should NOT be changed, even for debugging. It is much better to do the debugging
|
|
in the fast directory, or to duplicate the built-in directory, and not to rely on the
|
|
"make" mechanism for the debugging.
|