spack/var/spack/repos/builtin/packages/hiop/package.py
Asher Mancinelli 078ee48c4b
ExaGO: add v1.4.0 and v1.4.1 (#29712)
* ExaGO: Handling of cuda architectures and amdgpu targets changed
  to effectively handle multiple targets. See #28441.
* Add ROCm support to ExaGO and update ROCm support in HiOp
* ExaGO+rocm requires HiOp+rocm
* Newer versions of CMake may set HIP_CLANG_INCLUDE_PATH incorrectly:
  add comments to the ExaGO/HiOp packages explaining how to address
  this problem if it occurs.
2022-03-29 10:01:30 -07:00

189 lines
8.1 KiB
Python

# Copyright 2013-2022 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 Hiop(CMakePackage, CudaPackage, ROCmPackage):
"""HiOp is an optimization solver for solving certain mathematical
optimization problems expressed as nonlinear programming problems.
HiOp is a lightweight HPC solver that leverages application"s existing
data parallelism to parallelize the optimization iterations by using
specialized linear algebra kernels."""
homepage = 'https://github.com/LLNL/hiop'
git = 'https://github.com/LLNL/hiop.git'
maintainers = ['ashermancinelli', 'CameronRutherford']
# Most recent tagged snapshot is the preferred version when profiling.
version('0.5.4', commit='a37a7a677884e95d1c0ad37936aef3778fc91c3e')
version('0.5.3', commit='698e8d0fdc0ff9975d8714339ff8c782b70d85f9')
version('0.5.2', commit='662ad76dee1f501f648a8bec9a490cb5881789e9')
version('0.5.1', commit='6789bbb55824e68e428c2df1009d647af81f9cf1')
version('0.5.0', commit='a39da8025037c7c8ae2eb31234eb80cc73bec2af')
version('0.4.6', commit='b72d163d52c9225c3196ceb2baebdc7cf09a69de')
version('0.4.5', commit='c353580456c4776c50811b97cf8ff802dc27b90c')
version('0.4.4', commit='e858eefa6b914f5c87c3717bbce811931ea69386')
version('0.4.3', commit='c0394af4d84ebb84b7d2b95283ad65ffd84e0d45')
version('0.4.2', commit='3fcb788d223eec24c0241680070c4a9a5ec71ef3')
version('0.4.1', commit='3f269560f76d5a89bcbd1d3c4f9f0e5acaa6fd64')
version('0.4', commit='91d21085a1149eacdb27cd738d4a74a7e412fcff')
version('0.3.99.3', commit='bed1dbef260e53a9d139ccfb77d2e83a98aab216')
version('0.3.99.2', commit='9eb026768bc5e0a2c1293d0487cc39913001ae19')
version('0.3.99.1', commit='220e32c0f318665d6d394ca3cd0735b9d26a65eb')
version('0.3.99.0', commit='589b9c76781447108fa55788d5fa1b83ff71a3d1')
version('0.3', commit='7e8adae9db757aed48e5c2bc448316307598258f')
version('0.2', commit='c52a6f6b9baaaa2d7f233a749aa98f901349723f')
version('0.1', commit='5f60e11b79d532115fb41694378b54c9c707aad9')
# Development branches
version('master', branch='master')
version('develop', branch='develop')
variant(
'jsrun', default=False, description='Enable/Disable jsrun command for testing'
)
variant('shared', default=False, description='Enable/Disable shared libraries')
variant('mpi', default=True, description='Enable/Disable MPI')
variant('raja', default=False, description='Enable/Disable RAJA')
variant('kron', default=False, description='Enable/Disable Kron reduction')
variant('sparse', default=False, description='Enable/Disable Sparse linear algebra')
variant(
'deepchecking',
default=False,
description='Ultra safety checks - '
'used for increased robustness and self-diagnostics',
)
depends_on('lapack')
depends_on('blas')
depends_on('cmake@3.18:', type='build')
depends_on('mpi', when='+mpi')
depends_on('magma+cuda', when='+cuda')
depends_on('magma+rocm', when='+rocm')
# Depends on Magma when +rocm or +cuda
magma_ver_constraints = (
('2.5.4', '0.4'),
('2.6.1', '0.4.6'),
('2.6.2', '0.5.4'),
)
for (magma_v, hiop_v) in magma_ver_constraints:
depends_on('magma@{0}:'.format(magma_v), when='@{0}:+cuda'.format(hiop_v))
depends_on('magma@{0}:'.format(magma_v), when='@{0}:+rocm'.format(hiop_v))
depends_on('raja', when='+raja')
depends_on('raja+openmp', when='+raja~cuda~rocm')
depends_on('raja@0.14.0:', when='@0.5.0:+raja')
depends_on('raja+cuda', when='+raja+cuda')
depends_on('raja+rocm', when='+raja+rocm')
depends_on('umpire', when='+raja')
depends_on('umpire+cuda~shared', when='+raja+cuda')
depends_on('umpire+rocm', when='+raja+rocm')
depends_on('umpire@6.0.0:', when='@0.5.0:+raja')
depends_on('hip', when='+rocm')
depends_on('hipblas', when='+rocm')
depends_on('hipsparse', when='+rocm')
depends_on('suite-sparse', when='+kron')
depends_on('coinhsl+blas', when='+sparse')
depends_on('metis', when='+sparse')
conflicts(
'+shared',
when='+cuda+raja',
msg='umpire+cuda exports device code and requires static libs',
)
flag_handler = build_system_flags
def cmake_args(self):
args = []
spec = self.spec
use_gpu = '+cuda' in spec or '+rocm' in spec
if use_gpu:
args.extend([
self.define('HIOP_USE_GPU', True),
self.define('HIOP_USE_MAGMA', True),
self.define('HIOP_MAGMA_DIR', spec['magma'].prefix),
])
args.extend([
self.define('HIOP_BUILD_STATIC', True),
self.define('LAPACK_FOUND', True),
self.define('LAPACK_LIBRARIES', spec['lapack'].libs + spec['blas'].libs),
self.define_from_variant('HIOP_BUILD_SHARED', 'shared'),
self.define_from_variant('HIOP_USE_MPI', 'mpi'),
self.define_from_variant('HIOP_DEEPCHECKS', 'deepchecking'),
self.define_from_variant('HIOP_USE_CUDA', 'cuda'),
self.define_from_variant('HIOP_USE_HIP', 'rocm'),
self.define_from_variant('HIOP_USE_RAJA', 'raja'),
self.define_from_variant('HIOP_USE_UMPIRE', 'raja'),
self.define_from_variant('HIOP_WITH_KRON_REDUCTION', 'kron'),
self.define_from_variant('HIOP_SPARSE', 'sparse'),
self.define_from_variant('HIOP_USE_COINHSL', 'sparse'),
self.define_from_variant('HIOP_TEST_WITH_BSUB', 'jsrun'),
])
# NOTE: If building with spack develop on a cluster, you may want to
# change the ctest launch command to use your job scheduler like so:
#
# args.append(
# self.define('HIOP_CTEST_LAUNCH_COMMAND', 'srun -t 10:00'))
if '+mpi' in spec:
args.extend([
self.define('MPI_HOME', spec['mpi'].prefix),
self.define('MPI_C_COMPILER', spec['mpi'].mpicc),
self.define('MPI_CXX_COMPILER', spec['mpi'].mpicxx),
self.define('MPI_Fortran_COMPILER', spec['mpi'].mpifc),
])
# NOTE: On Cray platforms, libfabric is occasionally not picked up
# by Spack, causing HiOp's CMake code to fail to find MPI Fortran
# libraries. If this is the case, adding the following lines may
# resolve the issue. Searching <builddir>/CMakeFiles/CMakeError.log
# for MPI Fortran errors is the fastest way to check for this error.
#
# args.append(
# self.define('MPI_Fortran_LINK_FLAGS',
# '-L/path/to/libfabric/lib64/ -lfabric'))
if '+cuda' in spec:
cuda_arch_list = spec.variants['cuda_arch'].value
if cuda_arch_list[0] != 'none':
args.append(self.define('CMAKE_CUDA_ARCHITECTURES', cuda_arch_list))
# NOTE: if +rocm, some HIP CMake variables may not be set correctly.
# Namely, HIP_CLANG_INCLUDE_PATH. If the configure phase fails due to
# this variable being undefined, adding the following line typically
# resolves this issue:
#
# args.append(
# self.define('HIP_CLANG_INCLUDE_PATH',
# '/opt/rocm-X.Y.Z/llvm/lib/clang/14.0.0/include/'))
if '+rocm' in spec:
rocm_arch_list = spec.variants['amdgpu_target'].value
if rocm_arch_list[0] != 'none':
args.append(self.define('GPU_TARGETS', rocm_arch_list))
args.append(self.define('AMDGPU_TARGETS', rocm_arch_list))
if '+kron' in spec:
args.append(self.define('HIOP_UMFPACK_DIR', spec['suite-sparse'].prefix))
# Unconditionally disable strumpack, even when +sparse. This may be
# used in place of COINHSL for sparse interface, however this is not
# fully supported in spack at the moment.
args.append(self.define('HIOP_USE_STRUMPACK', False))
if '+sparse' in spec:
args.append(self.define('HIOP_COINHSL_DIR', spec['coinhsl'].prefix))
return args