I would like to add CUDA to my JUCE projects. The main steps of adding CUDA to an existing project as I understand are:
- The
.cufiles are supposed to contain the CUDA kernel code. - We are supposed to right click these files in Visual Studio > Properties > General > Item Type > Select Cuda C/C++ from the list.
- We are supposed to also add the CUDA build steps by right clicking the project (_SharedCode I presume) > Build Dependencies > Build Customizations > Select CUDA from the list.
That is all that is suggested is supposed to be needed to build.
This is based on the explanation here:
What I understand is then supposed to happen is nvcc (Nvidia compiler) is supposed to compile the .cu CUDA files as a first step, and then the rest of the program builds after that.
Old instructions to add CUDA to existing projects, as best I could research them before I found this more recent guide are here. However, that all seems obsolete though perhaps helpful for understanding what is happening.
I believe all of this centers around needing to have some custom build steps or rules to make nvcc process the .cu files first before the rest of the build proceeds.
I have tried adding some basic CUDA test scripts to a JUCE project. I marked the .cu files as binary in Projucer (correct?). I added the CUDA header folder to the Header Search Paths (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include).
Then I opened in Visual Studio and made the two fixes above.
However, when I try to build I get:
|Severity|Code|Description|Project|File|Line|Suppression State|Details|
|—|—|—|—|—|—|—|—|
|Error|MSB3721|The command C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\bin\nvcc.exe -gencode=arch=compute_52,code=\sm_52,compute_52\ –use-local-env -ccbin C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.41.34120\bin\HostX64\x64 -x cu -I…..\JuceLibraryCode -ID:\JUCE\app\modules -IC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include -ID:/JUCE/ASIO/common -IC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include -IC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include -G --keep-dir Synth.C18B78B3\x64\Debug -maxrregcount=0 --machine 64 --compile -cudart static -g -D_CRT_SECURE_NO_WARNINGS -DWIN32 -D_WINDOWS -DDEBUG -D_DEBUG -DJUCER_VS2022_78A503E=1 -DJUCE_APP_VERSION=1.0.0 -DJUCE_APP_VERSION_HEX=0x10000 -DJucePlugin_Build_VST=0 -DJucePlugin_Build_VST3=0 -DJucePlugin_Build_AU=0 -DJucePlugin_Build_AUv3=0 -DJucePlugin_Build_AAX=0 -DJucePlugin_Build_Standalone=1 -DJucePlugin_Build_Unity=0 -DJucePlugin_Build_LV2=0 -DJUCE_SHARED_CODE=1 -D_LIB -Xcompiler /EHsc /W4 /nologo /Od /FS /Zi /MDd /GR -Xcompiler /Fd\x64\Debug\Shared Code\Synth.pdb\ -o D:\JUCE\PROJECTS\Synth-Private\Builds\VisualStudio2022\Synth.C18B78B3\x64\Debug\CudaKernel.cu.obj D:\JUCE\PROJECTS\Synth-Private\Source\CUDA\CudaKernel.cu exited with code 2.|Synth_SharedCode|C:\Program Files\Microsoft Visual Studio\2022\Community\MSBuild\Microsoft\VC\v170\BuildCustomizations\CUDA 12.6.targets|799|||
|Error|C1189|#error: JUCE requires C++17 or later|Synth_SharedCode|D:\JUCE\app\modules\juce_core\system\juce_CompilerSupport.h|124|||
Any ideas? Thanks