Cufftdx compilation fails on windows - "FFT description must be complete to calculate queried information"

[Ido-Work]

Hi,
trying to compile the examples of cufftdx yeilds the following:

PS C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\example\cufftdx> cmake .
-- Building for: Visual Studio 17 2022
-- Selecting Windows SDK version 10.0.22621.0 to target Windows 10.0.26100.
-- The CXX compiler identification is MSVC 19.43.34810.0
-- The CUDA compiler identification is NVIDIA 12.8.61 with host compiler MSVC 19.43.34810.0
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: C:/Program Files/Microsoft Visual Studio/2022/Community/VC/Tools/MSVC/14.43.34808/bin/Hostx64/x64/cl.exe - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Check for working CUDA compiler: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.8/bin/nvcc.exe - skipped
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- Using mathDx package to find cuFFTDx
-- mathdx: Found CUTLASS (NvidiaCutlass) dependency: C:/Users/IdoLev/Desktop/nvidia-mathdx-25.06.1/nvidia/mathdx/25.06/external/cutlass/include
-- cufftdx: Found CUTLASS (NvidiaCutlass) dependency: C:/Users/IdoLev/Desktop/nvidia-mathdx-25.06.1/nvidia/mathdx/25.06/external/cutlass/include
-- cufftdx: Found commondx dependency
-- Found cufftdx: (Version: 1.5.1, Include dirs: C:/Users/IdoLev/Desktop/nvidia-mathdx-25.06.1/nvidia/mathdx/25.06/include)
-- mathDx: cuFFTDx found: C:/Users/IdoLev/Desktop/nvidia-mathdx-25.06.1/nvidia/mathdx/25.06/include
-- Targeted cuFFTDx CUDA Architectures: 80-real
-- Found CUDAToolkit: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.8/include (found version "12.8.61")
-- Configuring done (33.0s)
-- Generating done (1.0s)
-- Build files have been written to: C:/Users/IdoLev/Desktop/nvidia-mathdx-25.06.1/nvidia/mathdx/25.06/example/cufftdx

when opening in visual studio and compiling: 00_introduction_example_introduction_example
i get the following error:

2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\example\cufftdx>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\bin\nvcc.exe"  --use-local-env -ccbin "C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.43.34808\bin\HostX64\x64" -x cu   -I"C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include" -I"C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\external\cutlass\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\include"     --keep-dir 00_intro.26E594B1\x64\Release  -maxrregcount=0    --machine 64 --compile -cudart static -Xfatbin -compress-all -Xcudafe --display_error_number -std=c++20 --generate-code=arch=compute_80,code=[sm_80] -Xfatbin -compress-all -Xcompiler="/EHsc /WX /Zc:__cplusplus -Ob2"   -D_WINDOWS -DNDEBUG -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_WARNINGS -DNOMINMAX -DCUFFTDX_EXAMPLE_CMAKE -DCUFFTDX_EXAMPLE_ENABLE_SM_80 -D"CMAKE_INTDIR=\"Release\"" -D_MBCS -D"CMAKE_INTDIR=\"Release\"" -Xcompiler "/EHsc /W3 /nologo /O2 /FS   /MD " -Xcompiler "/Fd00_introduction_example_introduction_example.dir\Release\vc143.pdb" -o 00_introduction_example_introduction_example.dir\Release\introduction_example.obj "C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\example\cufftdx\00_introduction_example\introduction_example.cu"
2>introduction_example.cu
2>tmpxft_0000fbd0_00000000-7_introduction_example.cudafe1.cpp
2>C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.43.34808\include\type_traits(1278): error C2139: 'cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>': an undefined class is not allowed as an argument to compiler intrinsic type trait '__is_base_of'
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\cufftdx/detail/fft_execution.hpp(358): note: see declaration of 'cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>'
2>C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.43.34808\include\type_traits(1278): note: the template instantiation context (the oldest one first) is
2>C:\Users\IdoLev\AppData\Local\Temp\tmpxft_0000fbd0_00000000-7_introduction_example.cudafe1.stub.c(7): note: see reference to class template instantiation 'cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>' being compiled
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\cufftdx/detail/fft_execution.hpp(940): note: while compiling class template member function 'unsigned int cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>::get_storage_size(void)'
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\cufftdx/detail/fft_execution.hpp(973): note: see the first reference to 'cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>::get_storage_size'
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\cufftdx/detail/fft_execution.hpp(942): note: see reference to alias template instantiation 'cufftdx::detail::check_and_get_fft_implementation_t<cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>>' being compiled
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\cufftdx/traits/detail/check_and_get_trait.hpp(366): note: see reference to class template instantiation 'cufftdx::detail::check_and_get_fft_implementation<cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>>' being compiled
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\cufftdx/traits/detail/check_and_get_trait.hpp(358): note: see reference to class template instantiation 'cufftdx::detail::get_block_config_impl::helper<Description,false>' being compiled
2>        with
2>        [
2>            Description=cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>
2>        ]
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\cufftdx/traits/detail/check_and_get_trait.hpp(214): note: see reference to class template instantiation 'cufftdx::detail::is_complete_description<Description>' being compiled
2>        with
2>        [
2>            Description=cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>
2>        ]
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\cufftdx/traits/detail/description_traits.hpp(132): note: see reference to class template instantiation 'commondx::detail::is_description_expression<cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>>' being compiled
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\commondx/traits/detail/description_traits.hpp(45): note: see reference to class template instantiation 'std::is_base_of<commondx::detail::description_expression,T>' being compiled
2>        with
2>        [
2>            T=cufftdx::detail::fft_block_execution<cufftdx::Size<128>,cufftdx::Precision<float>,cufftdx::Type<cufftdx::fft_type::c2c>,cufftdx::Direction<cufftdx::fft_direction::forward>,cufftdx::ElementsPerThread<8>,cufftdx::FFTsPerBlock<2>,commondx::SM<800,void>,cufftdx::Block>
2>        ]
2>C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include\cufftdx/traits/detail/check_and_get_trait.hpp(215): error C2338: static_assert failed: 'FFT description must be complete to calculate queried information'
2>C:\Program Files\Microsoft Visual Studio\2022\Community\MSBuild\Microsoft\VC\v170\BuildCustomizations\CUDA 12.8.targets(800,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\bin\nvcc.exe"  --use-local-env -ccbin "C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.43.34808\bin\HostX64\x64" -x cu   -I"C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\include" -I"C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\external\cutlass\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\include"     --keep-dir 00_intro.26E594B1\x64\Release  -maxrregcount=0    --machine 64 --compile -cudart static -Xfatbin -compress-all -Xcudafe --display_error_number -std=c++20 --generate-code=arch=compute_80,code=[sm_80] -Xfatbin -compress-all -Xcompiler="/EHsc /WX /Zc:__cplusplus -Ob2"   -D_WINDOWS -DNDEBUG -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_WARNINGS -DNOMINMAX -DCUFFTDX_EXAMPLE_CMAKE -DCUFFTDX_EXAMPLE_ENABLE_SM_80 -D"CMAKE_INTDIR=\"Release\"" -D_MBCS -D"CMAKE_INTDIR=\"Release\"" -Xcompiler "/EHsc /W3 /nologo /O2 /FS   /MD " -Xcompiler "/Fd00_introduction_example_introduction_example.dir\Release\vc143.pdb" -o 00_introduction_example_introduction_example.dir\Release\introduction_example.obj "C:\Users\IdoLev\Desktop\nvidia-mathdx-25.06.1\nvidia\mathdx\25.06\example\cufftdx\00_introduction_example\introduction_example.cu"" exited with code 2.
2>Done building project "00_introduction_example_introduction_example.vcxproj" -- FAILED.
========== Build: 1 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========
========== Build completed at 22:54 and took 59.567 seconds ==========

[mferreravila]

hi @Ido-Work,

Windows support is not yet available (coming in a future release), but in the meantime we will take a look to see if there is a possible work-around for the issue you shared. ETA is next week.

Just to be clear, the code in the sample has not been modifyied in any way, correct? (equivalently, the same sample compiles without issues on linux)