CuRidge

Ridge Regression written in CUDA

Usage

First make sure you have installed CUDA.

To use the regression just copy the files in src/ into your project and use it like shown in the examples.

There are three possible ways to include ridge into your project:

1. The first way just includes the project-file into an cuda file/project.

   Do the following to run it:

   Compile:

   nvcc -std=c++11  -o out first_example.cu -lcublas -lcusolver

   And run:

   ./out

2. The second way uses forward declaration in an c++ file without CMake. This means that we can call the cuda code in c++ projects.

   Do the following to run it:

   Compile:

   g++ -c second_example.cpp
   nvcc -c path_to_ridge/ridge.cu
   nvcc -std=c++11  -o out second_example.o ridge.o -lcublas -lcusolver

   And run:

   ./out

3. The third way uses forward declaration in a c++ file with CMake.

   Just copy the CMakeLists.txt inside the src folder to your ridge location and include ridge similar as shown in this repository in your CMakeLists.txt files.