blob: 0a70d7c5a5e3f5b932beec5e72f5ae3822bfb802 (
plain) (
blame)
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
|
# Copyright 2013-2023 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 Nvcomp(CMakePackage, CudaPackage):
"""Last open source release of A library for fast lossless compression
/decompression on NVIDIA GPUs
forked from: https://github.com/NVIDIA/nvcomp after NVIDIA made this closed source
"""
homepage = "https://github.com/NVIDIA/nvcomp"
url = "https://github.com/NVIDIA/nvcomp/archive/refs/tags/v2.0.2.tar.gz"
# pinned to the last open source release+a few minor patches
git = "https://github.com/robertu94/nvcomp"
maintainers("robertu94")
license("BSD-3-Clause")
version("2.2.0", commit="3737f6e5028ed1887b0023ad0fc033e139d57574")
version("2.0.2", commit="5d5c194f3449486d989057f632d10954b8d11d75")
depends_on("cuda")
conflicts("~cuda")
def cmake_args(self):
args = ["-DBUILD_EXAMPLES=OFF", "-DBUILD_BENCHMARKS=OFF"]
cuda_arch_list = self.spec.variants["cuda_arch"].value
args.append("CMAKE_CUDA_ARCHITECTURES={0}".format(";".join(cuda_arch_list)))
return args
|
