Good day everyone,
I refer to compiling a DLL for the multiview deconvolution of SPIM data in Fiji on a windows PC, as described here: ⚡ CUDA Deconvolution
The first command seems to run perfectly: cmake.exe -G “Visual Studio 16 2019” -DCMAKE_INSTALL_PREFIX=“C:/Users/SPIM data Processor/Desktop/compiled-cuda-fft-lib” …
However, the second command I ran seems to return an output that the GPU architecture is not supported
cmake.exe --build . --target ALL_BUILD --config Release
Returns an error:
nvcc fatal : Unsupported gpu architecture ‘compute_30’
If someone can please help confirm if this is the case, please. I have attached a detailed report
@schmiedc @StephanPreibisch @hoerldavid @haesleinhuepf
Sorry to bother you all.
GPU_Code-Gideon.txt (7.0 KB)