Package Details: dlib-cuda 19.24.3-1

Git Clone URL: (read-only, click to copy)
Package Base: dlib-cuda
Description: A general purpose cross-platform C++ library designed using contract programming and modern C++ techniques
Upstream URL:
Keywords: c++ cuda deep learning machine
Licenses: custom
Conflicts: dlib
Provides: dlib
Submitter: xantares
Maintainer: swiftscythe (pingplug)
Last Packager: swiftscythe
Votes: 3
Popularity: 0.000004
First Submitted: 2021-01-18 14:19 (UTC)
Last Updated: 2024-03-10 23:30 (UTC)

Required by (14)

Sources (1)

Latest Comments

thales.barretto commented on 2024-05-28 01:58 (UTC) (edited on 2024-05-28 02:06 (UTC) by thales.barretto)

@AchmadFathoni indeed, using gcc11 works *(see NOTE).

But beyond adding -DCUDA_HOST_COMPILER=/usr/bin/gcc-11 also avoid doing LTO, in the PKGBUILD include the line options=( '!lto' ).

*NOTE:Now we have gcc14 in arch wich cannot compile gcc11. So this can get tricky: 1. with gcc14, build gcc13. 2. Use gcc13 to build gcc11 (GOTCHA: patch the PKGBUILD to use gcc13, see aur comments) 3. Use gcc11 to build dlib-cuda.

girlbossceo commented on 2023-12-05 02:06 (UTC)

Will this package be updated? Package has been marked out of date for some months now. 19.24.2 has been out since May. This PKGBUILD should just do what the dlib AUR package does and get the tarball from the GitHub which is 19.24.2:

Additionally, this package isn't erroring when the CUDA test compilation fails which results in building dlib without CUDA

-- Disabling CUDA support for dlib. DLIB WILL NOT USE CUDA

Can these be fixed please?

AchmadFathoni commented on 2022-09-30 12:26 (UTC) (edited on 2022-09-30 12:26 (UTC) by AchmadFathoni)

Add gcc11 as makedepend & add -DCUDA_HOST_COMPILER=/usr/bin/gcc-11 should fix this

AchmadFathoni commented on 2022-09-30 12:22 (UTC)

-- Building a CUDA test project to see if your compiler is compatible with CUDA... -- **************** -- CUDA was found but your compiler failed to compile a simple CUDA program so dlib isn't going to use CUDA. -- The output of the failed CUDA test compile is shown below: -- -- * Change Dir: /home/toni/.cache/yay/dlib-cuda/src/build/dlib/cuda_test_build

Run Build Command(s):/usr/bin/ninja && [1/2] Building NVCC (Device) object CMakeFiles/cuda_test.dir/ FAILED: CMakeFiles/cuda_test.dir/ /home/toni/.cache/yay/dlib-cuda/src/build/dlib/cuda_test_build/CMakeFiles/cuda_test.dir/ cd /home/toni/.cache/yay/dlib-cuda/src/build/dlib/cuda_test_build/CMakeFiles/cuda_test.dir && /usr/bin/cmake -E make_directory /home/toni/.cache/yay/dlib-cuda/src/build/dlib/cuda_test_build/CMakeFiles/cuda_test.dir//. && /usr/bin/cmake -D verbose:BOOL=OFF -D build_configuration:STRING= -D generated_file:STRING=/home/toni/.cache/yay/dlib-cuda/src/build/dlib/cuda_test_build/CMakeFiles/cuda_test.dir//./ -D generated_cubin_file:STRING=/home/toni/.cache/yay/dlib-cuda/src/build/dlib/cuda_test_build/CMakeFiles/cuda_test.dir//./ -P /home/toni/.cache/yay/dlib-cuda/src/build/dlib/cuda_test_build/CMakeFiles/cuda_test.dir// /usr/include/c++/12.2.0/bits/locale_facets_nonio.tcc: In member function ‘_InIter std::__cxx11::time_get<_CharT, _InIter>::get(iter_type, iter_type, std::ios_base&, std::ios_base::iostate&, tm, const char_type, const char_type) const’: /usr/include/c++/12.2.0/bits/locale_facets_nonio.tcc:1477:77: error: invalid type argument of unary ‘’ (have ‘int’) 1477 | if ((void)(this->(&time_get::do_get)) == (void)(&time_get::do_get)) | ^
CMake Error at (message): Error generating file * /home/toni/.cache/yay/dlib-cuda/src/build/dlib/cuda_test_build/CMakeFiles/cuda_test.dir//./

*** ninja: build stopped: subcommand failed.

-- ***************** -- Disabling CUDA support for dlib. DLIB WILL NOT USE CUDA -- C++11 activated. -- Configuring done -- Generating done -- Build files have been written to: /home/toni/.cache/yay/dlib-cuda/src/build [14/63] Building CXX object dlib/CMakeFiles/dlib.dir/logger/logger_config_file.cpp.o^C ninja: build stopped: interrupted by use