Commit b317d509 authored by David Goz's avatar David Goz 😴
Browse files

cuda-ompgpu examples

parent 46b8b006
Loading
Loading
Loading
Loading
+62 −0
Original line number Diff line number Diff line
//////////////////////////////////////////////////////////////////////////////////////////////////
// Assigment : write a CUDA code corresponding to the
// following sequential C code
//
// #include <stdio.h>
// #define N 100
// int main()
// {
//   for (int i=0 ; i<N ; i++)
//     printf("%d\n", (i * i));

//   return 0;
// }
//////////////////////////////////////////////////////////////////////////////////////////////////

//////////////////////////////////////////////////////////////////////////////////////////////////
// Author: David Goz
// mail  : david.goz@inaf.it
// date  : 06.07.2024
//
// - Compile the code:
//   $ nvcc classwork_1.cu -o classwork_1_cuda
// - Run the code:
//   $ ./classwork_1_cuda
// - Check the result:
//   $ ./classwork_1_cuda | tail -n 100 | sort -nk 5

//////////////////////////////////////////////////////////////////////////////////////////////////

#include <stdio.h>
#include <cuda.h>

#define N        100
#define NThreads 1024

__global__ void GPUkernel(const int size)
{
  const int myID = threadIdx.x + (blockIdx.x * blockDim.x);
  
  if (myID >= size)
    return;

  // C printf is supported on CUDA
  // C++ cout class is not supported in CUDA
  printf("Hello from CUDA thread: %d - result %d\n", myID, (myID * myID));

  return;
}

int main()
{
  printf("\n\t The host issues the kernel on the GPU \n");
  
  // kernel lunch
  GPUkernel<<<1, NThreads>>>(N);

  printf("\n\t cudaDeviceSynchronize \n");
  // GPU synchronization
  cudaDeviceSynchronize();

  return 0;
}
+79 −0
Original line number Diff line number Diff line
//////////////////////////////////////////////////////////////////////////////////////////////////
// Assigment : write a CUDA code corresponding to the
// following sequential C code
//
// #include <stdio.h>
// #define N 100
// int main()
// {
//   int A[N];
//
//   for (int i=0 ; i<N ; i++)
//     A[i] = (i * i);
//
//   return 0;
// }
//////////////////////////////////////////////////////////////////////////////////////////////////

//////////////////////////////////////////////////////////////////////////////////////////////////
// Author: David Goz
// mail  : david.goz@inaf.it
// date  : 17.11.2022
//
// - Compile the code:
//   $ nvcc classwork_2.cu -o classwork_2
// - Run the code:
//   $ ./classwork_2
// - Check the result:
//   $ ./classwork_2 | tail -n 100 | sort -nk 5

//////////////////////////////////////////////////////////////////////////////////////////////////

#include <iostream>
#include <stdlib.h>
#include <cuda.h>

#define N        100
#define NThreads 1024

__global__ void GPUkernel(      int *A,
			  const int size)
{
  const int myID = threadIdx.x + (blockIdx.x * blockDim.x);
  
  if (myID < size)
    A[myID] = (myID * myID);

  return;
}

int main()
{
  // allocate array that allows direct access of both host and device
  //  CUDA is responsible 
  int *A;
  const size_t size = (N * sizeof(int));
  cudaError error = cudaMallocManaged(&A, size);
  if (!error)
    std::cout << "Memory allocated for the host/device" << std::endl;
  else
    {
      std::cout << "Cannot allocate memory for the host/device. CUDA error : " << error << " ... aborting" << std::endl;
      exit(EXIT_FAILURE);
    }
  
  // kernel lunch
  GPUkernel<<<1, NThreads>>>(A, N);
  
  // device synchronization
  cudaDeviceSynchronize();
  
  // check the result
  for (size_t i=0 ; i<N ; i++)
    std::cout << "A[" << i << "] - Result: " << A[i] << std::endl;

  // free the memory
  cudaFree(A);
  
  return 0;
}
+62 −0
Original line number Diff line number Diff line
//////////////////////////////////////////////////////////////////////////////////////////////////
//
// OpenMP GPU Offload is available only on systems with NVIDIA GPUs with compute capability '>= cc70'
//
// Assigment : write a OMP-GPU code corresponding to the
// following sequential C code
//
// #include <stdio.h>
// #define N 100
// int main()
// {
//   for (int i=0 ; i<N ; i++)
//     printf("%d\n", (i * i));

//   return 0;
// }
//////////////////////////////////////////////////////////////////////////////////////////////////

//////////////////////////////////////////////////////////////////////////////////////////////////
// Author: David Goz
// mail  : david.goz@inaf.it
// date  : 06.07.2024
//
// - Compile the code to run on :
//   $ nvcc classwork_1.c -o classwork_1_omp
// - Run the code:
//   $ ./classwork_1_omp
// - Check the result:
//   $ ./classwork_1_omp | tail -n 100 | sort -nk 5

//////////////////////////////////////////////////////////////////////////////////////////////////

#include <stdio.h>
#include <omp.h>

#define N        100
#define NThreads 1024

void GPUkernelSerial(const int size)
{
#pragma omp target
  {
    if (!omp_is_initial_device())
      printf("\n\t GPU is executing GPUkernelSerial\n" );
    else
      printf("\n\t CPU is executing GPUkernelSerial\n" );
    
    for (int i=0 ; i<size ; i++)
      printf("Hello from OMP-GPU thread: %d - result %d\n", i, (i * i));
  }
  return;
}

int main()
{
  printf("\n\t The host issues the kernel on the GPU \n");
  
  /* kernel lunch using one GPU thread */
  GPUkernelSerial(N);

  return 0;
}