fblock_opencl.c 2.2 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667
  1. /* StarPU --- Runtime system for heterogeneous multicore architectures.
  2. *
  3. * Copyright (C) 2010 Centre National de la Recherche Scientifique
  4. * Copyright (C) 2011 Université de Bordeaux 1
  5. *
  6. * StarPU is free software; you can redistribute it and/or modify
  7. * it under the terms of the GNU Lesser General Public License as published by
  8. * the Free Software Foundation; either version 2.1 of the License, or (at
  9. * your option) any later version.
  10. *
  11. * StarPU is distributed in the hope that it will be useful, but
  12. * WITHOUT ANY WARRANTY; without even the implied warranty of
  13. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
  14. *
  15. * See the GNU Lesser General Public License in COPYING.LGPL for more details.
  16. */
  17. #include <starpu.h>
  18. #include <starpu_opencl.h>
  19. extern struct starpu_opencl_program opencl_program;
  20. void opencl_func(void *buffers[], void *cl_arg)
  21. {
  22. int id, devid, err;
  23. cl_kernel kernel;
  24. cl_command_queue queue;
  25. cl_event event;
  26. int *factor = cl_arg;
  27. cl_mem block = (cl_mem)STARPU_BLOCK_GET_PTR(buffers[0]);
  28. int nx = (int)STARPU_BLOCK_GET_NX(buffers[0]);
  29. int ny = (int)STARPU_BLOCK_GET_NY(buffers[0]);
  30. int nz = (int)STARPU_BLOCK_GET_NZ(buffers[0]);
  31. unsigned ldy = STARPU_BLOCK_GET_LDY(buffers[0]);
  32. unsigned ldz = STARPU_BLOCK_GET_LDZ(buffers[0]);
  33. id = starpu_worker_get_id();
  34. devid = starpu_worker_get_devid(id);
  35. err = starpu_opencl_load_kernel(&kernel, &queue, &opencl_program, "fblock_opencl", devid);
  36. if (err != CL_SUCCESS) STARPU_OPENCL_REPORT_ERROR(err);
  37. err = 0;
  38. err = clSetKernelArg(kernel, 0, sizeof(block), &block);
  39. err = clSetKernelArg(kernel, 1, sizeof(nx), &nx);
  40. err = clSetKernelArg(kernel, 2, sizeof(ny), &ny);
  41. err = clSetKernelArg(kernel, 3, sizeof(nz), &nz);
  42. err = clSetKernelArg(kernel, 4, sizeof(ldy), &ldy);
  43. err = clSetKernelArg(kernel, 5, sizeof(ldz), &ldz);
  44. err |= clSetKernelArg(kernel, 6, sizeof(*factor), factor);
  45. if (err) STARPU_OPENCL_REPORT_ERROR(err);
  46. {
  47. size_t global=nx*ny*nz;
  48. err = clEnqueueNDRangeKernel(queue, kernel, 1, NULL, &global, NULL, 0, NULL, &event);
  49. if (err != CL_SUCCESS) STARPU_OPENCL_REPORT_ERROR(err);
  50. }
  51. clFinish(queue);
  52. starpu_opencl_collect_stats(event);
  53. clReleaseEvent(event);
  54. starpu_opencl_release_kernel(kernel);
  55. }