I am trying to build the samples shipped with CUDAToolkit 10.
On macOS, these samples are installed here:
<U+2068>/Developer<U+2069>/NVIDIA<U+2069>/CUDA-10.0<U+2069>/samples<U+2069>
Modified MakeFile in 0_Simple/vectorAdd, line# 251, to read:
# Gencode arguments
SMS ?= 30 35 37 50 52 60 61 70 75
Then, compilation of 0_Simple/vectorAdd fails with an error:
nvcc fatal : Unsupported gpu architecture 'compute_75'
make: *** [vectorAdd.o] Error 1
nvcc version:
$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Sat_Aug_25_21:08:56_CDT_2018
Cuda compilation tools, release 10.0, V10.0.130
According to:
https://docs.nvidia.com/cuda/turing-compatibility-guide/index.html#building-turing-compatible-apps-using-cuda-10-0
I didn’t expect this error.
This is on macOS 10.14. According to the following post, I didn’t expect a successful build. Though, I didn’t expect the posted failure message either.
https://devtalk.nvidia.com/default/topic/1042279/cuda-setup-and-installation/cuda-10-and-macos-10-14/