Compilation issue with CUDA 12.6

[smbct]

When compiling with CUDA version 12.6, the linker produces multiples errors (tested on Ubuntu 24.04). The issue seems to be linked to a change with nvcc in newer versions, see this stackoverflow post : https://stackoverflow.com/questions/79148211/cuda-thrust-object-shared-between-cu-and-cpp-files (when linking object files from g++ and nvcc).

The workaround mentioned in the post allow to compile but a cleaner fix would probably be preferable.

[ 21%] Linking CXX executable bdd_solver_cl
/usr/bin/ld : libbdd_solver.a(bdd_solver.cpp.o) : in function « LPMP::bdd_solver::construct_solver(nlohmann::json_abi_v3_11_3::basic_json<std::map, std::vector, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, bool, long, unsigned long, double, std::allocator, nlohmann::json_abi_v3_11_3::adl_serializer, std::vector<unsigned char, std::allocator<unsigned char> >, void> const&, BDD::bdd_collection const&, std::vector<double, std::allocator<double> > const&) » :
bdd_solver.cpp:(.text+0x17db) : undefined reference to « LPMP::lbfgs<LPMP::bdd_cuda_parallel_mma<double>, thrust::THRUST_200500___CUDA_ARCH_LIST___NS::device_vector<double, thrust::THRUST_200500___CUDA_ARCH_LIST___NS::device_allocator<double> >, double, thrust::THRUST_200500___CUDA_ARCH_LIST___NS::device_vector<char, thrust::THRUST_200500___CUDA_ARCH_LIST___NS::device_allocator<char> >, true>::lbfgs(BDD::bdd_collection const&, std::vector<double, std::allocator<double> > const&, int, double, double, double, double) »