Simple GPU runtime support library based on driver API

This library allows running CUDA applications using only the driver API (https://docs.nvidia.com/cuda/archive/9.0/cuda-driver-api/index.html). It's functional enough to run simple CUDA applications, but is not tested beyond that.

NOTE: Google is not planning to make any further changes to this library. It's a proof-of-concept implementation released in hope that it will be useful on platforms where the CUDA runtime library is not available.

Limitations

This library only works for CUDA code compiled with CUDA-9.0 or older. More recent CUDA versions use slightly different API to set up and launch kernels and will need to implement a handful of additional functions.

The library has only been tested on Linux.

Building

Prerequisites: clang-7.0.0 or newer.

$ # git clone <...>/gpu-runtime.git
$ cd gpu-runtime
$ git submodule update --init --recursive
$ mkdir build
$ cd build
$ cmake cmake -DCUDA_ROOT=/path/to/cuda-9.0 -DCMAKE_CXX_COMPILER=/path/to/recent/clang++ ../
$ make -j 8
$ make test

Using

Link your application with gpu-runtime.o instead of -lcudart.