vector_scal_opencl.c 2.1 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263
  1. /* StarPU --- Runtime system for heterogeneous multicore architectures.
  2. *
  3. * Copyright (C) 2010-2021 Université de Bordeaux, CNRS (LaBRI UMR 5800), Inria
  4. *
  5. * StarPU 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. * StarPU 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. #include <starpu.h>
  17. extern struct starpu_opencl_program programs;
  18. void vector_scal_opencl(void *buffers[], void *_args)
  19. {
  20. float *factor = _args;
  21. int id, devid;
  22. cl_int err;
  23. cl_kernel kernel;
  24. cl_command_queue queue;
  25. /* length of the vector */
  26. unsigned int n = STARPU_VECTOR_GET_NX(buffers[0]);
  27. /* OpenCL copy of the vector pointer */
  28. cl_mem val = (cl_mem) STARPU_VECTOR_GET_DEV_HANDLE(buffers[0]);
  29. id = starpu_worker_get_id();
  30. devid = starpu_worker_get_devid(id);
  31. err = starpu_opencl_load_kernel(&kernel, &queue, &programs, "vector_mult_opencl", devid);
  32. if (err != CL_SUCCESS) STARPU_OPENCL_REPORT_ERROR(err);
  33. err = clSetKernelArg(kernel, 0, sizeof(n), &n);
  34. err = clSetKernelArg(kernel, 1, sizeof(val), &val);
  35. err |= clSetKernelArg(kernel, 2, sizeof(*factor), factor);
  36. if (err) STARPU_OPENCL_REPORT_ERROR(err);
  37. {
  38. size_t global=n;
  39. size_t local;
  40. size_t s;
  41. cl_device_id device;
  42. starpu_opencl_get_device(devid, &device);
  43. err = clGetKernelWorkGroupInfo (kernel, device, CL_KERNEL_WORK_GROUP_SIZE, sizeof(local), &local, &s);
  44. if (err != CL_SUCCESS) STARPU_OPENCL_REPORT_ERROR(err);
  45. if (local > global) local=global;
  46. else global = (global + local-1) / local * local;
  47. err = clEnqueueNDRangeKernel(queue, kernel, 1, NULL, &global, &local, 0, NULL, NULL);
  48. if (err != CL_SUCCESS) STARPU_OPENCL_REPORT_ERROR(err);
  49. }
  50. starpu_opencl_release_kernel(kernel);
  51. }