summaryrefslogtreecommitdiff
path: root/var/spack/repos/builtin/packages/accfft/package.py
blob: 7009695240453cb70583ea34b193f928ad6d409e (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
# Copyright 2013-2020 Lawrence Livermore National Security, LLC and other
# Spack Project Developers. See the top-level COPYRIGHT file for details.
#
# SPDX-License-Identifier: (Apache-2.0 OR MIT)

from spack import *


class Accfft(CMakePackage, CudaPackage):
    """AccFFT extends existing FFT libraries for CUDA-enabled
    Graphics Processing Units (GPUs) to distributed memory clusters
    """

    homepage = "http://accfft.org"
    git      = "https://github.com/amirgholami/accfft.git"

    version('develop', branch='master')

    variant('pnetcdf', default=True, description='Add support for parallel NetCDF')
    variant('shared', default=True, description='Enables the build of shared libraries')

    # See: http://accfft.org/articles/install/#installing-dependencies
    depends_on('fftw precision=float,double ~mpi+openmp')

    depends_on('parallel-netcdf', when='+pnetcdf')

    # fix error [-Wc++11-narrowing]
    patch('fix_narrowing_error.patch')

    parallel = False

    def cmake_args(self):
        spec = self.spec
        args = [
            '-DFFTW_ROOT={0}'.format(spec['fftw'].prefix),
            '-DFFTW_USE_STATIC_LIBS=false',
            '-DBUILD_GPU={0}'.format('true' if '+cuda' in spec else 'false'),
            '-DBUILD_SHARED={0}'.format(
                'true' if '+shared' in spec else 'false'
            ),
        ]

        if '+cuda' in spec:
            cuda_arch = [x for x in spec.variants['cuda_arch'].value if x]
            if cuda_arch:
                args.append('-DCUDA_NVCC_FLAGS={0}'.format(
                    ' '.join(self.cuda_flags(cuda_arch))
                ))

        return args