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
|
# 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.package import *
class Spla(CMakePackage):
"""Specialized Parallel Linear Algebra, providing distributed GEMM
functionality for specific matrix distributions with optional GPU
acceleration."""
homepage = "https://github.com/eth-cscs/spla"
url = "https://github.com/eth-cscs/spla/archive/v1.0.0.tar.gz"
git = 'https://github.com/eth-cscs/spla.git'
maintainers = ['AdhocMan', 'haampie']
version('1.5.4', sha256='de30e427d24c741e2e4fcae3d7668162056ac2574afed6522c0bb49d6f1d0f79')
version('1.5.3', sha256='527c06e316ce46ec87309a16bfa4138b1abad23fd276fe789c78a2de84f05637')
version('1.5.2', sha256='344c34986dfae182ec2e1eb539c9a57f75683aaa7a61a024fd0c594d81d97016')
version('1.5.1', sha256='2021a30b7cbb10bd660e5d94e1cc7bc6a428c87ea507e09d1e57e455685da421')
version('1.5.0', sha256='bea782d46ce615e1c40efc2bfb19d95e3b59f332fc9ca83ac7e6684b8ac2dd93')
version('1.4.0', sha256='364a9fe759fddec8a0839cf79f1cf0619fc36f4d4c15f1c2b1f437249d7840c6')
version('1.3.0', sha256='ff05a22bd655607ff941f3228ac8605a813e1eec6eaa49fbcf7b58a3a4cf5f00')
version('1.2.1', sha256='4d7237f752dc6257778c84ee19c9635072b1cb8ce8d9ab6e34a047f63a736b29')
version('1.2.0', sha256='96ddd13c155ef3d7e40f87a982cdb439cf9e720523e66b6d20125d346ffe8fca')
version('1.1.1', sha256='907c374d9c53b21b9f67ce648e7b2b09c320db234a1013d3f05919cd93c95a4b')
version('1.1.0', sha256='b0c4ebe4988abc2b3434e6c50e7eb0612f3f401bc1aa79ad58a6a92dc87fa65b')
version('1.0.0', sha256='a0eb269b84d7525b223dc650de12170bba30fbb3ae4f93eb2b5cbdce335e4da1')
version('develop', branch='develop')
version('master', branch='master')
variant('openmp', default=True, description="Build with OpenMP support")
variant('static', default=False, description="Build as static library")
variant('cuda', default=False, description="CUDA backend")
variant('rocm', default=False, description="ROCm backend")
variant('fortran', default=False, description="Build fortran module")
conflicts("+cuda", when="+rocm", msg="+cuda and +rocm are mutually exclusive")
depends_on('mpi')
depends_on('blas')
depends_on('cmake@3.10:', type='build')
depends_on('cuda', when='+cuda')
depends_on('rocblas', when='+rocm')
depends_on('hip', when='+rocm')
# Propagate openmp to blas
depends_on('openblas threads=openmp', when='+openmp ^openblas')
depends_on('amdblis threads=openmp', when='+openmp ^amdblis')
depends_on('blis threads=openmp', when='+openmp ^blis')
depends_on('intel-mkl threads=openmp', when='+openmp ^intel-mkl')
# Fix CMake find module for AMD BLIS,
# which uses a different library name for the multi-threaded version
patch('0001-amd_blis.patch', when='@1.3.0:1.4.0 ^amdblis')
def cmake_args(self):
args = [
self.define_from_variant('SPLA_OMP', 'openmp'),
self.define_from_variant('SPLA_FORTRAN', 'fortran'),
self.define_from_variant('SPLA_STATIC', 'static')
]
if '+cuda' in self.spec:
args += ["-DSPLA_GPU_BACKEND=CUDA"]
elif '+rocm' in self.spec:
args += ["-DSPLA_GPU_BACKEND=ROCM"]
else:
args += ["-DSPLA_GPU_BACKEND=OFF"]
if self.spec['blas'].name == 'openblas':
args += ["-DSPLA_HOST_BLAS=OPENBLAS"]
elif self.spec['blas'].name in ['amdblis', 'blis']:
args += ["-DSPLA_HOST_BLAS=BLIS"]
elif self.spec['blas'].name == 'atlas':
args += ["-DSPLA_HOST_BLAS=ATLAS"]
elif self.spec['blas'].name == 'intel-mkl':
args += ["-DSPLA_HOST_BLAS=MKL"]
elif self.spec['blas'].name == 'netlib-lapack':
args += ["-DSPLA_HOST_BLAS=GENERIC"]
elif self.spec['blas'].name == 'cray-libsci':
args += ["-DSPLA_HOST_BLAS=CRAY_LIBSCI"]
return args
|
