cmake -G "MinGW Makefiles" -DGGML_CUBLAS=ON .. error in windows 10

[wuxianjun666]

GNU ld (GNU Binutils) 2.30
-- x86 detected
-- Found CUDAToolkit: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/include (found version "11.6.55")
-- cuBLAS found
-- The CUDA compiler identification is NVIDIA 11.6.55
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - failed
-- Check for working CUDA compiler: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/bin/nvcc.exe
-- Check for working CUDA compiler: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/bin/nvcc.exe - broken
CMake Error at D:/cmake/share/cmake-3.28/Modules/CMakeTestCUDACompiler.cmake:59 (message):
The CUDA compiler

"C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/bin/nvcc.exe"

is not able to compile a simple test program.

It fails with the following output:

Change Dir: 'E:/c++/ggml/build/CMakeFiles/CMakeScratch/TryCompile-69z2j5'

Run Build Command(s): D:/cmake/bin/cmake.exe -E env VERBOSE=1 D:/mingw64/bin/make.exe -f Makefile cmTC_54b8d/fast
D:/mingw64/bin/make.exe  -f CMakeFiles/cmTC_54b8d.dir/build.make CMakeFiles/cmTC_54b8d.dir/build
make.exe[1]: Entering directory 'E:/c++/ggml/build/CMakeFiles/CMakeScratch/TryCompile-69z2j5'
Building CUDA object CMakeFiles/cmTC_54b8d.dir/main.cu.obj
"C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/bin/nvcc.exe" -forward-unknown-to-host-compiler   -D_WINDOWS -Xcompiler=" /GR /EHsc"  -Xcompiler=" -Zi -Ob0 -Od /RTC1" -Xcompiler=-MDd -MD -MT CMakeFiles/cmTC_54b8d.dir/main.cu.obj -MF CMakeFiles/cmTC_54b8d.dir/main.cu.obj.d -x cu -c E:/c++/ggml/build/CMakeFiles/CMakeScratch/TryCompile-69z2j5/main.cu -o CMakeFiles/cmTC_54b8d.dir/main.cu.obj -Xcompiler=-FdCMakeFiles/cmTC_54b8d.dir/,-FS
main.cu
Linking CUDA executable cmTC_54b8d.exe
D:/cmake/bin/cmake.exe -E vs_link_exe --intdir=CMakeFiles/cmTC_54b8d.dir --rc=D:/mingw64/bin/windres.exe --mt="" --manifests -- D:/mingw64/bin/ld.exe /nologo @CMakeFiles/cmTC_54b8d.dir/objects1.rsp  /out:cmTC_54b8d.exe /implib:cmTC_54b8d.lib /pdb:E:/c++/ggml/build/CMakeFiles/CMakeScratch/TryCompile-69z2j5/cmTC_54b8d.pdb /version:0.0 /debug /INCREMENTAL @CMakeFiles/cmTC_54b8d.dir/linkLibs.rsp -LIBPATH:"C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/lib/x64"
Visual Studio Non-Incremental Link
LINK:
D:/mingw64/bin/ld.exe C:/Program Files/Git/nologo @CMakeFiles/cmTC_54b8d.dir/objects1.rsp /out:cmTC_54b8d.exe /implib:cmTC_54b8d.lib /pdb:E:/c++/ggml/build/CMakeFiles/CMakeScratch/TryCompile-69z2j5/cmTC_54b8d.pdb /version:0.0 C:/Program Files/Git/debug C:/Program Files/Git/INCREMENTAL @CMakeFiles/cmTC_54b8d.dir/linkLibs.rsp -LIBPATH:C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/lib/x64 /MANIFEST:EMBED,ID=1
LINK: command "D:/mingw64/bin/ld.exe C:/Program Files/Git/nologo @CMakeFiles/cmTC_54b8d.dir/objects1.rsp /out:cmTC_54b8d.exe /implib:cmTC_54b8d.lib /pdb:E:/c++/ggml/build/CMakeFiles/CMakeScratch/TryCompile-69z2j5/cmTC_54b8d.pdb /version:0.0 C:/Program Files/Git/debug C:/Program Files/Git/INCREMENTAL @CMakeFiles/cmTC_54b8d.dir/linkLibs.rsp -LIBPATH:C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/lib/x64 /MANIFEST:EMBED,ID=1" failed (exit code 1) with the following output:
D:/mingw64/bin/ld.exe: cannot find C:/Program Files/Git/nologo: No such file or directory
D:/mingw64/bin/ld.exe: cannot find /out:cmTC_54b8d.exe: No such file or directory
D:/mingw64/bin/ld.exe: cannot find /implib:cmTC_54b8d.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find /pdb:E:/c++/ggml/build/CMakeFiles/CMakeScratch/TryCompile-69z2j5/cmTC_54b8d.pdb: Invalid argument
D:/mingw64/bin/ld.exe: cannot find /version:0.0: No such file or directory
D:/mingw64/bin/ld.exe: cannot find C:/Program Files/Git/debug: No such file or directory
D:/mingw64/bin/ld.exe: cannot find C:/Program Files/Git/INCREMENTAL: No such file or directory
D:/mingw64/bin/ld.exe: cannot find cudadevrt.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find cudart_static.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find kernel32.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find user32.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find gdi32.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find winspool.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find shell32.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find ole32.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find oleaut32.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find uuid.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find comdlg32.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find advapi32.lib: No such file or directory
D:/mingw64/bin/ld.exe: cannot find /MANIFEST:EMBED,ID=1: No such file or directory
make.exe[1]: *** [CMakeFiles/cmTC_54b8d.dir/build.make:101: cmTC_54b8d.exe] Error 127
make.exe[1]: Leaving directory 'E:/c++/ggml/build/CMakeFiles/CMakeScratch/TryCompile-69z2j5'
make.exe: *** [Makefile:127: cmTC_54b8d/fast] Error 2

CMake will not be able to correctly generate this project.
Call Stack (most recent call first):
src/CMakeLists.txt:212 (enable_language)

-- Configuring incomplete, errors occurred!

[wuxianjun666]

if i use Visual Studio 16 2019 is ok ,when i use cmake -G "MinGW Makefiles" it broken
-- Building for: Visual Studio 16 2019
-- Selecting Windows SDK version 10.0.19041.0 to target Windows 10.0.19045.
-- The C compiler identification is MSVC 19.29.30152.0
-- The CXX compiler identification is MSVC 19.29.30152.0
-- x86 detected
-- Found CUDAToolkit: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/include (found version "11.6.55")
-- cuBLAS found
-- The CUDA compiler identification is NVIDIA 11.6.55
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Check for working CUDA compiler: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/bin/nvcc.exe - skipped
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- GGML CUDA sources found
-- GGML Configuring CUDA architectures 52;61;70

[slaren]

I don't think that CUDA supports mingw.