GitLab

The Intel Fortran compiler accepts the flag "-fopenmp" for compilation
with OpenMP. However, the Intel MPI compiler wrapper does not.
With the Intel compiler, this leads to the fact, that if ELPA is compiled
with the "-fopenmp" flag a not thread-save version of the Intel MPI
library is used and the test (with make check) fails.

Intel promised to solve this in the future.

However, for now the problem is solved in the user friendly way that no
manipulation of the MPI compiler wrappers have to be done:

For detecting the OpenMP compiler flags, instead of the predefined
macro "AC_OPENMP" of autoconf a modified macro "AX_ELPA_OPENMP"
is used, which first checks "-openmp" and only then "-fopenmp".
Thus it is ensured that the Intel compiler (and mpi compiler wrapper)
does not get confused.

This is invisible for users calling "configure" during the installation
process.

If the ELPA library is compiled with OpenMP, the tests check
whether the MPI library provides the neccessary threading level.

There has been the error, that if the required threading level
was not available the test programs aborted, put no explicit
error code was set. This is changed now.

It is more user friendly if the configure steps aborts in
the case that the user did not define which ELPA kernel to use

It is more user friendly if the configure steps aborts in the case
that the user did not define which ELPA kernel to use

A few kernels could not be used together with OpenMP:
an interface error prohibited the compilation.

Removed a forgotten C-preprocessor error flag.
Futhermore, describe in the documentation why at the moment a
build of Bluegene P/Q kernels and OpenMP is not allowed.

Removed a forgotten C-preprocessor error flag.
Futhermore, describe in the documentation why at the moment a
build of Bluegene P/Q kernels and OpenMP is not allowed.

Automake does normally not understand the dependencies of Fortran 90
modules, thus a parallel build will almost certainly fail.

As in ELPA 2011.12.002, the ELPA_development_version_qr  now
includes scripts from the fdep project (see git.schokokeks.org/fdep)
which cirumvents this problem.

Thanks to Lorenz Huedepohl, now a parallel build works.

Automake does normally not understand the dependencies of Fortran 90
modules, thus a parallel build will almost certainly fail.

As in ELPA 2011.12.002, ELPA 2013.08.004  now includes scripts from
the fdep project (see git.schokokeks.org/fdep) which cirumvents this
problem.

Thanks to Lorenz Huedepohl, now a parallel build works.

Automake does normally not understand the dependencies of Fortran 90
modules, thus a parallel build will almost certainly fail.

ELPA 2011.12.002 now includes scripts from the fdep project
(see git.schokokeks.org/fdep) which cirumvents this problem.

Thanks to Lorenz Huedepohl, now a parallel build works.

Identical to ELPA 2013.11.001 the build procedure is updated

  - updated INSTALL documentation
  - improved description of kernels src/elpa2_kernels/README_elpa2_kernels.txt
  - improved OpenMP builds:
      * usage of OpenMP can now be specified with
        "--enable-openmp" or "--with-openmp"
  - improved checks for AVX build (identical to ELPA 2013.08.003):
     * if AVX kernels are specified AVX compilation is tested
     * the option "--with-avx-optimization" automatically updates
       CFLAGS and CXXFLAGS with optimization flags

The Fortran standard requires a maximum line length of 132 characters.
Linefeeds are introduced where necessary

- updated INSTALL documentation
- improved description of kernels src/elpa2_kernels/README_elpa2_kernels.txt
- improved OpenMP builds:
   * usage of OpenMP can now be specified with
     "--enable-openmp" or "--with-openmp"
- improved checks for AVX build (identical to ELPA 2013.08.003):
   * if AVX kernels are specified AVX compilation is tested
   * the option "--with-avx-optimization" automatically updates
     CFLAGS and CXXFLAGS with optimization flags

 - The INSTALL documentation was updated a bit
 - the documentation of the ELPA kernels was improved
 - the configure script was improved:

   if usage of an AVX kernel is specified, at configure time
   it is checked whether such a kernel can be build
   If not, it is checked whether the kernel can be build if the
   option "-mavx" is added to the CFLAGS and CXXFLAGS. If this is
   still not possible an error is thrown.

   if the option "--with-avx-optimization" than the CFLAGS and CXXFLAGS
   are automatically updated with some necessary flags (which are
   described in the kernel documentation file)

Now, the branch ELPA_development_version_qr contains the full QR
functionality, which was introduced by Auckenthaler in the (obsolete)
ELPA_development_version_branch

Changes are:
- build procedure with configure like the other branches/releases
- renaming of a few files

The QR decomposition was tested for a few matrices and gave correct
results

- the files in the sub-dir src/elpa_qr from the ELPA_development_branch
  are introduced in the branch ELPA_development_version_qr

- the file tum_utils.f90 is renamed to qr_utils.f90

This is the release of the ELPA_development_version_OpenMP
If OpenMP support is not used, this version has the same
functionality as ELPA_2013.08.

If OpenMP support is used, obviously, a hybrid version of ELPA will be
build.

Allthough this is a release, version ELPA_2013.11 is far from complete!
During the next week optimizations of the OpenMP part will be published,
however, the basic functionality is set by this commit

If ELPA is compiled with OpenMP the MPI_INIT_THREAD with
level THREAD_MULIPLE must be called. This is changed in the test
programs

- install (if build) test binaries in installdir/bin
- finish the test examples with a clean blacs_grid_exit

ELPA 2013.08.002

- install (if build) test binaries in installdir/bin
- finish the test examples with a clean blacs_grid_exit

ELPA_development_version_OPENMP
- install (if build) test binaries in installdir/bin
- rename files from *.f90 to *.F90 (later needed)

A developement version for MRRR algorithm is created;
Based on ELPA_development_version_OpenMP

ELPA_development_version_branch

Since the new, tested branch ELPA_development_version_OpenMP is
now in place, the still buggy OpenMP functionality from the branch
ELPA_development_version is removed.

The branch ELPA_development_version now only contains the untested
support of the MRRR algorithm

The test examples are a little bit more verbose if OpenMP is
used

This commit introduces OpenMP functionality in the
ELPA_development_version_OpenMP branch.

It contains several bugfixes to the OpenMP functionality in the
branch "ELPA_development_version", the later will soon be deleted
since the new branch is the new reference implementation.

The current branch contains the following features/bugfixes:
- building of the OpenMP version of ELPA via configure and the
  "--with-openmp" flag. The build library contains a "_mt"
  (multi-threaded) in its name.
  The configure procedure should (hopefully) determine for each
  compiler the neccessary OpenMP flags.
  If the "--with-openmp" flag is ommitted exactly the same code
  as in the ELPA 2013.08.001 release is used and build in the
  same way
- The example test cases print which kernels have been used and
  how many OpenMP threads are used at runtime
- correct handling of OpenMP stack arrays: the previous implementation
  caused compiler dependent segmentation faults
- OpenMP capability with all available kernels: the correctness of
  the computations have been checked for all kernels except the
  Bluegene (P/Q) versions

Based on the ELPA 2013.08.001 a development version for ELPA OpenMP
has been created

For debugging purposes the test examples of the ELPA library can
now write the Eigenvectors in an ascii file, e.g
a call of "test_real 4000 1500 16 output" will envoke the test
case for real matrices of the ELPA1 library. A matrix of 4000x4000
values will be created an 1500 EV's will be computet with a 16 blocking.
And the results will be --- due to the output flag --- written to a
text file

A bug in ELPA_2013.08 was fixed: due to typo in a preprocessor statement
the kernels optimized for BLUGENE/P could not be used correctly.

In course of this action, also a bugfix versioning scheme was
introduced: in case of bugfixes they are versioned with an extra 3 digit
number in the ELPA release name, i.e ELPA 2013.08 -> 2013.08.001

Consistently any new release of ELPA will from now on start on the minor
level "000".

Due to a wrong preprocessor setting, dependig on the chosen elpa2
kernel, no kernel was called from within elpa2.F90, i. e. the
real computation actually did nothing. The complex case was not
affected.

It is pointed out that setting the rpath in the linkline for
the blacs/scalapack installation is useful

The configure process in ELPA_2013.08 did not work, since
I forgot to upload some files

It is pointed out that it is mandatory to set the variables
"BLACS_LDFLAGS" and "BLACS_FCFLAGS" at build time.

Furthermore, it is clarified how to set individual optimization
flags at build time

Thanks to a comment to the Debian community, the versioning
of the ELPA library and its API are changed:

- the API version nummer starts now with 0.0.0 and not
  anymore with the version number of the first ELPA release,
  i.e. 2011.12

- if ELPA is installed as library, the files are now named
  elpa.so.API_VERSION. All references to the ELPA version
  and the used compiler vanished, since these are considered
  harmeful.

Note, that at a later stage, if neccessary, the library name might
also contain the uses MPI version.

This commit releases ELPA version 2013.08

It is identical with the latest commit version 2013.02_BETA, except:

- rename ELPA_2013.02.BETA -> ELPA_2013.08
- update of configure.ac
- remove of src/elpa2.f90_save
- remove of test/Makefile

This version has been tested extensivly, however, there might always
be some bugs.
In case of questions please contact elpa-library@rzg.mpg.de or via
the ELPA forum on http://elpa-lib.fhi-berlin.mpg.de/forum/index.php

This is the version as obtained from Peter Messmer from Nvidia.
This version is not yet consolidated or very well tested.
The build process is a little tricky.

But next commits will start cleaning up

Similiar to ELPA_2011.12 this version of ELPA now supports automake.

To the _source_ code of ELPA the following changes had to be made:

-elpa2.F90 now use the create "config-f90.h" file to choose via
 preprocessor switches which kernel was defined by the user

- the SSE-AVX intrinsic kernels of A. Heinecke had to be changed such
  that they support "internal linkage" for the building of a shared
  library.

The following other changes were made to the documentation

- elpa2_kernels/README_elpa2_kernels.txt was updated; the new configure
  targets are explained

- README, INSTALL were updated

To high propability this commit is the last one, before this beta
version	will become an official release