Merge remote-tracking branch 'origin/master_pre_stage' into pkus/generalized

generalized tests disabled (files taken from master_pre_stage due to
many conflicts), have to be re-introduced

.ci-env-vars

-if [ "$HOST" = "buildtest-rzg" ]; then module load impi/5.1.3 intel/16.0 gcc/6.3 mkl/11.3 autotools pkg-config; fi
-if [ "$HOST" = "knl1" -o "$HOST" = "knl2" -o "$HOST" = "knl3" -o "$HOST" = "knl4" ] ; then module load impi/5.1.3 intel/17.0 gcc/6.3 mkl/2017 pkg-config; fi
-if [ "$HOST" = "gp02" ] ; then module load impi/2017.2 intel/17.0 gcc/6.3 mkl/2017 pkg-config; fi
-if [ "$HOST" != "hydra03" -a "$HOST" != "hydra04" -a "$HOST" != "hydra05"  -a "$HOST" != "hydra06"  -a "$HOST" != "hydra07"  -a "$HOST" != "knl1" -a "$HOST" != "knl2" -a "$HOST" != "knl3" -a "$HOST" != "knl4" -a "$HOST" != "gp02" -a "$HOST" != "draco01" -a "$HOST" != "draco02" -a "$HOST" != "draco03" -a "$HOST" != "draco04"  ] ; then module load impi/5.1.3 intel/16.0 gcc/6.4 mkl/11.3 autotools pkg-config; fi
-if [ "$HOST" = "draco01" -o "$HOST" = "draco02" -o "$HOST" = "draco03" -o "$HOST" = "draco04" ] ; then module load impi/2017.3 intel/17.0 gcc/6.3 mkl/2017 pkg-config; fi
-module list
-if [ "$HOST" != "buildtest-rzg" ]; then  pushd $HOME/bin/reserve_timeslot && git pull && popd ; fi
+if [ "$(hostname)" == "buildtest-rzg" ]; then module load impi/5.1.3 intel/16.0 gcc/6.3 mkl/11.3 autotools pkg-config; fi
+
+if [ "$(hostname)" == "knl1" -o "$(hostname)" == "knl2" -o "$(hostname)" == "knl3" -o "$(hostname)" == "knl4" ] ; then module load impi/2017.4 intel/17.0 gcc/7.2 mkl/2017 pkg-config; fi
+if [ "$(hostname)" == "maik" ]; then module load impi/5.1.3 intel intel/17.0 gcc/6.3 mkl/2017 pkg-config/0.29.1; fi
+
+if [ "$(hostname)" == "gp02" ] ; then module load impi/5.1.3 intel/17.0 gcc/6.3 mkl/2017 pkg-config; fi
+
+if [ "$(hostname)" == "amarek-elpa-gitlab-runner-1" ]; then module load intel/16.0 gcc mkl/11.3 autoconf automake libtool impi/5.1.3; fi
+if [ "$(hostname)" == "amarek-elpa-gitlab-runner-2" ]; then module load intel/16.0 gcc mkl/11.3 autoconf automake libtool impi/5.1.3; fi
+if [ "$(hostname)" == "amarek-elpa-gitlab-runner-3" ]; then module load intel/16.0 gcc mkl/11.3 autoconf automake libtool impi/5.1.3; fi
+if [ "$(hostname)" == "amarek-elpa-gitlab-runner-4" ]; then module load intel/16.0 gcc mkl/11.3 autoconf automake libtool impi/5.1.3; fi
+
+if [ "$(hostname)" == "dvl01" ]; then module load intel/17.0 gcc/5.4 mkl/2017 impi/2017.2 gcc/5.4 cuda/8.0; fi
+if [ "$(hostname)" == "dvl02" ]; then module load intel/17.0 gcc/5.4 mkl/2017 impi/2017.2 gcc/5.4 cuda/8.0; fi
+
+if [ "$(hostname)" == "miy01" ]; then module purge && module load gcc/5.4 smpi essl/5.5 cuda pgi/17.9 && export LD_LIBRARY_PATH=/opt/ibm/spectrum_mpi/lib:/opt/ibm/spectrum_mpi/profilesupport/lib:$LD_LIBRARY_PATH && export PATH=/opt/ibm/spectrum_mpi/bin:$PATH && export OMPI_CC=gcc && export OMPI_FC=gfortran; fi
+if [ "$(hostname)" == "miy02" ]; then module load gcc/5.4 pgi/17.9 ompi/pgi/17.9/1.10.2 essl/5.5 cuda && export LD_LIBRARY_PATH=/opt/ibm/spectrum_mpi/lib:/opt/ibm/spectrum_mpi/profilesupport/lib:$LD_LIBRARY_PATH && export PATH=/opt/ibm/spectrum_mpi/bin:$PATH; fi
+if [ "$(hostname)" == "miy03" ]; then module load gcc/5.4 pgi/17.9 ompi/pgi/17.9/1.10.2 essl/5.5 cuda && export LD_LIBRARY_PATH=/opt/ibm/spectrum_mpi/lib:/opt/ibm/spectrum_mpi/profilesupport/lib:$LD_LIBRARY_PATH && export PATH=/opt/ibm/spectrum_mpi/bin:$PATH; fi
 
+
+if [ "$(hostname)" != "miy01" -a "$(hostname)" != "miy02" -a "$(hostname)" != "miy03" ]; then
 export MKL_INTEL_SCALAPACK_MPI_NO_OMP_BASELINE="-L$MKL_HOME/lib/intel64 -lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lmkl_blacs_intelmpi_lp64 -lpthread"
 export MKL_INTEL_SCALAPACK_FCFLAGS_MPI_NO_OMP="-I$MKL_HOME/include/intel64/lp64"
 export MKL_INTEL_SCALAPACK_LDFLAGS_MPI_NO_OMP="$MKL_INTEL_SCALAPACK_MPI_NO_OMP_BASELINE -Wl,-rpath,$MKL_HOME/lib/intel64"
@@ -40,3 +53,9 @@ export MKL_GFORTRAN_SCALAPACK_LDFLAGS_NO_MPI_OMP="$MKL_GFORTRAN_SCALAPACK_NO_MPI
 
 export ASAN_OPTIONS=suppressions=no_asan_for_mpi.supp,fast_unwind_on_malloc=0
 export LSAN_OPTIONS=suppressions=no_lsan_for_mpi.supp
+
+fi
+
+if [ "$(hostname)" != "knl1" -a "$(hostname)" != "knl2" -a "$(hostname)" != "knl3" -a "$(hostname)" != "knl4" -a  "$(hostname)" != "maik" ] ; then export I_MPI_DEBUG=5; fi
+
+module list

Changelog

+Changelog for ELPA 2017.11.001
+
+- added new compute kernels for IBM Power8 and Fujistu Sparc64
+  processors
+- a first implementation of autotuning capability
+- correct some type statements in Fortran
+- correct detection of PAPI in configure step
+
+Changelog for ELPA 2017.05.003
+
+- remove bug in invert_triangular, which had been introduced
+  in ELPA 2017.05.002
+
 Changelog for ELPA 2017.05.002
 
 Mainly bugfixes for ELPA 2017.05.001:

DEPRECATED_FEATURES.md

@@ -17,12 +17,12 @@ following listed interfaces will be removed at some time.
 In order to unfiy the namespace of the *ELPA* public interfaces, several interfaces
 have been replaced by new names. The old interfaces will be removed
 
-Deprecated interface             Replacement
-===================================================
-get_elpa_row_col_coms            elpa_get_communicators
-get_elpa_communicators           elpa_get_communicators
-solve_evp_real                   elpa_solve_evp_real_1stage_double
-solve_evp_complex                elpa_solve_evp_complex_1stage_double
+Deprecated interface             Replacement                            Comment
+==================================================================================================
+get_elpa_row_col_coms            elpa_get_communicators                (removed since 2017.11.001)
+get_elpa_communicators           elpa_get_communicators                (removed since 2017.11.001)
+solve_evp_real                   elpa_solve_evp_real_1stage_double     (removed since 2017.11.001)
+solve_evp_complex                elpa_solve_evp_complex_1stage_double  (removed since 2017.11.001)
 solve_evp_real_1stage            elpa_solve_evp_real_1stage_double
 solve_evp_complex_1stage         elpa_solve_evp_complex_1stage_double
 solve_evp_real_2stage            elpa_solve_evp_real_2stage_double

INSTALL.md

 # Installation guide for the *ELPA* library#
 
-## Preamle ##
+## Preamble ##
 
-This file provides documentation on how to build the *ELPA* library in **version ELPA-2017.05.002**.
+This file provides documentation on how to build the *ELPA* library in **version ELPA-2017.11.001**.
 With release of **version ELPA-2017.05.001** the build process has been significantly simplified,
 which makes it easier to install the *ELPA* library
 
@@ -28,7 +28,7 @@ for the documentation how to proceed.
 *ELPA* can be installed with the build steps
 - configure
 - make
-- make check
+- make check   | or make check CHECK_LEVEL=extended
 - make install
 
 Please look at configure --help for all available options.

ISSUES.md

@@ -3,6 +3,12 @@
 For more details and recent updates please visit the online [issue system] (https://gitlab.mpcdf.mpg.de/elpa/elpa/issues)
 Issues which are not mentioned in a newer release are (considered as) solved
 
+### ELPA 2017.11.001 release ###
+- the elpa autotune print functions cannot print at the moment
+
+### ELPA 2017.05.003 release ###
+- at the moment no issues are known
+
 ### ELPA 2017.05.002 release ###
 - at the moment no issues are known
 

LIBRARY_INTERFACE

@@ -73,3 +73,9 @@ https://www.gnu.org/software/libtool/manual/html_node/Updating-version-info.html
   NO incompatible API changes w.r.t. to the previous version. Only some functions have
   been added. The state of release 2017.05.001 defines this interface
 
+- 11
+  Incompatible API changes w.r.t. to the previous version (only in the so called
+  "legacy interface", since as anounced some deprecated function aliases have been
+  removed). For the current interface all changes since 2017.05.001 are
+  compatible, since only some functions have been added. 
+  The state of release 2017.11.001.(rc1) defines this interface

Makefile.am

@@ -35,6 +35,7 @@ noinst_LTLIBRARIES += libelpa@SUFFIX@_private.la
 libelpa@SUFFIX@_private_la_FCFLAGS = $(AM_FCFLAGS) $(FC_MODOUT)private_modules $(FC_MODINC)private_modules
 libelpa@SUFFIX@_private_la_SOURCES = \
   src/elpa_impl.F90 \
+  src/elpa_autotune_impl.F90 \
   src/elpa_abstract_impl.F90 \
   src/helpers/mod_precision.F90 \
   src/helpers/mod_mpi.F90 \
@@ -203,6 +204,20 @@ endif
 endif
 endif
 
+if WITH_REAL_SPARC64_BLOCK2_KERNEL
+  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_sparc64_2hv_double_precision.c
+#if WANT_SINGLE_PRECISION_REAL
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_sparc64_2hv_single_precision.c
+#endif
+endif
+
+if WITH_REAL_VSX_BLOCK2_KERNEL
+  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_vsx_2hv_double_precision.c
+if WANT_SINGLE_PRECISION_REAL
+  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_vsx_2hv_single_precision.c
+endif
+endif
+
 if WITH_REAL_SSE_BLOCK2_KERNEL
   libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_sse_2hv_double_precision.c
 if WANT_SINGLE_PRECISION_REAL
@@ -231,6 +246,19 @@ if WANT_SINGLE_PRECISION_REAL
 endif
 endif
 
+if WITH_REAL_SPARC64_BLOCK4_KERNEL
+  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_sparc64_4hv_double_precision.c
+#if WANT_SINGLE_PRECISION_REAL
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_sparc64_4hv_single_precision.c
+#endif
+endif
+
+if WITH_REAL_VSX_BLOCK4_KERNEL
+  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_vsx_4hv_double_precision.c
+if WANT_SINGLE_PRECISION_REAL
+  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_vsx_4hv_single_precision.c
+endif
+endif
 
 if WITH_REAL_SSE_BLOCK4_KERNEL
   libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_sse_4hv_double_precision.c
@@ -260,7 +288,19 @@ if WANT_SINGLE_PRECISION_REAL
 endif
 endif
 
+if WITH_REAL_SPARC64_BLOCK6_KERNEL
+  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_sparc64_6hv_double_precision.c
+#if WANT_SINGLE_PRECISION_REAL
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_sparc64_6hv_single_precision.c
+#endif
+endif
 
+if WITH_REAL_VSX_BLOCK6_KERNEL
+  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_vsx_6hv_double_precision.c
+if WANT_SINGLE_PRECISION_REAL
+  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_vsx_6hv_single_precision.c
+endif
+endif
 
 if WITH_REAL_SSE_BLOCK6_KERNEL
   libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/real_sse_6hv_double_precision.c
@@ -290,6 +330,19 @@ if WANT_SINGLE_PRECISION_REAL
 endif
 endif
 
+#if WITH_COMPLEX_SPARC64_BLOCK1_KERNEL
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_sparc64_1hv_double_precision.c
+#if WANT_SINGLE_PRECISION_COMPLEX
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_sparc64_1hv_single_precision.c
+#endif
+#endif
+#
+#if WITH_COMPLEX_VSX_BLOCK1_KERNEL
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_vsx_1hv_double_precision.c
+#if WANT_SINGLE_PRECISION_COMPLEX
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_vsx_1hv_single_precision.c
+#endif
+#endif
 
 if WITH_COMPLEX_SSE_BLOCK1_KERNEL
   libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_sse_1hv_double_precision.c
@@ -320,6 +373,20 @@ if WANT_SINGLE_PRECISION_COMPLEX
 endif
 endif
 
+#if WITH_COMPLEX_SPARC64_BLOCK2_KERNEL
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_sparc64_2hv_double_precision.c
+#if WANT_SINGLE_PRECISION_COMPLEX
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_sparc64_2hv_single_precision.c
+#endif
+#endif
+#
+#if WITH_COMPLEX_VSX_BLOCK2_KERNEL
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_vsx_2hv_double_precision.c
+#if WANT_SINGLE_PRECISION_COMPLEX
+#  libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_vsx_2hv_single_precision.c
+#endif
+#endif
+
 if WITH_COMPLEX_SSE_BLOCK2_KERNEL
   libelpa@SUFFIX@_private_la_SOURCES += src/elpa2/kernels/complex_sse_2hv_double_precision.c
 if WANT_SINGLE_PRECISION_COMPLEX
@@ -390,9 +457,7 @@ dist_man_MANS = \
 
 if ENABLE_LEGACY
 dist_man_MANS += \
-  man/solve_evp_real.3 \
   man/solve_evp_real_1stage_double.3 \
-  man/solve_evp_complex.3 \
   man/solve_evp_complex_1stage_double.3 \
   man/solve_evp_real_2stage_double.3 \
   man/solve_evp_complex_2stage_double.3 \
@@ -400,8 +465,7 @@ dist_man_MANS += \
   man/elpa_solve_evp_complex_1stage_double.3 \
   man/elpa_solve_evp_real_2stage_double.3 \
   man/elpa_solve_evp_complex_2stage_double.3 \
-  man/get_elpa_row_col_comms.3 \
-  man/get_elpa_communicators.3 \
+  man/elpa_get_communicators.3 \
   man/elpa_mult_at_b_real_double.3 \
   man/elpa_mult_at_b_real_single.3 \
   man/elpa_mult_ah_b_complex_double.3 \
@@ -413,7 +477,11 @@ dist_man_MANS += \
   man/elpa_solve_evp_real_double.3 \
   man/elpa_solve_evp_real_single.3 \
   man/elpa_solve_evp_complex_double.3 \
-  man/elpa_solve_evp_complex_single.3
+  man/elpa_solve_evp_complex_single.3 \
+  man/elpa_autotune_setup.3 \
+  man/elpa_autotune_step.3 \
+  man/elpa_autotune_set_best.3 \
+  man/elpa_autotune_deallocate.3
 endif
 
 
@@ -489,26 +557,26 @@ include legacy_test_programs.am
 endif
 
 noinst_PROGRAMS += double_instance@SUFFIX@
-check_SCRIPTS += double_instance@SUFFIX@.sh
+check_SCRIPTS += double_instance@SUFFIX@_default.sh
 double_instance@SUFFIX@_SOURCES = test/Fortran/elpa2/double_instance.F90
 double_instance@SUFFIX@_LDADD = $(test_program_ldadd)
 double_instance@SUFFIX@_FCFLAGS = $(AM_FCFLAGS) $(FC_MODINC)test_modules $(FC_MODINC)modules
 
 noinst_PROGRAMS += real_2stage_banded@SUFFIX@
-check_SCRIPTS += real_2stage_banded@SUFFIX@.sh
+check_SCRIPTS += real_2stage_banded@SUFFIX@_default.sh
 real_2stage_banded@SUFFIX@_SOURCES = test/Fortran/elpa2/real_2stage_banded.F90
 real_2stage_banded@SUFFIX@_LDADD = $(test_program_ldadd)
 real_2stage_banded@SUFFIX@_FCFLAGS = $(AM_FCFLAGS) $(FC_MODINC)test_modules $(FC_MODINC)modules
 
 noinst_PROGRAMS += complex_2stage_banded@SUFFIX@
-check_SCRIPTS += complex_2stage_banded@SUFFIX@.sh
+check_SCRIPTS += complex_2stage_banded@SUFFIX@_default.sh
 complex_2stage_banded@SUFFIX@_SOURCES = test/Fortran/elpa2/complex_2stage_banded.F90
 complex_2stage_banded@SUFFIX@_LDADD = $(test_program_ldadd)
 complex_2stage_banded@SUFFIX@_FCFLAGS = $(AM_FCFLAGS) $(FC_MODINC)test_modules $(FC_MODINC)modules
 
 if WANT_SINGLE_PRECISION_REAL
 noinst_PROGRAMS += single_real_2stage_banded@SUFFIX@
-check_SCRIPTS += single_real_2stage_banded@SUFFIX@.sh
+check_SCRIPTS += single_real_2stage_banded@SUFFIX@_default.sh
 single_real_2stage_banded@SUFFIX@_SOURCES = test/Fortran/elpa2/single_real_2stage_banded.F90
 single_real_2stage_banded@SUFFIX@_LDADD = $(test_program_ldadd)
 single_real_2stage_banded@SUFFIX@_FCFLAGS = $(AM_FCFLAGS) $(FC_MODINC)test_modules $(FC_MODINC)modules
@@ -516,14 +584,14 @@ endif
 
 if WANT_SINGLE_PRECISION_COMPLEX
 noinst_PROGRAMS += single_complex_2stage_banded@SUFFIX@
-check_SCRIPTS += single_complex_2stage_banded@SUFFIX@.sh
+check_SCRIPTS += single_complex_2stage_banded@SUFFIX@_default.sh
 single_complex_2stage_banded@SUFFIX@_SOURCES = test/Fortran/elpa2/single_complex_2stage_banded.F90
 single_complex_2stage_banded@SUFFIX@_LDADD = $(test_program_ldadd)
 single_complex_2stage_banded@SUFFIX@_FCFLAGS = $(AM_FCFLAGS) $(FC_MODINC)test_modules $(FC_MODINC)modules
 endif
 
 
-# test scriptsT
+# test scripts
 TASKS ?= 2
 if WITH_MPI
   wrapper=$(MPI_BINARY) -n $${TASKS:-$(TASKS)}
@@ -531,7 +599,12 @@ else
   wrapper=
 endif
 TESTS = $(check_SCRIPTS)
-%.sh: %
+%_extended.sh: %
+	@echo "#!/bin/bash" > $@
+	@echo 'if [ "$$CHECK_LEVEL" = "extended" ] ; then $(wrapper) ./$^ $$TEST_FLAGS ; else exit 77; fi' >> $@
+	@chmod +x $@
+
+%_default.sh: %
 	@echo "#!/bin/bash" > $@
 	@echo '$(wrapper)' ./$^ '$$TEST_FLAGS' >> $@
 	@chmod +x $@
@@ -622,6 +695,9 @@ EXTRA_DIST = \
   src/elpa2/kernels/real_avx512_2hv_template.c \
   src/elpa2/kernels/real_avx512_4hv_template.c \
   src/elpa2/kernels/real_avx512_6hv_template.c \
+  src/elpa2/kernels/real_vsx_2hv_template.c \
+  src/elpa2/kernels/real_vsx_4hv_template.c \
+  src/elpa2/kernels/real_vsx_6hv_template.c \
   src/elpa2/kernels/real_sse_2hv_template.c \
   src/elpa2/kernels/real_sse_4hv_template.c \
   src/elpa2/kernels/real_sse_6hv_template.c \
@@ -646,12 +722,12 @@ EXTRA_DIST = \
   test/shared/test_check_correctness_template.F90 \
   test/shared/test_prepare_matrix_template.F90 \
   test/shared/test_analytic_template.F90 \
-  test_project/Makefile.am \
-  test_project/autogen.sh \
-  test_project/configure.ac \
-  test_project/fdep \
-  test_project/m4 \
-  test_project/src/test_real.F90 \
+  test_project_1stage/Makefile.am \
+  test_project_1stage/autogen.sh \
+  test_project_1stage/configure.ac \
+  test_project_1stage/fdep \
+  test_project_1stage/m4 \
+  test_project_1stage/src/test_real.F90 \
   test_project_2stage/Makefile.am \
   test_project_2stage/autogen.sh \
   test_project_2stage/configure.ac \
@@ -664,7 +740,19 @@ EXTRA_DIST = \
   test_project_C/fdep \
   test_project_C/m4 \
   test_project_C/src/test_real.c \
-  test_project_C/src/test_blacs_infrastructure.F90
+  test_project_C/src/test_blacs_infrastructure.F90\
+  test_project_1stage_legacy_api/Makefile.am \
+  test_project_1stage_legacy_api/autogen.sh \
+  test_project_1stage_legacy_api/configure.ac \
+  test_project_1stage_legacy_api/fdep \
+  test_project_1stage_legacy_api/m4 \
+  test_project_1stage_legacy_api/src/test_real.F90 \
+  test_project_2stage_legacy_api/Makefile.am \
+  test_project_2stage_legacy_api/autogen.sh \
+  test_project_2stage_legacy_api/configure.ac \
+  test_project_2stage_legacy_api/fdep \
+  test_project_2stage_legacy_api/m4 \
+  test_project_2stage_legacy_api/src/test_real2.F90
 
 if ENABLE_LEGACY
 EXTRA_DIST += \

README.md

@@ -2,10 +2,19 @@
 
 ## Current Release ##
 
-The current release is ELPA 2017.05.002. The current supported API version
-is 20170403. This release supports the earliest API version 20170403.
+The current release is ELPA 2017.11.001 The current supported API version
+is 20171201. This release supports the earliest API version 20170403.
 
-## About *ELPA*
+[![Build 
+status](https://gitlab.mpcdf.mpg.de/elpa/elpa/badges/master/build.svg)](https://gitlab.mpcdf.mpg.de/elpa/elpa/commits/master)
+
+[![Code 
+coverage](https://gitlab.mpcdf.mpg.de/elpa/badges/master/coverage.svg)](http://elpa.pages.mpcdf.de/elpa/coverage_summary
+
+[![License: LGPL v3][license-badge]](LICENSE)
+[license-badge]: https://img.shields.io/badge/License-LGPL%20v3-blue.svg
+
+## About *ELPA* ##
 
 The computation of selected or all eigenvalues and eigenvectors of a symmetric
 (Hermitian) matrix has high relevance for various scientific disciplines.

RELEASE_NOTES

-This file contains the release notes for the ELPA 2017.05.002 version
+This file contains the release notes for the ELPA 2017.11.001 version
 
 What is new?
 -------------
-For detailed information about changes since release ELPA 2016.11.001 please have a look at the Changelog file
-Highlights are:
+For detailed information about changes since release ELPA 2017.05.03 please have a look at the Changelog file
 
-- a more generic and more flexible API, which allows easy implementation of upcoming features
-- faster GPU implementation, especially for ELPA 1stage
-- the restriction of the block-cyclic distribution blocksize = 128 in the GPU
-  case is relaxed
-- Faster CPU implementation due to better blocking
-- support of already banded matrices (new API only!)
-- improved KNL support
+- This release contains new compute kernels for IBM Power 8 and Fujistu Sparc64 processors
+- A first version of autotuning has been implemented
 
 
 ABI change
 ---------------------
 
-Since release 2016.05.001 the ABI has not changed. Only additional interfaces habe been added, please have also a look
-at the DEPRECTATED_FEATURES file
-
+Since release 2017.05.003 the ABI has changed.
 
 Any incompatibilities to previous version?
 ---------------------------------------
 
-As mentioned before, the ABI of ELPA was not changed; There is no
-incompatibility with the previous version ELPA 2016.11.001.
+As mentioned before, the ABI of ELPA was changed; 
+This change only affectes the  old "lecgacy interface", since as 
+anounced some deprecated function aliases have been removed.
+For the current interface there is no incompatibility with the
+previous version ELPA 2017.05.003.
 

SWITCHING_TO_NEW_INTERFACE.md

@@ -124,8 +124,6 @@ the user application):
 
    "debug"    can be one of { 0 | 1 }, will give more information case of an error if set to 1
 
-   "eigenvalues_only" can be one of { 0 | 1 }, if set to 1 solve will only compute the eigenvalues
-
 
 6. do the desired task with the *ELPA* library, which could be
    a) e%eigenvectors                  ! solve EV problem with solver as set by "set" method; computes eigenvalues AND eigenvectors
@@ -151,7 +149,7 @@ Local documentation (via man pages) should be available (if *ELPA* has been inst
 For example "man elpa2_print_kernels" should provide the documentation for the *ELPA* program which prints all
 the available kernels.
 
-Also a [online doxygen documentation] (http://elpa.mpcdf.mpg.de/html/Documentation/ELPA-2017.05.002/html/index.html)
+Also a [online doxygen documentation] (http://elpa.mpcdf.mpg.de/html/Documentation/ELPA-2017.11.001/html/index.html)
 for each *ELPA* release is available.
 
 

USERS_GUIDE.md

@@ -9,7 +9,7 @@ Local documentation (via man pages) should be available (if *ELPA* has been inst
 For example "man elpa2_print_kernels" should provide the documentation for the *ELPA* program which prints all
 the available kernels.
 
-Also a [online doxygen documentation] (http://elpa.mpcdf.mpg.de/html/Documentation/ELPA-2017.05.002/html/index.html)
+Also a [online doxygen documentation] (http://elpa.mpcdf.mpg.de/html/Documentation/ELPA-2017.11.001/html/index.html)
 for each *ELPA* release is available.
 
 
@@ -125,7 +125,7 @@ of a simple example program can be found in ./test_project/src.
 
 
    ! All ELPA routines need MPI communicators for communicating within
-   ! rows or columns of processes, these are set in get_elpa_communicators
+   ! rows or columns of processes, these are set in elpa_get_communicators
 
    success = elpa_get_communicators(mpi_comm_world, my_prow, my_pcol, &
                                     mpi_comm_rows, mpi_comm_cols)
@@ -216,8 +216,8 @@ SYNOPSIS
        integer, intent(in)    ldq:           leading dimension of matrix q which stores the eigenvectors
        integer, intent(in)    nblk:          blocksize of block cyclic distributin, must be the same in both directions
        integer, intent(in)    matrixCols:    number of columns of locally distributed matrices a and q
-       integer, intent(in)    mpi_comm_rows: communicator for communication in rows. Constructed with get_elpa_communicators(3)
-       integer, intent(in)    mpi_comm_cols: communicator for communication in colums. Constructed with get_elpa_communicators(3)
+       integer, intent(in)    mpi_comm_rows: communicator for communication in rows. Constructed with elpa_get_communicators(3)
+       integer, intent(in)    mpi_comm_cols: communicator for communication in colums. Constructed with elpa_get_communicators(3)
 
        logical                success:       return value indicating success or failure
 
@@ -238,14 +238,14 @@ SYNOPSIS
        int     ldq:           leading dimension of matrix q which stores the eigenvectors
        int     nblk:          blocksize of block cyclic distributin, must be the same in both directions
        int     matrixCols:    number of columns of locally distributed matrices a and q
-       int     mpi_comm_rows: communicator for communication in rows. Constructed with get_elpa_communicators(3)
-       int     mpi_comm_cols: communicator for communication in colums. Constructed with get_elpa_communicators(3)
+       int     mpi_comm_rows: communicator for communication in rows. Constructed with elpa_get_communicators(3)
+       int     mpi_comm_cols: communicator for communication in colums. Constructed with elpa_get_communicators(3)
 
        int     success:       return value indicating success (1) or failure (0)
 
 DESCRIPTION
        Solve the real eigenvalue problem with the 1-stage solver. The ELPA communicators mpi_comm_rows and mpi_comm_cols are obtained with the
-       get_elpa_communicators(3) function. The distributed quadratic marix a has global dimensions na x na, and a local size lda x matrixCols.
+       elpa_get_communicators(3) function. The distributed quadratic marix a has global dimensions na x na, and a local size lda x matrixCols.
        The solver will compute the first nev eigenvalues, which will be stored on exit in ev. The eigenvectors corresponding to the eigenvalues
        will be stored in q. All memory of the arguments must be allocated outside the call to the solver.
 
@@ -265,8 +265,8 @@ DESCRIPTION
        integer,     intent(in)    ldq:           leading dimension of matrix q which stores the eigenvectors
        integer,     intent(in)    nblk:          blocksize of block cyclic distributin, must be the same in both directions
        integer,     intent(in)    matrixCols:    number of columns of locally distributed matrices a and q
-       integer,     intent(in)    mpi_comm_rows: communicator for communication in rows. Constructed with get_elpa_communicators(3)
-       integer, intent(in)        mpi_comm_cols: communicator for communication in colums. Constructed with get_elpa_communicators(3)
+       integer,     intent(in)    mpi_comm_rows: communicator for communication in rows. Constructed with elpa_get_communicators(3)
+       integer, intent(in)        mpi_comm_cols: communicator for communication in colums. Constructed with elpa_get_communicators(3)
 
        logical                    success:       return value indicating success or failure
 
@@ -288,14 +288,14 @@ DESCRIPTION
        int             ldq:           leading dimension of matrix q which stores the eigenvectors
        int             nblk:          blocksize of block cyclic distributin, must be the same in both directions
        int             matrixCols:    number of columns of locally distributed matrices a and q
-       int             mpi_comm_rows: communicator for communication in rows. Constructed with get_elpa_communicators(3)
-       int             mpi_comm_cols: communicator for communication in colums. Constructed with get_elpa_communicators(3)
+       int             mpi_comm_rows: communicator for communication in rows. Constructed with elpa_get_communicators(3)
+       int             mpi_comm_cols: communicator for communication in colums. Constructed with elpa_get_communicators(3)
 
        int             success:       return value indicating success (1) or failure (0)
 
 DESCRIPTION
        Solve the complex eigenvalue problem with the 1-stage solver. The ELPA communicators mpi_comm_rows and mpi_comm_cols are obtained with the
-       get_elpa_communicators(3) function. The distributed quadratic marix a has global dimensions na x na, and a local size lda x matrixCols.
+       elpa_get_communicators(3) function. The distributed quadratic marix a has global dimensions na x na, and a local size lda x matrixCols.
        The solver will compute the first nev eigenvalues, which will be stored on exit in ev. The eigenvectors corresponding to the eigenvalues
        will be stored in q. All memory of the arguments must be allocated outside the call to the solver.
 
@@ -357,8 +357,8 @@ SYNOPSIS
        integer, intent(in)            ldq:           leading dimension of matrix q which stores the eigenvectors
        integer, intent(in)            nblk:          blocksize of block cyclic distributin, must be the same in both directions
        integer, intent(in)            matrixCols:    number of columns of locally distributed matrices a and q
-       integer, intent(in)            mpi_comm_rows: communicator for communication in rows. Constructed with get_elpa_communicators(3)
-       integer, intent(in)            mpi_comm_cols: communicator for communication in colums. Constructed with get_elpa_communicators(3)
+       integer, intent(in)            mpi_comm_rows: communicator for communication in rows. Constructed with elpa_get_communicators(3)
+       integer, intent(in)            mpi_comm_cols: communicator for communication in colums. Constructed with elpa_get_communicators(3)
        integer, intent(in)            mpi_comm_all:  communicator for all processes in the processor set involved in ELPA
        logical, intent(in), optional: useQR:         optional argument; switches to QR-decomposition if set to .true.
        logical, intent(in), optional: useGPU:        decide whether GPUs should be used ore not
@@ -382,8 +382,8 @@ SYNOPSIS
        int     ldq:           leading dimension of matrix q which stores the eigenvectors
        int     nblk:          blocksize of block cyclic distributin, must be the same in both directions
        int     matrixCols:    number of columns of locally distributed matrices a and q
-       int     mpi_comm_rows: communicator for communication in rows. Constructed with get_elpa_communicators(3)
-       int     mpi_comm_cols: communicator for communication in colums. Constructed with get_elpa_communicators(3)
+       int     mpi_comm_rows: communicator for communication in rows. Constructed with elpa_get_communicators(3)
+       int     mpi_comm_cols: communicator for communication in colums. Constructed with elpa_get_communicators(3)
        int     mpi_comm_all:  communicator for all processes in the processor set involved in ELPA
        int     useQR:         if set to 1 switch to QR-decomposition
        int     useGPU:        decide whether the GPU version should be used or not
@@ -393,7 +393,7 @@ SYNOPSIS
 
 DESCRIPTION
        Solve the real eigenvalue problem with the 2-stage solver. The ELPA communicators mpi_comm_rows and mpi_comm_cols are obtained with the
-       get_elpa_communicators(3) function. The distributed quadratic marix a has global dimensions na x na, and a local size lda x matrixCols.
+       elpa_get_communicators(3) function. The distributed quadratic marix a has global dimensions na x na, and a local size lda x matrixCols.
        The solver will compute the first nev eigenvalues, which will be stored on exit in ev. The eigenvectors corresponding to the eigenvalues
        will be stored in q. All memory of the arguments must be allocated outside the call to the solver.
 

configure.ac

@@ -27,12 +27,13 @@ AM_SILENT_RULES([yes])
 #    by the current interface, as they are ABI compatible (e.g. only new symbols
 #    were added by the new interface)
 #
-AC_SUBST([ELPA_SO_VERSION], [10:0:2])
+AC_SUBST([ELPA_SO_VERSION], [11:0:0])
 
 # API Version
 AC_DEFINE([EARLIEST_API_VERSION], [20170403], [Earliest supported ELPA API version])
-AC_DEFINE([CURRENT_API_VERSION], [20170403], [Current ELPA API version])
-
+AC_DEFINE([CURRENT_API_VERSION], [20171201], [Current ELPA API version])
+# Autotune Version
+AC_DEFINE([EARLIEST_AUTOTUNE_VERSION], [20171201], [Earliest ELPA API version, which supports autotuning])
 
 AX_CHECK_GNU_MAKE()
 if test x$_cv_gnu_make_command = x ; then
@@ -142,7 +143,7 @@ fi
 
 dnl check which MPI binray invokes a MPI job
 if test x"$with_mpi" = x"yes"; then
-  AC_CHECK_PROGS([MPI_BINARY], [srun mpiexec.hydra mpiexec mpirun poe runjob], [no])
+  AC_CHECK_PROGS([MPI_BINARY], [mpiexec.hydra mpiexec mpirun poe runjob srun], [no])
   if test x"$MPI_BINARY" = x"no"; then
     AC_MSG_ERROR([Could not find either of the MPI binaries: mpiexec.hydra, mpiexec, mpirun, poe, runjob, srun])
   fi
@@ -188,6 +189,7 @@ fi
 AM_CONDITIONAL([HAVE_DETAILED_TIMINGS], [test x"$enable_timings" = x"yes"])
 
 dnl PAPI for ftimings
+AC_LANG_PUSH([C])
 AC_ARG_WITH([papi],
             [AS_HELP_STRING([--with-papi],
                             [Use PAPI to also measure flop count in the detailed timing (--enable-timing), disabled by default])],
@@ -202,6 +204,7 @@ if test x"${enable_timings}" = x"yes"; then
     AC_DEFINE([HAVE_LIBPAPI], [1], [Use the PAPI library])
   fi
 fi
+AC_LANG_POP([C])
 
 save_FCFLAGS=$FCFLAGS
 save_LDFLAGS=$LDFLAGS
@@ -466,6 +469,18 @@ m4_define(elpa_m4_sse_kernels, [
         complex_sse_block2
 ])
 
+m4_define(elpa_m4_sparc64_kernels, [
+        real_sparc64_block2
+        real_sparc64_block4
+        real_sparc64_block6
+])
+
+m4_define(elpa_m4_vsx_kernels, [
+        real_vsx_block2
+        real_vsx_block4
+        real_vsx_block6
+])
+
 m4_define(elpa_m4_avx_kernels, [
         real_avx_block2
         real_avx_block4
@@ -505,7 +520,7 @@ m4_define(elpa_m4_gpu_kernels, [
         complex_gpu
 ])
 
-m4_define(elpa_m4_kernel_types, [generic sse sse_assembly avx avx2 avx512 bgp bgq gpu])
+m4_define(elpa_m4_kernel_types, [generic sparc64 vsx sse sse_assembly avx avx2 avx512 bgp bgq gpu])
 
 m4_define(elpa_m4_all_kernels,
           m4_foreach_w([elpa_m4_type],
@@ -538,6 +553,8 @@ AC_DEFUN([ELPA_SELECT_KERNELS], [