CUDA C/C++ on Google Colaboratory

CUDA

What is CUDA?

Compute Unified Device Architecture is a parallel computing platform and programming model developed by Nvidia for general computing on its own GPUs. CUDA enables developers to speed up compute-intensive applications by harnessing the power of GPUs for the parallelizable part of the computation.

CUDA code doesn’t run on AMD CPU or Intel HD graphics unless you have a NVIDIA Hardware installed in your Machine.

What is Google Colab?

Google Colab is a free cloud service and the most important feature able to distinguish Colab from other free cloud services is; Colab offers GPU and is completely free! With Colab you can work on the GPU with CUDA C/C++ for free!

Does that mean not every one can use it? Not really, we can use other resources for that like Google Collaboratory So, let’s start with that!

Procedure:

• Go to https://colab.research.google.com in Browser and Click on New Notebook

Creating new notebook

• Switch runtime from CPU to GPU:
  Click on Runtime > Change runtime type > Hardware Accelerator > GPU > Save

CPU to GPU

• Completely uninstall any previous CUDA versions.We need to refresh the Cloud Instance of CUDA:

!apt-get --purge remove cuda nvidia* libnvidia-*
!dpkg -l | grep cuda- | awk '{print $2}' | xargs -n1 dpkg --purge
!apt-get remove cuda-*
!apt autoremove
!apt-get update
*Write code in a separate code Block and Run that code.Every line that starts with ‘!’, it will be executed as a command line command.

• Install CUDA Version 9:

!wget https://developer.nvidia.com/compute/cuda/9.2/Prod/local_installers/cuda-repo-ubuntu1604-9-2-local_9.2.88-1_amd64 -O cuda-repo-ubuntu1604-9-2-local_9.2.88-1_amd64.deb
!dpkg -i cuda-repo-ubuntu1604-9-2-local_9.2.88-1_amd64.deb
!apt-key add /var/cuda-repo-9-2-local/7fa2af80.pub
!apt-get update
!apt-get install cuda-9.2

• Check your CUDA installation by running the command given below :

!nvcc --version

Verifying installation

• Install a small extension to run nvcc from the Notebook cells:

!pip install git+https://github.com/andreinechaev/nvcc4jupyter.git

Extension installation

• Load the extension using the code given below:

%load_ext nvcc_plugin

loading extension

Now we are ready to run CUDA C/C++ code right in your Notebook.

• Checking if CUDA is working or not:
  To run the code in your notebook, add the %%cu extension at the beginning of your code.

%%cu
#include <stdio.h>
#include <stdlib.h>
__global__ void add(int *a, int *b, int *c) {
    *c = *a + *b;
}
int main() {
    int a, b, c;
// host copies of variables a, b & c
    int *d_a, *d_b, *d_c;
// device copies of variables a, b & c
    int size = sizeof(int);
// Allocate space for device copies of a, b, c
    cudaMalloc((void **)&d_a, size);
    cudaMalloc((void **)&d_b, size);
    cudaMalloc((void **)&d_c, size);
// Setup input values
    c = 0;
    a = 3;
    b = 5;
// Copy inputs to device
    cudaMemcpy(d_a, &a, size, cudaMemcpyHostToDevice);
    cudaMemcpy(d_b, &b, size, cudaMemcpyHostToDevice);
// Launch add() kernel on GPU
    add<<<1,1>>>(d_a, d_b, d_c);
// Copy result back to host
    cudaError err = cudaMemcpy(&c, d_c, size,    cudaMemcpyDeviceToHost);
if(err!=cudaSuccess) {
       printf("CUDA error copying to Host: %s\n", cudaGetErrorString(err));
    }
printf("result is %d\n",c);
// Cleanup
    cudaFree(d_a);
    cudaFree(d_b);
    cudaFree(d_c);
return 0;
}

code output

and YOU DID IT!