dbcsr: expose all options, check openblas feats (#16034)

* dbcsr: expose all options, check openblas feats

* dbcsr: use Ninja to build, ensure serialized tests

* dbcsr: add myself as maintainer

1 files changed, 60 insertions, 7 deletions

diff --git a/var/spack/repos/builtin/packages/dbcsr/package.py b/var/spack/repos/builtin/packages/dbcsr/package.py
index c917ba010b..028efb0e1d 100644
--- a/var/spack/repos/builtin/packages/dbcsr/package.py
+++ b/var/spack/repos/builtin/packages/dbcsr/package.py
@@ -6,36 +6,89 @@
 from spack import *
 
 
-class Dbcsr(CMakePackage):
+class Dbcsr(CMakePackage, CudaPackage):
     """Distributed Block Compressed Sparse Row matrix library."""
 
     homepage = "https://github.com/cp2k/dbcsr"
     git      = "https://github.com/cp2k/dbcsr.git"
 
+    maintainers = ['dev-zero']
+
     version('develop', branch='develop')
 
     variant('mpi',    default=True,  description='Compile with MPI')
     variant('openmp', default=False, description='Build with OpenMP support')
+    variant('shared', default=True,  description='Build shared library')
+    variant('smm', default='libxsmm', values=('libxsmm', 'blas'),
+            description='Library for small matrix multiplications')
+    variant('cuda_arch',
+            description='CUDA architecture',
+            default='none',
+            values=('none', '35', '37', '60', '70'),
+            multi=False)
+    variant('cuda_arch_35_k20x', default=False,
+            description=('CP2K (resp. DBCSR) has specific parameter sets for'
+                         ' different GPU models. Enable this when building'
+                         ' with cuda_arch=35 for a K20x instead of a K40'))
 
     depends_on('blas')
     depends_on('lapack')
     depends_on('mpi', when='+mpi')
-    depends_on('py-fypp')
+    depends_on('libxsmm@1.11:~header-only', when='smm=libxsmm')
+
+    depends_on('cmake@3.12:', type='build')
+    depends_on('py-fypp', type='build')
+    depends_on('pkgconfig', type='build')
+    depends_on('python@3.6:', type='build', when='+cuda')
+
+    conflicts('+cuda', when='cuda_arch=none')
+
+    generator = 'Ninja'
+    depends_on('ninja@1.10:', type='build')
 
     def cmake_args(self):
+        if ('+openmp' in self.spec
+            and '^openblas' in self.spec
+            and '^openblas threads=openmp' not in self.spec):
+            raise InstallError(
+                '^openblas threads=openmp required for dbcsr+openmp')
+
         spec = self.spec
         args = [
+            '-DUSE_SMM=%s' % ('libxsmm' if 'smm=libxsmm' in spec else 'blas'),
             '-DUSE_MPI=%s' % ('ON' if '+mpi' in spec else 'OFF'),
             '-DUSE_OPENMP=%s' % (
                 'ON' if '+openmp' in spec else 'OFF'),
-            # C API need MPI
+            # C API needs MPI
             '-DWITH_C_API=%s' % ('ON' if '+mpi' in spec else 'OFF'),
-            '-DLAPACK_FOUND=true',
-            '-DLAPACK_LIBRARIES=%s' % spec['lapack'].libs.joined(';'),
             '-DBLAS_FOUND=true',
-            '-DBLAS_LIBRARIES=%s' % spec['blas'].libs.joined(';'),
+            '-DBLAS_LIBRARIES=%s' % (spec['blas'].libs.joined(';')),
+            '-DLAPACK_FOUND=true',
+            '-DLAPACK_LIBRARIES=%s' % (spec['lapack'].libs.joined(';')),
             '-DWITH_EXAMPLES=ON',
-            '-DBUILD_SHARED_LIBS=ON'
+            '-DBUILD_SHARED_LIBS=%s' % ('ON' if '+shared' in spec else 'OFF'),
         ]
 
+        if '+cuda' in self.spec:
+            cuda_arch = self.spec.variants['cuda_arch'].value
+
+            gpuver = {
+                '35': 'K40',
+                '37': 'K80',
+                '60': 'P100',
+                '70': 'V100',
+            }[cuda_arch]
+
+            if (cuda_arch == '35'
+                    and self.spec.satisfies('+cuda_arch_35_k20x')):
+                gpuver = 'K20X'
+
+            args += ['-DWITH_GPU=%s' % gpuver]
+
         return args
+
+    def check(self):
+        """Override CMakePackage's check() to enforce seralized test runs
+           since they are already parallelized"""
+        with working_dir(self.build_directory):
+            self._if_ninja_target_execute('test', parallel=False)