GitLab

This closes issue #51.

This closes issue #51.

The various MPI libraries seem all to have some memory leaks, thus we
do not seem to be able to use LSAN in these tests.

Disable any MPI internal failures with a suppresion file, as the various
MPI libraries (unfortunately) are not ASAN-safe.

in elpa2_trans_ev_tridi_to_band_template

in elpa2_trans_ev_tridi_to_band_template

in elpa2_trans_ev_tridi_to_band_template

unlike elpa 2stage, scalapack computes only part of eigenvalues
if only part of eigenvectors are required. Check only those.

The blocking factor can be set now with the new interface
set method. The default is 3

Autotuning still has to be implemented.

This closses issue #44

This release contains bugfixes for ELPA 2017.05.001

adding variants for single precision and complex math datatype
for the scalapack test

in elpa2_trans_ev_band_to_full_template

in elpa2_trans_ev_band_to_full_template

in elpa2_trans_ev_band_to_full_template

in elpa2_trans_ev_band_to_full_template

taking into account, that computed eigenvectors can be arbitrarily
"rotated" by multiplying by complex number. Hopefully dealt with in
a robust way. Enabling analytic complex double and single

reshaped 1D array produced transposed array withouth the order specified
also making printing matrices and checking module sanity work for complex

test analytic has been transformed to template to allow all
real/complex and single/double variants. However, at this
commit, only real double and real single variants are enabled

This closses issue #40.