spack/var/spack/repos/builtin/packages/npb/package.py
Todd Gamblin 62927654dd checksums: use sha256 checksums everywhere
We'd like to use a consistent checksum scheme everywhere so that we can:

    a) incorporate archive checksums into our specs and have a
       consistent hashing algorithm across all specs.

    b) index mirrors with a consistent type of checksum, and not one that
       is dependent on how spack packages are written.

- [x] convert existing md5, sha224, sha512, sha1 checksums to sha256
2019-10-12 07:19:43 -07:00

213 lines
7.7 KiB
Python

# Copyright 2013-2019 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)
import numbers
from spack import *
def is_integral(x):
"""Any integer value"""
try:
return isinstance(int(x), numbers.Integral) and not isinstance(x, bool)
except ValueError:
return False
class Npb(MakefilePackage):
"""The NAS Parallel Benchmarks (NPB) are a small set of programs
designed to help evaluate the performance of parallel supercomputers.
The benchmarks are derived from computational fluid dynamics (CFD)
applications and consist of five kernels and three pseudo-applications
in the original "pencil-and-paper" specification (NPB 1). The benchmark
suite has been extended to include new benchmarks for unstructured
adaptive mesh, parallel I/O, multi-zone applications, and computational
grids. Problem sizes in NPB are predefined and indicated as different
classes. Reference implementations of NPB are available in commonly-used
programming models like MPI and OpenMP (NPB 2 and NPB 3)."""
homepage = "https://www.nas.nasa.gov/publications/npb.html"
url = "https://www.nas.nasa.gov/assets/npb/NPB3.3.1.tar.gz"
version('3.3.1', sha256='4a8ea679b1df69f583c544c47198b3c26a50ec2bb6f8f69aef66c04c9a747d2d')
# Valid Benchmark Names
valid_names = (
'is', # Integer Sort, random memory access
'ep', # Embarrassingly Parallel
'cg', # Conjugate Gradient, irregular memory access and communication
'mg', # Multi-Grid on a sequence of meshes, long- and short-distance
# communication, memory intensive
'ft', # discrete 3D fast Fourier Transform, all-to-all communication
'bt', # Block Tri-diagonal solver
'sp', # Scalar Penta-diagonal solver
'lu', # Lower-Upper Gauss-Seidel solver
)
# Valid Benchmark Classes
valid_classes = (
'S', # Small for quick test purposes
'W', # Workstation size
'A', 'B', 'C', # standard test problems
# ~4X size increase going from one class to the next
'D', 'E', # large test problems
# ~16X size increase from each of the previous classes
)
# TODO: Combine these into a single mutually exclusive variant
variant(
'implementation',
default='mpi',
values=('serial', 'mpi', 'openmp'),
description='Selects one among the available implementations'
)
variant(
'names',
default=','.join(valid_names),
values=valid_names,
multi=True,
description='Benchmark names (comma separated list)'
)
variant(
'classes',
default=','.join(valid_classes),
values=valid_classes,
multi=True,
description='Benchmark classes (comma separated list)'
)
# This variant only applies to the MPI implementation
variant(
'nprocs',
default='1,2,4,8,16,32,64,128',
values=is_integral,
multi=True,
description='Number of processes (comma separated list)'
)
depends_on('mpi@2:', when='implementation=mpi')
phases = ['edit', 'install']
# Cannot be built in parallel
parallel = False
@property
def build_directory(self):
if 'implementation=mpi' in self.spec:
implementation = 'MPI'
elif 'implementation=openmp' in self.spec:
implementation = 'OMP'
elif 'implementation=serial' in self.spec:
implementation = 'SER'
else:
raise RuntimeError('You must choose an implementation to build')
return 'NPB{0}-{1}'.format(self.version.up_to(2), implementation)
def edit(self, spec, prefix):
names = spec.variants['names'].value
classes = spec.variants['classes'].value
nprocs = spec.variants['nprocs'].value
if 'implementation=mpi' in spec:
definitions = {
# Parallel Fortran
'MPIF77': spec['mpi'].mpif77,
'FLINK': spec['mpi'].mpif77,
'FMPI_LIB': spec['mpi'].libs.ld_flags,
'FMPI_INC': '-I' + spec['mpi'].prefix.include,
'FFLAGS': '-O3',
'FLINKFLAGS': '-O3',
# Parallel C
'MPICC': spec['mpi'].mpicc,
'CLINK': spec['mpi'].mpicc,
'CMPI_LIB': spec['mpi'].libs.ld_flags,
'CMPI_INC': '-I' + spec['mpi'].prefix.include,
'CFLAGS': '-O3',
'CLINKFLAGS': '-O3',
# Utilities C
'CC': spack_cc + ' -g',
'BINDIR': prefix.bin,
'RAND': 'randi8',
}
elif 'implementation=openmp' in spec:
definitions = {
# Parallel Fortran
'F77': spack_f77,
'FLINK': spack_f77,
'F_LIB': '',
'F_INC': '',
'FFLAGS': '-O3 ' + self.compiler.openmp_flag,
'FLINKFLAGS': '-O3 ' + self.compiler.openmp_flag,
# Parallel C
'CC': spack_cc,
'CLINK': spack_cc,
'C_LIB': '-lm',
'C_INC': '',
'CFLAGS': '-O3 ' + self.compiler.openmp_flag,
'CLINKFLAGS': '-O3 ' + self.compiler.openmp_flag,
# Utilities C
'UCC': spack_cc,
'BINDIR': prefix.bin,
'RAND': 'randi8',
'WTIME': 'wtime.c',
}
elif 'implementation=serial' in spec:
definitions = {
# Parallel Fortran
'F77': spack_f77,
'FLINK': spack_f77,
'F_LIB': '',
'F_INC': '',
'FFLAGS': '-O3',
'FLINKFLAGS': '-O3',
# Parallel C
'CC': spack_cc,
'CLINK': spack_cc,
'C_LIB': '-lm',
'C_INC': '',
'CFLAGS': '-O3',
'CLINKFLAGS': '-O3',
# Utilities C
'UCC': spack_cc,
'BINDIR': prefix.bin,
'RAND': 'randi8',
'WTIME': 'wtime.c',
}
with working_dir(self.build_directory):
with open('config/make.def', 'w') as make_def:
for key in definitions:
make_def.write('{0} = {1}\n'.format(
key, definitions[key]))
with open('config/suite.def', 'w') as suite_def:
for name in names:
for classname in classes:
# Classes C, D and E are not available for DT
if name == 'dt' and classname in ('C', 'D', 'E'):
continue
# Class E is not available for IS
if name == 'is' and classname == 'E':
continue
if 'implementation=mpi' in spec:
for nproc in nprocs:
suite_def.write('{0}\t{1}\t{2}\n'.format(
name, classname, nproc))
else:
suite_def.write('{0}\t{1}\n'.format(
name, classname))
def install(self, spec, prefix):
mkdir(prefix.bin)
with working_dir(self.build_directory):
make('suite')