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
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
|
# 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)
class Minigmg(Package):
"""miniGMG is a compact benchmark for understanding the performance
challenges associated with geometric multigrid solvers
found in applications built from AMR MG frameworks
like CHOMBO or BoxLib when running
on modern multi- and manycore-based supercomputers.
It includes both productive reference examples as well as
highly-optimized implementations for CPUs and GPUs.
It is sufficiently general that it has been used to evaluate
a broad range of research topics including PGAS programming models
and algorithmic tradeoffs inherit in multigrid. miniGMG was developed
under the CACHE Joint Math-CS Institute.
Note, miniGMG code has been supersceded by HPGMG. """
homepage = "http://crd.lbl.gov/departments/computer-science/PAR/research/previous-projects/miniGMG/"
url = "https://crd.lbl.gov/assets/Uploads/FTG/Projects/miniGMG/miniGMG.tar.gz"
version('master', sha256='1c2d27496a881f655f5e849d6a7a132625e535739f82575991c511cc2cf899ac')
variant('vec', default='ompif', description='Which method of vectorisation to use',
values=('ompif', 'sse', 'avx', 'simde'), multi=False)
variant('opt', default=False, description='Enable optimization flags for improved OpenMP')
depends_on('mpi')
# Set up SIMD Everywhere config
depends_on('simde', when='vec=simde')
patch('simde.patch', when='vec=simde')
# Patch to add timer for Aarch64 rather than rdtsc
patch('aarch64_time.patch', when='target=aarch64:')
# Replaces inline with inline static, for correct syntax
patch('inline_static.patch')
def install(self, spec, prefix):
cc = Executable(spec['mpi'].mpicc)
args = []
# Default optimisation level
if spec.satisfies('+opt'):
if self.spec.satisfies('%nvhpc'):
args.append('-fast')
else:
args.append('-Ofast')
else:
args.append('-O3')
# Add OpenMP flag
args += [self.compiler.openmp_flag]
args += ['miniGMG.c', 'mg.c', 'box.c', 'solver.c']
# Set the correct operators file - using the vec variant
if spec.satisfies('vec=sse'):
args += ['operators.sse.c']
elif spec.satisfies('vec=avx'):
args += ['operators.avx.c']
elif spec.satisfies('vec=simde'):
args += ['operators.simde.c']
else:
args += ['operators.ompif.c']
# Switch out timer file (depends on patch)
if spec.satisfies('target=aarch64:'):
args += ['timer.aarch64.c']
else:
args += ['timer.x86.c']
args += ['-D__MPI']
if spec.satisfies('+opt'):
args += ['-D__PREFETCH_NEXT_PLANE_FROM_DRAM']
args += ['-D__FUSION_RESIDUAL_RESTRICTION']
else:
args += ['-D__COLLABORATIVE_THREADING=6']
args += ['-D__TEST_MG_CONVERGENCE', '-D__PRINT_NORM', '-D__USE_BICGSTAB']
args += ['-o', 'run.miniGMG', '-lm']
cc(*args)
mkdir(prefix.bin)
install('run.miniGMG', prefix.bin)
mkdir(prefix.jobs)
install('job*', prefix.jobs)
|
