Compiling Error Ubuntu RTX 5080: nvcc fatal : Unsupported gpu architecture 'compute_120'

[jeffklassen]

I'm trying to build with nvidia support:

> cmake -B build -DGGML_CUDA=1
-- Warning: ccache not found - consider installing it for faster compilation or disable this warning with GGML_CCACHE=OFF
-- CMAKE_SYSTEM_PROCESSOR: x86_64
-- Including CPU backend
-- x86 detected
-- Adding CPU backend variant ggml-cpu: -march=native
-- CUDA Toolkit found
-- Using CUDA architectures: native
-- CUDA host compiler is GNU 12.3.0

-- Including CUDA backend
-- Configuring done (0.3s)
-- Generating done (0.0s)
-- Build files have been written to: /home/jeff/Documents/workspace/whisper.cpp/build

and then:

>cmake --build build -j --config Release
...
nvcc fatal   : Unsupported gpu architecture 'compute_120'
gmake[2]: *** [ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/build.make:752: ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_4.cu.o] Error 1
nvcc fatal   : Unsupported gpu architecture 'compute_120'
gmake[2]: *** [ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/build.make:1067: ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/template-instances/mmq-instance-q3_k.cu.o] Error 1
nvcc fatal   : Unsupported gpu architecture 'compute_120'
gmake[2]: *** [ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/build.make:767: ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_8.cu.o] Error 1
nvcc fatal   : Unsupported gpu architecture 'compute_120'
gmake[2]: *** [ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/build.make:842: ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_8.cu.o] Error 1
nvcc fatal   : Unsupported gpu architecture 'compute_120'
gmake[2]: *** [ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/build.make:722: ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_2.cu.o] Error 1
nvcc fatal   : Unsupported gpu architecture 'compute_120'
gmake[2]: *** [ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/build.make:1292: ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/template-instances/fattn-vec-f16-instance-hs64-f16-f16.cu.o] Error 1
gmake[1]: *** [CMakeFiles/Makefile2:1133: ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/all] Error 2
gmake: *** [Makefile:146: all] Error 2

[jeffklassen]

Note: I got this to work with

cmake -B build -DGGML_CUDA=1 -DCMAKE_CUDA_ARCHITECTURES="86"
cmake --build build -j --config Release

This issue will be closed by merging: #3031