Shared library for intercepting CUDA Runtime API calls. This code was part of my Bachelor thesis: "A Study on the Computational Exploitation of Remote Virtualized Graphics Cards" (https://bit.ly/37tIG0D)
Prerequisites: -GNU/Linux for compilation -Set CUDA_PATH variable in the Makefile to the correct directory where cuda is installed.
Tested on: gcc-6,gcc-7
How to compile: $ make
To remove: $ make clean
How to run: $ LD_PRELOAD=/full_path_to_thecuda_intercept_directory/cuda_intercept/lib_cuda_intercept.so ./full_path_to_the_directory_of_the_CUDA_Program/your_cuda_program.cu
e.g. LD_PRELOAD=/home/cuda_intercept/lib_cuda_intercept.so /home/NVIDIA_CUDA-9.0_Samples/6_Advanced/transpose/transpose