
* exawind: updates to package to allow mixed device * Style. * Remove ninja variants. * Add conflict for amr-wind+hypre with mixed device. * Relax amr-wind~hypre requirement. * Move runtime variables to nalu-wind. * Update suggestions. * Remove umpire.
209 lines
8.4 KiB
Python
209 lines
8.4 KiB
Python
# Copyright 2013-2024 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.package import *
|
|
|
|
|
|
def _parse_float(val):
|
|
try:
|
|
return float(val) > 0.0
|
|
except ValueError:
|
|
return False
|
|
|
|
|
|
class NaluWind(CMakePackage, CudaPackage, ROCmPackage):
|
|
"""Nalu-Wind: Wind energy focused variant of Nalu."""
|
|
|
|
homepage = "https://nalu-wind.readthedocs.io"
|
|
git = "https://github.com/exawind/nalu-wind.git"
|
|
|
|
maintainers("jrood-nrel", "psakievich")
|
|
|
|
tags = ["ecp", "ecp-apps"]
|
|
|
|
version("master", branch="master")
|
|
version("2.0.0", tag="v2.0.0")
|
|
|
|
variant("pic", default=True, description="Position independent code")
|
|
variant(
|
|
"abs_tol",
|
|
default=1.0e-15,
|
|
values=_parse_float,
|
|
description="Absolute tolerance for regression tests",
|
|
)
|
|
variant(
|
|
"rel_tol",
|
|
default=1.0e-12,
|
|
values=_parse_float,
|
|
description="Relative tolerance for regression tests",
|
|
)
|
|
variant("openfast", default=False, description="Compile with OpenFAST support")
|
|
variant("tioga", default=False, description="Compile with Tioga support")
|
|
variant("hypre", default=True, description="Compile with Hypre support")
|
|
variant("trilinos-solvers", default=True, description="Compile with Trilinos Solvers support")
|
|
variant("catalyst", default=False, description="Compile with Catalyst support")
|
|
variant("shared", default=True, description="Build shared libraries")
|
|
variant("fftw", default=False, description="Compile with FFTW support")
|
|
variant("fsi", default=False, description="Enable fluid-structure-interaction models")
|
|
variant("boost", default=False, description="Enable Boost integration")
|
|
variant("gpu-aware-mpi", default=False, description="gpu-aware-mpi")
|
|
variant("wind-utils", default=False, description="Build wind-utils")
|
|
variant("umpire", default=False, description="Enable Umpire")
|
|
|
|
depends_on("mpi")
|
|
depends_on("yaml-cpp@0.5.3:")
|
|
depends_on("openfast@4.0.0:+cxx+netcdf", when="+fsi")
|
|
depends_on("trilinos@13.4.1", when="@=2.0.0")
|
|
depends_on("hypre@2.29.0:", when="@2.0.0:+hypre")
|
|
depends_on(
|
|
"trilinos@13:+exodus+tpetra+zoltan+stk~superlu-dist~superlu+hdf5+shards~hypre+gtest "
|
|
"gotype=long cxxstd=17"
|
|
)
|
|
depends_on("trilinos~cuda~wrapper", when="~cuda")
|
|
depends_on("openfast@2.6.0: +cxx", when="+openfast")
|
|
depends_on("tioga@master:", when="+tioga")
|
|
depends_on("hypre@2.18.2: ~int64+mpi~superlu-dist", when="+hypre")
|
|
depends_on("trilinos+muelu+belos+amesos2+ifpack2", when="+trilinos-solvers")
|
|
conflicts(
|
|
"~hypre~trilinos-solvers",
|
|
msg="nalu-wind: Must enable at least one of the linear-solvers: hypre or trilinos-solvers",
|
|
)
|
|
depends_on("kokkos-nvcc-wrapper", type="build", when="+cuda")
|
|
for _arch in CudaPackage.cuda_arch_values:
|
|
depends_on(
|
|
"trilinos~shared+cuda+cuda_rdc+wrapper cuda_arch={0}".format(_arch),
|
|
when="+cuda cuda_arch={0}".format(_arch),
|
|
)
|
|
depends_on(
|
|
"hypre@develop +mpi+cuda~int64~superlu-dist cuda_arch={0}".format(_arch),
|
|
when="+hypre+cuda cuda_arch={0}".format(_arch),
|
|
)
|
|
for _arch in ROCmPackage.amdgpu_targets:
|
|
depends_on(
|
|
"trilinos@13.4: ~shared+rocm amdgpu_target={0}".format(_arch),
|
|
when="+rocm amdgpu_target={0}".format(_arch),
|
|
)
|
|
depends_on(
|
|
"hypre+rocm amdgpu_target={0}".format(_arch),
|
|
when="+hypre+rocm amdgpu_target={0}".format(_arch),
|
|
)
|
|
|
|
depends_on("trilinos-catalyst-ioss-adapter", when="+catalyst")
|
|
depends_on("fftw+mpi", when="+fftw")
|
|
depends_on("nccmp")
|
|
# indirect dependency needed to make original concretizer work
|
|
depends_on("netcdf-c+parallel-netcdf")
|
|
depends_on("boost +filesystem +iostreams cxxstd=14", when="+boost")
|
|
depends_on("hypre+gpu-aware-mpi", when="+gpu-aware-mpi")
|
|
depends_on("hypre+umpire", when="+umpire")
|
|
depends_on("trilinos~shared", when="+trilinos-solvers platform=darwin")
|
|
|
|
conflicts(
|
|
"+shared",
|
|
when="+cuda",
|
|
msg="invalid device functions are generated with shared libs and cuda",
|
|
)
|
|
conflicts(
|
|
"+shared",
|
|
when="+rocm",
|
|
msg="invalid device functions are generated with shared libs and rocm",
|
|
)
|
|
conflicts("+cuda", when="+rocm")
|
|
conflicts("+rocm", when="+cuda")
|
|
conflicts("^hypre+cuda", when="~cuda")
|
|
conflicts("^hypre+rocm", when="~rocm")
|
|
conflicts("^hypre+sycl")
|
|
conflicts("^trilinos+cuda", when="~cuda")
|
|
conflicts("^trilinos+rocm", when="~rocm")
|
|
|
|
def setup_dependent_run_environment(self, env, dependent_spec):
|
|
spec = self.spec
|
|
if spec.satisfies("+cuda") or spec.satisfies("+rocm"):
|
|
env.set("CUDA_LAUNCH_BLOCKING", "1")
|
|
env.set("CUDA_MANAGED_FORCE_DEVICE_ALLOC", "1")
|
|
env.set("HIP_LAUNCH_BLOCKING", "1")
|
|
env.set("HIP_MANAGED_FORCE_DEVICE_ALLOC", "1")
|
|
|
|
def setup_build_environment(self, env):
|
|
spec = self.spec
|
|
env.append_flags("CXXFLAGS", "-DUSE_STK_SIMD_NONE")
|
|
if spec.satisfies("+cuda"):
|
|
env.set("OMPI_CXX", self.spec["kokkos-nvcc-wrapper"].kokkos_cxx)
|
|
env.set("MPICH_CXX", self.spec["kokkos-nvcc-wrapper"].kokkos_cxx)
|
|
env.set("MPICXX_CXX", self.spec["kokkos-nvcc-wrapper"].kokkos_cxx)
|
|
if spec.satisfies("+rocm"):
|
|
env.append_flags("CXXFLAGS", "-fgpu-rdc")
|
|
|
|
def cmake_args(self):
|
|
spec = self.spec
|
|
|
|
args = [
|
|
self.define_from_variant("CMAKE_POSITION_INDEPENDENT_CODE", "pic"),
|
|
self.define("CMAKE_CXX_COMPILER", spec["mpi"].mpicxx),
|
|
self.define("CMAKE_Fortran_COMPILER", spec["mpi"].mpifc),
|
|
self.define("Trilinos_DIR", spec["trilinos"].prefix),
|
|
self.define("YAML_DIR", spec["yaml-cpp"].prefix),
|
|
self.define_from_variant("ENABLE_CUDA", "cuda"),
|
|
self.define_from_variant("ENABLE_WIND_UTILS", "wind-utils"),
|
|
self.define_from_variant("ENABLE_BOOST", "boost"),
|
|
self.define_from_variant("CMAKE_CXX_STANDARD", "17"),
|
|
self.define_from_variant("BUILD_SHARED_LIBS", "shared"),
|
|
self.define_from_variant("ENABLE_OPENFAST", "openfast"),
|
|
self.define_from_variant("ENABLE_TIOGA", "tioga"),
|
|
self.define_from_variant("ENABLE_HYPRE", "hypre"),
|
|
self.define_from_variant("ENABLE_TRILINOS_SOLVERS", "trilinos-solvers"),
|
|
self.define_from_variant("ENABLE_PARAVIEW_CATALYST", "catalyst"),
|
|
self.define_from_variant("ENABLE_FFTW", "fftw"),
|
|
self.define_from_variant("ENABLE_UMPIRE", "umpire"),
|
|
]
|
|
|
|
if spec.satisfies("+openfast"):
|
|
args.append(self.define("OpenFAST_DIR", spec["openfast"].prefix))
|
|
|
|
if spec.satisfies("+tioga"):
|
|
args.append(self.define("TIOGA_DIR", spec["tioga"].prefix))
|
|
|
|
if spec.satisfies("+hypre"):
|
|
args.append(self.define("HYPRE_DIR", spec["hypre"].prefix))
|
|
|
|
if spec.satisfies("+catalyst"):
|
|
args.append(
|
|
self.define(
|
|
"PARAVIEW_CATALYST_INSTALL_PATH", spec["trilinos-catalyst-ioss-adapter"].prefix
|
|
)
|
|
)
|
|
|
|
if spec.satisfies("+fftw"):
|
|
args.append(self.define("FFTW_DIR", spec["fftw"].prefix))
|
|
|
|
args.append(self.define("ENABLE_TESTS", self.run_tests))
|
|
if self.run_tests:
|
|
args.extend(
|
|
[
|
|
self.define("TEST_TOLERANCE", spec.variants["abs_tol"].value),
|
|
self.define("TEST_REL_TOL", spec.variants["rel_tol"].value),
|
|
]
|
|
)
|
|
|
|
if spec.satisfies("+umpire"):
|
|
args.append(self.define("UMPIRE_DIR", spec["umpire"].prefix))
|
|
|
|
if spec.satisfies("+rocm"):
|
|
args.append(self.define("CMAKE_CXX_COMPILER", spec["hip"].hipcc))
|
|
args.append(self.define("ENABLE_ROCM", True))
|
|
targets = spec.variants["amdgpu_target"].value
|
|
args.append(self.define("GPU_TARGETS", ";".join(str(x) for x in targets)))
|
|
|
|
if "darwin" in spec.architecture:
|
|
args.append(self.define("CMAKE_MACOSX_RPATH", "ON"))
|
|
|
|
return args
|
|
|
|
@run_before("cmake")
|
|
def add_submodules(self):
|
|
if self.run_tests or self.spec.satisfies("+wind-utils"):
|
|
git = which("git")
|
|
git("submodule", "update", "--init", "--recursive")
|