Neko  0.9.0
A portable framework for high-order spectral element flow simulations
opr_dudxyz.c
Go to the documentation of this file.
1 /*
2  Copyright (c) 2021-2022, The Neko Authors
3  All rights reserved.
4 
5  Redistribution and use in source and binary forms, with or without
6  modification, are permitted provided that the following conditions
7  are met:
8 
9  * Redistributions of source code must retain the above copyright
10  notice, this list of conditions and the following disclaimer.
11 
12  * Redistributions in binary form must reproduce the above
13  copyright notice, this list of conditions and the following
14  disclaimer in the documentation and/or other materials provided
15  with the distribution.
16 
17  * Neither the name of the authors nor the names of its
18  contributors may be used to endorse or promote products derived
19  from this software without specific prior written permission.
20 
21  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22  "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23  LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
24  FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
25  COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
26  INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
27  BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
28  LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
29  CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
31  ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
32  POSSIBILITY OF SUCH DAMAGE.
33 */
34 
35 #ifdef __APPLE__
36 #include <OpenCL/cl.h>
37 #else
38 #include <CL/cl.h>
39 #endif
40 
41 #include <stdio.h>
42 #include <device/device_config.h>
43 #include <device/opencl/jit.h>
44 #include <device/opencl/prgm_lib.h>
45 #include <device/opencl/check.h>
46 
47 #include "dudxyz_kernel.cl.h"
48 
52 void opencl_dudxyz(void *du, void *u,
53  void *dr, void *ds, void *dt,
54  void *dx, void *dy, void *dz,
55  void *jacinv, int *nel, int *lx) {
56  cl_int err;
57 
58  if (dudxyz_program == NULL)
59  opencl_kernel_jit(dudxyz_kernel, (cl_program *) &dudxyz_program);
60 
61  const size_t global_item_size = 256 * (*nel);
62  const size_t local_item_size = 256;
63 
64 #define STR(X) #X
65 #define CASE(LX) \
66  case LX: \
67  { \
68  cl_kernel kernel = clCreateKernel(dudxyz_program, \
69  STR(dudxyz_kernel_lx##LX), &err); \
70  CL_CHECK(err); \
71  \
72  CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), (void *) &du)); \
73  CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), (void *) &u)); \
74  CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), (void *) &dr)); \
75  CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_mem), (void *) &ds)); \
76  CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), (void *) &dt)); \
77  CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_mem), (void *) &dx)); \
78  CL_CHECK(clSetKernelArg(kernel, 6, sizeof(cl_mem), (void *) &dy)); \
79  CL_CHECK(clSetKernelArg(kernel, 7, sizeof(cl_mem), (void *) &dz)); \
80  CL_CHECK(clSetKernelArg(kernel, 8, sizeof(cl_mem), (void *) &jacinv)); \
81  \
82  CL_CHECK(clEnqueueNDRangeKernel((cl_command_queue) glb_cmd_queue, \
83  kernel, 1, NULL, &global_item_size, \
84  &local_item_size, 0, NULL, NULL)); \
85  } \
86  break
87 
88  switch(*lx) {
89  CASE(2);
90  CASE(3);
91  CASE(4);
92  CASE(5);
93  CASE(6);
94  CASE(7);
95  CASE(8);
96  CASE(9);
97  CASE(10);
98  CASE(11);
99  CASE(12);
100  }
101 }
102 
u
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ u
Definition: ax_helm_full_kernel.h:46
dx
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dx
Definition: ax_helm_full_kernel.h:49
dz
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dz
Definition: ax_helm_full_kernel.h:51
dy
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dy
Definition: ax_helm_full_kernel.h:50
jacinv
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ jacinv
Definition: ax_helm_full_kernel.h:62
dr
__global__ void const T *__restrict__ const T *__restrict__ dr
Definition: cdtp_kernel.h:107
ds
__global__ void const T *__restrict__ const T *__restrict__ const T *__restrict__ ds
Definition: cdtp_kernel.h:108
dt
__global__ void const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dt
Definition: cdtp_kernel.h:109
device_config.h
opencl_kernel_jit
void opencl_kernel_jit(const char *kernel, cl_program *program)
Definition: jit.c:50
CASE
#define CASE(LX)
opencl_dudxyz
void opencl_dudxyz(void *du, void *u, void *dr, void *ds, void *dt, void *dx, void *dy, void *dz, void *jacinv, int *nel, int *lx)
Definition: opr_dudxyz.c:52
prgm_lib.h
dudxyz_program
void * dudxyz_program