- 02 Feb, 2015 2 commits
-
-
Andreas Marek authored
As obvious from the previous commits, this release of ELPA introduces a (optional) QR-decomposition for real valued matrices. This option can be used at run-time by either setting an environment variable, or by calling the ELPA-2 solver for real matrices with an additional flag. Thus the ABI changed, w.r.t. previous versions. Furthermore, the build process of ELPA has been made more consistent. All optimization flags (especially O1, O2 etc.) have to be set at build time by the user via the CFLAGS, FCFLAGS, and CXXFLAGS variables. The configure script does not set automatically the "O-Flags" anymore.
-
Andreas Marek authored
- cleanup of the file - add more (optional) timing information
-
- 29 Jan, 2015 1 commit
-
-
Andreas Marek authored
The qr decomposition is now available as a runtime choice. Some testing has still to be done
-
- 28 Jan, 2015 2 commits
-
-
Andreas Marek authored
-
Andreas Marek authored
-
- 27 Jan, 2015 1 commit
-
-
Lorenz Huedepohl authored
-
- 25 Aug, 2014 2 commits
-
-
Andreas Marek authored
At build time it can be specified that the ELPA test programs give more detailed timing information, which is usefull for performace measurements
-
Andreas Marek authored
If specified in the configure step, the test programs redirect their stdout and stderr output of each MPI task in a seperate file, which will be stored in a subdirectory "mpi_stdout". This will only be done if the environment variable "REDIRECT_ELPA_TEST_OUTPUT" is set to "true"
-
- 08 Aug, 2014 1 commit
-
-
Andreas Marek authored
These programs have not been updated correctly
-
- 21 Jul, 2014 1 commit
-
-
Lorenz Huedepohl authored
Train fdep to recognize module procedure ... lines. Also, previoulsy an fdep internal error aborted make, however the partially generated dependency file stayed and was used in that form on a subsequent make call.
-
- 03 Jul, 2014 1 commit
-
-
Andreas Marek authored
-
- 02 Jul, 2014 5 commits
-
-
Lorenz Huedepohl authored
It was decided that the version number shall be all of 2014.06.001, that is including the release version. Sadly, pkg-config is not too flexible with different versions of libraries, therefore the pkg-config file will have the full version number including the release. This might be a bit cumbersome for dependent projects.
-
Andreas Marek authored
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
It was decided that the library file shall be called libelpa.so.* without an explicit version string. Version information will be encoded by using the -version-info parameter of libtool, resulting in an actual filename as defined by your flavor of Unix. The file LIBRARY_INTERFACE documents the changes, the current interface has number 1 and is binary incompatible with the previous version 0.
-
Andreas Marek authored
-
- 01 Jul, 2014 7 commits
-
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
Conflicts happened only on the print statements with the supported MPI thread level error message
-
Lorenz Huedepohl authored
Small helper module for better error diagnostics
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
_mt implies thread-safety, which is not what the OpenMP version does. Additionally, only install a subset of the test-programs
-
Lorenz Huedepohl authored
-
- 30 Jun, 2014 2 commits
-
-
Lorenz Huedepohl authored
Multiple rules for the same dummy-file were created for modules used by sources of more than one program. Corrected by including the program name into the dummy-files
-
Andreas Marek authored
Build servers recommend against calling "abort" functions in a library. Thus, from now on, ELPA returns a success flag, which is a change of ABI! In order to force the user to honour this, the ELPA drivers are converted to functions, which implies an adaptation and rebuild of the user code. The driver functions return a logical "success", which, if it is false, indicates an error within the library. It is advised to check catch and check this return value in the user code, which calls ELPA.
-
- 27 Jun, 2014 7 commits
-
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
The value of $TEST_FLAGS at the time make check was called was hard-coded in test*.sh files due to wrong quoting.
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
By passing the makefile variable TEST_FLAGS, one can adjust the computional cost of the test, e.g. reduce it considerably via make check TEST_FLAGS="1500 50 16"
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
-
- 25 Jun, 2014 6 commits
-
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
These are saved in the .log files by the test-driver anyway and do not clutter the screen output for the user. But having them can be useful to find out how some test failed.
-
Andreas Marek authored
- the ascii files "INSTALL", "README", and "RELEASE_NOTES" are updated - the Makefile.example in the test directory is updated
-
Lorenz Huedepohl authored
-
Lorenz Huedepohl authored
The fdep magic only works within a single Automake target, add the remaining dependencies by hand. Parallel compilation should now work also for "make all" again.
-
Lorenz Huedepohl authored
-
- 23 Jun, 2014 1 commit
-
-
Andreas Marek authored
Since the tar-balls are availabel on the ELPA website , http://elpa.rzg.mpg.de/ , this directory is not useful anymore and just wastes disk space
-
- 18 Jun, 2014 1 commit
-
-
Andreas Marek authored
-