Merge pull request #520 from stanmoore1/kokkos_update

# Change Log

## [2.03.05](https://github.com/kokkos/kokkos/tree/2.03.05) (2017-05-27)

[Full Changelog](https://github.com/kokkos/kokkos/compare/2.03.00...2.03.05)

**Implemented enhancements:**

- Harmonize Custom Reductions over nesting levels [\#802](https://github.com/kokkos/kokkos/issues/802)

- Prevent users directly including KokkosCore\_config.h [\#815](https://github.com/kokkos/kokkos/issues/815)

- DualView aborts on concurrent host/device modify \(in debug mode\) [\#814](https://github.com/kokkos/kokkos/issues/814)

- Abort when running on a NVIDIA CC5.0 or higher architecture with code compiled for CC \< 5.0 [\#813](https://github.com/kokkos/kokkos/issues/813)

- Add "name" function to ExecSpaces [\#806](https://github.com/kokkos/kokkos/issues/806)

- Allow null Future in task spawn dependences [\#795](https://github.com/kokkos/kokkos/issues/795)

- Add Unit Tests for Kokkos::complex [\#785](https://github.com/kokkos/kokkos/issues/785)

- Add pow function for Kokkos::complex [\#784](https://github.com/kokkos/kokkos/issues/784)

- Square root of a complex [\#729](https://github.com/kokkos/kokkos/issues/729)

- Command line processing of --threads argument prevents users from having any commandline arguments starting with --threads [\#760](https://github.com/kokkos/kokkos/issues/760)

- Protected deprecated API with appropriate macro [\#756](https://github.com/kokkos/kokkos/issues/756)

- Allow task scheduler memory pool to be used by tasks [\#747](https://github.com/kokkos/kokkos/issues/747)

- View bounds checking on host-side performance: constructing a std::string [\#723](https://github.com/kokkos/kokkos/issues/723)

- Add check for AppleClang as compiler distinct from check for Clang. [\#705](https://github.com/kokkos/kokkos/issues/705)

- Uninclude source files for specific configurations to prevent link warning. [\#701](https://github.com/kokkos/kokkos/issues/701)

- Add --small option to snapshot script [\#697](https://github.com/kokkos/kokkos/issues/697)

- CMake Standalone Support [\#674](https://github.com/kokkos/kokkos/issues/674)

- CMake build unit test and install [\#808](https://github.com/kokkos/kokkos/issues/808)

- CMake: Fix having kokkos as a subdirectory in a pure cmake project [\#629](https://github.com/kokkos/kokkos/issues/629)

- Tribits macro assumes build directory is in top level source directory [\#654](https://github.com/kokkos/kokkos/issues/654)

- Use bin/nvcc\_wrapper, not config/nvcc\_wrapper [\#562](https://github.com/kokkos/kokkos/issues/562)

- Allow MemoryPool::allocate\(\) to be called from multiple threads per warp. [\#487](https://github.com/kokkos/kokkos/issues/487)

- Allow MemoryPool::allocate\\(\\) to be called from multiple threads per warp. [\#487](https://github.com/kokkos/kokkos/issues/487)

- Move OpenMP 4.5 OpenMPTarget backend into Develop [\#456](https://github.com/kokkos/kokkos/issues/456)

- Testing on ARM testbed [\#288](https://github.com/kokkos/kokkos/issues/288)

**Fixed bugs:**

- Fix label in OpenMP parallel\_reduce verify\_initialized [\#834](https://github.com/kokkos/kokkos/issues/834)

- TeamScratch Level 1 on Cuda hangs [\#820](https://github.com/kokkos/kokkos/issues/820)

- \[bug\] memory pool. [\#786](https://github.com/kokkos/kokkos/issues/786)

- Some Reduction Tests fail on Intel 18 with aggressive vectorization on [\#774](https://github.com/kokkos/kokkos/issues/774)

- Error copying dynamic view on copy of memory pool [\#773](https://github.com/kokkos/kokkos/issues/773)

- CUDA stack overflow with TaskDAG test [\#758](https://github.com/kokkos/kokkos/issues/758)

- ThreadVectorRange Customized Reduction Bug [\#739](https://github.com/kokkos/kokkos/issues/739)

- set\_scratch\_size overflows  [\#726](https://github.com/kokkos/kokkos/issues/726)

- Get wrong results for compiler checks in Makefile on OS X. [\#706](https://github.com/kokkos/kokkos/issues/706)

- Fix check if multiple host architectures enabled. [\#702](https://github.com/kokkos/kokkos/issues/702)

- Threads Backend Does not Pass on Cray Compilers [\#609](https://github.com/kokkos/kokkos/issues/609)

- Rare bug in memory pool where allocation can finish on superblock in empty state [\#452](https://github.com/kokkos/kokkos/issues/452)

- LDFLAGS in core/unit\_test/Makefile: potential "undefined reference" to pthread lib [\#148](https://github.com/kokkos/kokkos/issues/148)

## [2.03.00](https://github.com/kokkos/kokkos/tree/2.03.00) (2017-04-25)

[Full Changelog](https://github.com/kokkos/kokkos/compare/2.02.15...2.03.00)

ENDIF()

IF(NOT KOKKOS_HAS_TRILINOS)

  CMAKE_MINIMUM_REQUIRED(VERSION 2.8.11 FATAL_ERROR)

  INCLUDE(cmake/tribits.cmake)

  SET(CMAKE_CXX_STANDARD 11)

ENDIF()

  cmake_minimum_required(VERSION 3.1 FATAL_ERROR)

  project(Kokkos CXX)

  INCLUDE(cmake/kokkos.cmake)

ELSE()

#------------------------------------------------------------------------------

#

# A) Forward delcare the package so that certain options are also defined for

# subpackages

TRIBITS_PACKAGE_DECL(Kokkos) # ENABLE_SHADOWING_WARNINGS)

#------------------------------------------------------------------------------

#

# B) Define the common options for Kokkos first so they can be used by

# subpackages as well.

#

# mfh 01 Aug 2016: See Issue #61:

#

# https://github.com/kokkos/kokkos/issues/61

  )

TRIBITS_ADD_OPTION_AND_DEFINE(

  Kokkos_ENABLE_Qthreads

  Kokkos_ENABLE_QTHREAD

  KOKKOS_HAVE_QTHREADS

  "Enable Qthreads support in Kokkos."

  "${TPL_ENABLE_QTHREADS}"

  "${TPL_ENABLE_QTHREAD}"

  )

# TODO: No longer an option in Kokkos.  Needs to be removed.

TRIBITS_ADD_OPTION_AND_DEFINE(

  Kokkos_ENABLE_CXX11

  KOKKOS_HAVE_CXX11

  "${TPL_ENABLE_HWLOC}"

  )

# TODO: This is currently not used in Kokkos.  Should it be removed?

TRIBITS_ADD_OPTION_AND_DEFINE(

  Kokkos_ENABLE_MPI

  KOKKOS_HAVE_MPI

  "${Kokkos_ENABLE_Debug_Bounds_Check_DEFAULT}"

  )

TRIBITS_ADD_OPTION_AND_DEFINE(

  Kokkos_ENABLE_Debug_DualView_Modify_Check

  KOKKOS_ENABLE_DEBUG_DUALVIEW_MODIFY_CHECK

  "Enable abort when Kokkos::DualView modified on host and device without sync."

  "${Kokkos_ENABLE_DEBUG}"

  )

TRIBITS_ADD_OPTION_AND_DEFINE(

  Kokkos_ENABLE_Profiling

  KOKKOS_ENABLE_PROFILING_INTERNAL

  KOKKOS_ENABLE_PROFILING

  "Enable KokkosP profiling support for kernel data collections."

  "${TPL_ENABLE_DLlib}"

  )

TRIBITS_ADD_OPTION_AND_DEFINE(

  Kokkos_ENABLE_Profiling_Load_Print

  KOKKOS_ENABLE_PROFILING_LOAD_PRINT

  "Print to standard output which profiling library was loaded."

  OFF

  )

# placeholder for future device...

TRIBITS_ADD_OPTION_AND_DEFINE(

  Kokkos_ENABLE_Winthread

  "${TPL_ENABLE_Winthread}"

  )

# TODO: No longer an option in Kokkos.  Needs to be removed.

# use new/old View

TRIBITS_ADD_OPTION_AND_DEFINE(

  Kokkos_USING_DEPRECATED_VIEW

  OFF

  )

#------------------------------------------------------------------------------

#

# C) Install Kokkos' executable scripts

#

# nvcc_wrapper is Kokkos' wrapper for NVIDIA's NVCC CUDA compiler.

# Kokkos needs nvcc_wrapper in order to build.  Other libraries and

# executables also need nvcc_wrapper.  Thus, we need to install it.

TRIBITS_PROCESS_SUBPACKAGES()

#------------------------------------------------------------------------------

#

# E) If Kokkos itself is enabled, process the Kokkos package

#

  )

TRIBITS_PACKAGE_POSTPROCESS()

ENDIF()

endif

ifeq ($(KOKKOS_INTERNAL_USE_OPENMP), 1)

Kokkos_OpenMPexec.o: $(KOKKOS_CPP_DEPENDS) $(KOKKOS_PATH)/core/src/OpenMP/Kokkos_OpenMPexec.cpp

	$(CXX) $(KOKKOS_CPPFLAGS) $(KOKKOS_CXXFLAGS) $(CXXFLAGS) -c $(KOKKOS_PATH)/core/src/OpenMP/Kokkos_OpenMPexec.cpp

Kokkos_OpenMP_Exec.o: $(KOKKOS_CPP_DEPENDS) $(KOKKOS_PATH)/core/src/OpenMP/Kokkos_OpenMP_Exec.cpp

	$(CXX) $(KOKKOS_CPPFLAGS) $(KOKKOS_CXXFLAGS) $(CXXFLAGS) -c $(KOKKOS_PATH)/core/src/OpenMP/Kokkos_OpenMP_Exec.cpp

Kokkos_OpenMP_Task.o: $(KOKKOS_CPP_DEPENDS) $(KOKKOS_PATH)/core/src/OpenMP/Kokkos_OpenMP_Task.cpp

	$(CXX) $(KOKKOS_CPPFLAGS) $(KOKKOS_CXXFLAGS) $(CXXFLAGS) -c $(KOKKOS_PATH)/core/src/OpenMP/Kokkos_OpenMP_Task.cpp

endif

ifeq ($(KOKKOS_INTERNAL_USE_OPENMPTARGET), 1)

Kokkos_OpenMPTarget_Exec.o: $(KOKKOS_CPP_DEPENDS) $(KOKKOS_PATH)/core/src/OpenMPTarget/Kokkos_OpenMPTarget_Exec.cpp

	$(CXX) $(KOKKOS_CPPFLAGS) $(KOKKOS_CXXFLAGS) $(CXXFLAGS) -c $(KOKKOS_PATH)/core/src/OpenMPTarget/Kokkos_OpenMPTarget_Exec.cpp

Kokkos_OpenMPTargetSpace.o: $(KOKKOS_CPP_DEPENDS) $(KOKKOS_PATH)/core/src/OpenMPTarget/Kokkos_OpenMPTargetSpace.cpp

	$(CXX) $(KOKKOS_CPPFLAGS) $(KOKKOS_CXXFLAGS) $(CXXFLAGS) -c $(KOKKOS_PATH)/core/src/OpenMPTarget/Kokkos_OpenMPTargetSpace.cpp

#Kokkos_OpenMPTarget_Task.o: $(KOKKOS_CPP_DEPENDS) $(KOKKOS_PATH)/core/src/OpenMPTarget/Kokkos_OpenMPTarget_Task.cpp

#       $(CXX) $(KOKKOS_CPPFLAGS) $(KOKKOS_CXXFLAGS) $(CXXFLAGS) -c $(KOKKOS_PATH)/core/src/OpenMPTarget/Kokkos_OpenMPTarget_Task.cpp

endif

Kokkos_HBWSpace.o: $(KOKKOS_CPP_DEPENDS) $(KOKKOS_PATH)/core/src/impl/Kokkos_HBWSpace.cpp

	$(CXX) $(KOKKOS_CPPFLAGS) $(KOKKOS_CXXFLAGS) $(CXXFLAGS) -c $(KOKKOS_PATH)/core/src/impl/Kokkos_HBWSpace.cpp

void KOKKOS_ALGORITHMS_SRC_DUMMY_PREVENT_LINK_ERROR() {}