Install this version:
emerge -a =sci-libs/spqr-4.3.6-r1
If this version is masked, you can unmask it using the autounmask tool or standard emerge options:
autounmask =sci-libs/spqr-4.3.6-r1
Or alternatively:
emerge --autounmask-write -a =sci-libs/spqr-4.3.6-r1
| Version | EAPI | Keywords | Slot |
|---|---|---|---|
| 4.3.6-r1 | 8 | ~alpha ~amd64 ~arm ~arm64 ~hppa ~loong ~mips ~ppc ~ppc64 ~riscv ~sparc ~x86 | 0/$(ver_cut 1) |
# Copyright 1999-2026 Gentoo Authors
# Distributed under the terms of the GNU General Public License v2
EAPI=8
inherit cmake cuda toolchain-funcs
Sparse_PV="7.12.2"
Sparse_P="SuiteSparse-${Sparse_PV}"
DESCRIPTION="Multithreaded multifrontal sparse QR factorization library"
HOMEPAGE="https://people.engr.tamu.edu/davis/suitesparse.html"
SRC_URI="https://github.com/DrTimothyAldenDavis/SuiteSparse/archive/refs/tags/v${Sparse_PV}.tar.gz -> ${Sparse_P}.gh.tar.gz"
S="${WORKDIR}/${Sparse_P}/${PN^^}"
LICENSE="GPL-2+"
SLOT="0/$(ver_cut 1)"
KEYWORDS="~alpha ~amd64 ~arm ~arm64 ~hppa ~loong ~mips ~ppc ~ppc64 ~riscv ~sparc ~x86"
IUSE="cuda doc test"
RESTRICT="!test? ( test )"
DEPEND="
>=sci-libs/suitesparseconfig-${Sparse_PV}:=
>=sci-libs/cholmod-5.3.4:=
virtual/blas
cuda? (
dev-util/nvidia-cuda-toolkit:=
x11-drivers/nvidia-drivers
)
"
RDEPEND="${DEPEND}"
BDEPEND="
doc? (
virtual/latex-base
dev-texlive/texlive-plaingeneric
)
"
src_prepare() {
cmake_src_prepare
if use cuda; then
cuda_src_prepare
fi
}
src_configure() {
# Define SUITESPARSE_INCLUDEDIR_POSTFIX to "" otherwise it take
# the value suitesparse, and the include directory would be set to
# /usr/include/suitesparse
# This need to be set in all suitesparse ebuilds.
local mycmakeargs=(
-DBUILD_STATIC_LIBS=OFF
-DSPQR_USE_CUDA=$(usex cuda)
-DSUITESPARSE_DEMOS=$(usex test)
-DSUITESPARSE_INCLUDEDIR_POSTFIX=""
)
if has_version 'virtual/blas[index64]'; then
mycmakeargs+=( -DSUITESPARSE_USE_64BIT_BLAS=ON )
fi
if has_version 'virtual/blas[flexiblas]'; then
mycmakeargs+=( -DBLA_VENDOR=FlexiBLAS )
else
mycmakeargs+=( -DBLA_VENDOR=Generic )
fi
# TODO: Figure out how to make sci-libs/mkl work. Bug 974246
if use cuda; then
cuda_add_sandbox
addpredict "/dev/char/"
mycmakeargs+=(
-DSUITESPARSE_CUDA_ARCHITECTURES="${CUDAARCHS:-all-major}"
)
local -x CUDAHOSTCXX="$(cuda_gccdir)"
local -x CUDAHOSTLD="$(tc-getCXX)"
fi
cmake_src_configure
}
src_test() {
# Because we are not using cmake_src_test,
# we have to manually go to BUILD_DIR
cd "${BUILD_DIR}" || die
local simple=(
ash219
west0067
)
local demo=(
a04
a0
a1
a2
a4
ash219
b1_ss
bcspwr01
bfwa62
c2
c32
GD01_b
GD06_theory
GD98_a
lfat5b
LFAT5
lp_e226_transposed
lpi_galenet
lpi_itest6
lp_share1b
r2
Ragusa16
s32
Tina_AskCal
Tina_AskCal_perm
west0067
young1c
)
local simples=(
qrsimple
qrsimplec
qrsimplec_int32
)
local demos=(
qrdemo
qrdemo_int32
qrdemoc
qrdemoc_int32
)
declare -A testsuite
local i
local j
for i in "${simples[@]}"; do
testsuite["${i}"]=${simple[@]}
done
for i in "${demos[@]}"; do
testsuite["${i}"]=${demo[@]}
done
# Run demo files
for i in ${!testsuite[@]}; do
for j in ${testsuite[${i[@]}]}; do
./${i} < "${S}"/Matrix/${j}.mtx || die "failed testing ${i} with ${j}"
done
done
einfo "All tests passed"
}
src_install() {
if use doc; then
pushd "${S}/Doc" || die
emake clean
rm -rf *.pdf || die
emake
popd || die
DOCS="${S}/Doc/*.pdf"
fi
cmake_src_install
}
Manage flags for this package:
euse -i <flag> -p sci-libs/spqr |
euse -E <flag> -p sci-libs/spqr |
euse -D <flag> -p sci-libs/spqr
>=sci-libs/suitesparseconfig-7.12.2:= >=sci-libs/cholmod-5.3.4:= virtual/blas cuda? ( dev-util/nvidia-cuda-toolkit:= x11-drivers/nvidia-drivers )
>=sci-libs/suitesparseconfig-7.12.2:= >=sci-libs/cholmod-5.3.4:= virtual/blas cuda? ( dev-util/nvidia-cuda-toolkit:= x11-drivers/nvidia-drivers )
doc? ( virtual/latex-base dev-texlive/texlive-plaingeneric )
| Type | File | Size | Source URLs |
|---|---|---|---|
| DIST | SuiteSparse-7.12.2.gh.tar.gz | 95337908 bytes | https://github.com/DrTimothyAldenDavis/SuiteSparse/archive/refs/tags/v7.12.2.tar.gz |