Package Details: clfft 2.12.2-2

Git Clone URL: https://aur.archlinux.org/clfft.git (read-only, click to copy)
Package Base: clfft
Description: A software library containing FFT functions written in OpenCL
Upstream URL: https://github.com/clMathLibraries/clFFT
Licenses: Apache
Submitter: jdarch
Maintainer: None
Last Packager: acxz
Votes: 3
Popularity: 0.000000
First Submitted: 2013-11-15 22:13 (UTC)
Last Updated: 2022-07-02 17:04 (UTC)

Required by (1)

Sources (2)

Pinned Comments

Latest Comments

acxz commented on 2022-06-15 14:31 (UTC) (edited on 2022-07-02 16:56 (UTC) by acxz)

I am also not able to build this package and have the same error as @dale6998

Edit: adopted and fixed the package. Please try again @dale6998.

dale6998 commented on 2022-06-01 15:27 (UTC) (edited on 2022-06-01 15:30 (UTC) by dale6998)

this fails to build but clfft-git builds fine. please see the log:

/home/dale/.aur_build/trizen-dale/clfft/src/clFFT-2.12.2/src/library/generator.transpose.cpp: In function ‘clfftStatus clfft_transpose_generator::genSwapKernelGeneral(const FFTGeneratedTransposeNonSquareAction::Signature&, std::string&, std::string&, const size_t&, size_t)’:
/home/dale/.aur_build/trizen-dale/clfft/src/clFFT-2.12.2/src/library/generator.transpose.cpp:1092:95: error: taking address of rvalue [-fpermissive]
 1092 |         std::string smaller_dim_str = static_cast<std::ostringstream*>(&(std::ostringstream() << smaller_dim))->str();
      |                                                                         ~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~
/home/dale/.aur_build/trizen-dale/clfft/src/clFFT-2.12.2/src/library/generator.transpose.cpp:1093:93: error: taking address of rvalue [-fpermissive]
 1093 |         std::string dim_ratio_str = static_cast<std::ostringstream*>(&(std::ostringstream() << dim_ratio))->str();
      |                                                                       ~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~
make[2]: *** [library/CMakeFiles/clFFT.dir/build.make:160: library/CMakeFiles/clFFT.dir/generator.transpose.cpp.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:156: library/CMakeFiles/clFFT.dir/all] Error 2
make: *** [Makefile:156: all] Error 2
==> ERROR: A failure occurred in build().
    Aborting...
:: Unable to build clfft - makepkg exited with code: 4

lulingar commented on 2016-02-15 17:37 (UTC)

I have built the latest version as of Feb 2016, with the following changes to PKGBUILD: pkgver=2.10.0 md5sums=('f6f8605a4885dcd708b1fa24348e1803') sha512sums=('45f023bfdaf2294dc47704371dda2531aedde49977e7a9b810998a3be972fae70c44fb0ad6285e247fb6a428e00c056ab1698443d339edebdc84a37f091e528d')

jdarch commented on 2015-04-02 21:33 (UTC)

Thanks, I have just made a small modification to your script, it appears to build fine, have not tested the functionality of the resulting lib though.

rubenvb commented on 2015-03-31 07:19 (UTC)

I changed the header install directory to match the binary releases (without the clFFT/ subdirectory). Here is an updated PKGBUILD for version 2.4: pkgname=clfft pkgver=2.4 pkgrel=1 pkgdesc="A software library containing FFT functions written in OpenCL" arch=("i686" "x86_64") url="https://github.com/clMathLibraries/clFFT" license=('APACHE') depends=("libcl") makedepends=("opencl-headers") source=("https://github.com/clMathLibraries/clFFT/archive/v${pkgver}.tar.gz") md5sums=('32f1c5e18d0827550a96ac5f0d369120') sha512sums=('3450f3ee962e1722c765e962178c57df0bf298508603ac5b4ec12bd6396dc10279136d5b91383b9da56d95cebd0e552bf78e2c01ccaeb8cd43ed947b7d536209') if [ "$CARCH" == "x86_64" ]; then _bits=64 elif [ "$CARCH" == "i686" ]; then _bits=32 fi build() { export DESTDIR="${pkgdir}" cd "${srcdir}/clFFT-${pkgver}/src" cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/usr ./ make clFFT } package() { cd "${srcdir}/clFFT-${pkgver}/src" make install rm -rf "${pkgdir}/usr/bin${_bits}" mv "${pkgdir}/usr/lib${_bits}" "${pkgdir}/usr/lib" rm "${pkgdir}/usr/lib/libStatTimer.so" }