CudaAllocator报错

环境 VS2022 cuda12.2

代码为08课04节
https://github.com/parallel101/course/blob/2d30da61b442008c003f69225e6feca20a4ca7df/08/04_sugar/01/main.cu#L7

报错信息

[build] MSBuild version 17.4.0+18d5aef85 for .NET Framework
[build]   Compiling CUDA source file ..\src\allocator.cu...
[build]   
[build]   C:\Dev\mgxpbd\build>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.2\bin\nvcc.exe"  --use-local-env -ccbin "C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.34.31933\bin\HostX64\x64" -x cu   -IC:\Dev\mgxpbd\include -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.2\include"     --keep-dir x64\Debug  -maxrregcount=0   --machine 64 --compile -cudart static --generate-code=arch=compute_52,code=[compute_52,sm_52] -std=c++17 -Xcompiler="/EHsc -Ob0 -Zi" -g  -D_WINDOWS -D"CMAKE_INTDIR=\"Debug\"" -D_MBCS -D"CMAKE_INTDIR=\"Debug\"" -Xcompiler "/EHsc /W1 /nologo /Od /FS /Zi /RTC1 /MDd " -Xcompiler "/Fdallocator.dir\Debug\vc143.pdb" -o allocator.dir\Debug\allocator.obj "C:\Dev\mgxpbd\src\allocator.cu" 
[build] C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.34.31933\include\vector(2125): error : no suitable user-defined conversion from "CudaAllocator<int>" to "CudaAllocator<std::_Container_proxy>" exists [C:\Dev\mgxpbd\build\allocator.vcxproj]
[build]             auto&& _Alproxy = static_cast<_Rebind_alloc_t<_Alty, _Container_proxy>>(_Al);
[build]                                                                                     ^
[build]             detected during:
[build]               instantiation of "void std::vector<_Ty, _Alloc>::_Construct_n(std::vector<_Ty, _Alloc>::size_type, _Valty &&...) [with _Ty=int, _Alloc=CudaAllocator<int>, _Valty=<>]" at line 683
[build]               instantiation of "std::vector<_Ty, _Alloc>::vector(std::vector<_Ty, _Alloc>::size_type, const _Alloc &) [with _Ty=int, _Alloc=CudaAllocator<int>]" at line 30 of C:\Dev\mgxpbd\src\allocator.cu
[build]   
[build] C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.34.31933\include\vector(832): error : no suitable user-defined conversion from "CudaAllocator<int>" to "CudaAllocator<std::_Container_proxy>" exists [C:\Dev\mgxpbd\build\allocator.vcxproj]
[build]             auto&& _Alproxy = static_cast<_Rebind_alloc_t<_Alty, _Container_proxy>>(_Getal());
[build]                                                                                     ^
[build]             detected during instantiation of "std::vector<_Ty, _Alloc>::~vector() noexcept [with _Ty=int, _Alloc=CudaAllocator<int>]" at line 30 of C:\Dev\mgxpbd\src\allocator.cu
[build]   
[build] C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.34.31933\include\vector(833): error : no instance of function template "std::_Delete_plain_internal" matches the argument list [C:\Dev\mgxpbd\build\allocator.vcxproj]
[build]               argument types are: (<error-type>, std::_Container_proxy *)
[build]             _Delete_plain_internal(_Alproxy, ::std:: exchange(_Mypair._Myval2._Myproxy, nullptr));
[build]             ^
[build]             detected during instantiation of "std::vector<_Ty, _Alloc>::~vector() noexcept [with _Ty=int, _Alloc=CudaAllocator<int>]" at line 30 of C:\Dev\mgxpbd\src\allocator.cu
[build]   
[build]   3 errors detected in the compilation of "C:/Dev/mgxpbd/src/allocator.cu".
[build]   allocator.cu
[build] C:\Program Files\Microsoft Visual Studio\2022\Community\MSBuild\Microsoft\VC\v170\BuildCustomizations\CUDA 12.2.targets(799,9): error MSB3721: 命令“"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.2\bin\nvcc.exe"  --use-local-env -ccbin "C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.34.31933\bin\HostX64\x64" -x cu   -IC:\Dev\mgxpbd\include 

Start with the referenced lesson code at 08/04_sugar/01/main.cu#L7 and the allocator instantiation reported at src/allocator.cu line 30. Reproduce the build with Visual Studio 2022 and CUDA 12.2, then compare the CudaAllocator usage with the compiler diagnostics in vector. Done means the lesson code builds successfully without the reported allocator conversion errors.