vector_scal_cuda.cu 1.4 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546
  1. /*
  2. * StarPU
  3. * Copyright (C) INRIA 2008-2009 (see AUTHORS file)
  4. *
  5. * This program is free software; you can redistribute it and/or modify
  6. * it under the terms of the GNU Lesser General Public License as published by
  7. * the Free Software Foundation; either version 2.1 of the License, or (at
  8. * your option) any later version.
  9. *
  10. * This program is distributed in the hope that it will be useful, but
  11. * WITHOUT ANY WARRANTY; without even the implied warranty of
  12. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
  13. *
  14. * See the GNU Lesser General Public License in COPYING.LGPL for more details.
  15. */
  16. /*
  17. * This example complements vector_scale.c: here we implement a CUDA version.
  18. */
  19. #include <starpu.h>
  20. static __global__ void vector_mult_cuda(float *val, unsigned n,
  21. float factor)
  22. {
  23. unsigned i;
  24. for(i = 0 ; i < n ; i++)
  25. val[i] *= factor;
  26. }
  27. extern "C" void scal_cuda_func(void *buffers[], void *_args)
  28. {
  29. float *factor = (float *)_args;
  30. struct starpu_vector_interface_s *vector = (struct starpu_vector_interface_s *) buffers[0];
  31. /* length of the vector */
  32. unsigned n = STARPU_GET_VECTOR_NX(vector);
  33. /* local copy of the vector pointer */
  34. float *val = (float *)STARPU_GET_VECTOR_PTR(vector);
  35. /* TODO: use more blocks and threads in blocks */
  36. vector_mult_cuda<<<1,1>>>(val, n, *factor);
  37. cudaThreadSynchronize();
  38. }