Cuda unsupported gcc/g++ or incompatible with version 10
Recently, I am trying to develop a GPU supported Python package to maximumly harvest the benefits of GPU fast calculation. To do this type of work, I have tested on the every package that I can find over the internet, mainly focusing on third party packages working like the wrapper for Python-CUDA
programing.
Thanks for those contributors' genius, diligent, and masterpiece works, I have found lots of resources sharing the knowledge on coding (perspective on software) and architecture (perspective on hardware).
Based on my testing, I have made my decision to choose Numba as my bilingual wrapper, Python
-and-C/C++
. However, it does not mean that other packages are not favorable.
Besides, it is also good to know that there are the “mutual-linking” with the Cuda and C compilers, please have a look on it question: https://stackoverflow.com/questions/6622454/cuda-incompatible-with-my-gcc-version
Also, it is good to know that there are modifications of environment variables on make
commands.. https://stackoverflow.com/questions/3239343/make-install-but-not-to-default-directories
..though, I do think I am a middle high-level C/C++ programer.. (awkward happens..)