GitLab

There was again a case where stack-variables were loaded with
instructions that needed properly aligned memory. This only surfaced
with the Intel C compiler, where the stack layout evidently was
sufficiently different to trigger this.

This was also the case for SSE kernels

- remove runtime by not executing unecessary runs
- make sure that on gp02 impi/2017 is NOT use
- adjust reserved time on knl systems

The new Makefile rule just dumped more and more lines into the test_x.sh
script, as the ">>" operator was used always.

Pavel Kus authored Jul 20, 2017 and Lorenz Huedepohl committed Jul 25, 2017

and moved the implementation to test_util.F90

Due to the extremely slow compilation on KNL this error now surfaced, as
libelpatest.la was not ready in time, there :)

Previously we checked ortonormality of eigenvectors by comparing
the matrix S^T * S to identity matrix. The new feature is also
checking just the diagonal of the matrix S^T * S. By that we get
the information, how far are the eigenvectors from having length 1.
If it turns out to be important, we could try to normalize them
at the end of elpa, which is simple (in contrast with enforcing
better orthogonality).

This stupid bug was there since commit ae45bbb3

From now on call make check with TASKS=x, e.g.

make check TASKS=2

for mpi_allreduce in test_check_correctness

of _GEMM and P_GEMM calls

The module is intended to be hidden, thus I moved it into the public
part of the legacy API.

Some test programs use it, thus the test programs have now also access
to ELPA's private modules. This is of course anyway necessary to test
ELPA internals that are not exposed via the public API.

The functions in elpa_utilities are considered "ELPA internal",
i.e. the should not be accessible by the users and thus not
be part of the API.

Should be done in a more systematic way. In this case, in certain
configuration 1stage_analytic test pased, while 2stage_analytic tests
failed due to error larger then tolerance