1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
|
//
// Copyright (c) 2017 The Khronos Group Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
#include "harness/compat.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/types.h>
#include <sys/stat.h>
#include "procs.h"
const char *wg_barrier_kernel_code =
"__kernel void compute_sum(__global int *a, int n, __global int *tmp_sum, __global int *sum)\n"
"{\n"
" int tid = get_local_id(0);\n"
" int lsize = get_local_size(0);\n"
" int i;\n"
"\n"
" tmp_sum[tid] = 0;\n"
" for (i=tid; i<n; i+=lsize)\n"
" tmp_sum[tid] += a[i];\n"
" \n"
" // updated to work for any workgroup size \n"
" for (i=hadd(lsize,1); lsize>1; i = hadd(i,1))\n"
" {\n"
" work_group_barrier(CLK_GLOBAL_MEM_FENCE);\n"
" if (tid + i < lsize)\n"
" tmp_sum[tid] += tmp_sum[tid + i];\n"
" lsize = i; \n"
" }\n"
"\n"
" //no barrier is required here because last person to write to tmp_sum[0] was tid 0 \n"
" if (tid == 0)\n"
" *sum = tmp_sum[0];\n"
"}\n";
static int
verify_sum(int *inptr, int *tmpptr, int *outptr, int n)
{
int i;
int reference = 0;
for (i=0; i<n; i++)
{
reference += inptr[i];
}
if (reference != outptr[0])
{
log_error("work_group_barrier test failed\n");
return -1;
}
log_info("work_group_barrier test passed\n");
return 0;
}
int
test_wg_barrier(cl_device_id device, cl_context context, cl_command_queue queue, int num_elements)
{
cl_mem streams[3];
cl_int *input_ptr = NULL, *output_ptr = NULL, *tmp_ptr =NULL;
cl_program program;
cl_kernel kernel;
size_t global_threads[3];
size_t local_threads[3];
int err;
int i;
size_t max_local_workgroup_size[3];
size_t max_threadgroup_size = 0;
MTdata d;
err = create_single_kernel_helper_with_build_options(
context, &program, &kernel, 1, &wg_barrier_kernel_code, "compute_sum",
nullptr);
test_error(err, "Failed to build kernel/program.");
err = clGetKernelWorkGroupInfo(kernel, device, CL_KERNEL_WORK_GROUP_SIZE,
sizeof(max_threadgroup_size), &max_threadgroup_size, NULL);
test_error(err, "clGetKernelWorkgroupInfo failed.");
err = clGetDeviceInfo(device, CL_DEVICE_MAX_WORK_ITEM_SIZES, sizeof(max_local_workgroup_size), max_local_workgroup_size, NULL);
test_error(err, "clGetDeviceInfo failed for CL_DEVICE_MAX_WORK_ITEM_SIZES");
// Pick the minimum of the device and the kernel
if (max_threadgroup_size > max_local_workgroup_size[0])
max_threadgroup_size = max_local_workgroup_size[0];
// work group size must divide evenly into the global size
while( num_elements % max_threadgroup_size )
max_threadgroup_size--;
input_ptr = (int*)malloc(sizeof(int) * num_elements);
output_ptr = (int*)malloc(sizeof(int));
streams[0] = clCreateBuffer(context, CL_MEM_READ_WRITE,
sizeof(cl_int) * num_elements, NULL, &err);
test_error(err, "clCreateBuffer failed.");
streams[1] =
clCreateBuffer(context, CL_MEM_READ_WRITE, sizeof(cl_int), NULL, &err);
test_error(err, "clCreateBuffer failed.");
streams[2] =
clCreateBuffer(context, CL_MEM_READ_WRITE,
sizeof(cl_int) * max_threadgroup_size, NULL, &err);
test_error(err, "clCreateBuffer failed.");
d = init_genrand( gRandomSeed );
for (i=0; i<num_elements; i++)
input_ptr[i] = (int)get_random_float(-0x01000000, 0x01000000, d);
free_mtdata(d); d = NULL;
err = clEnqueueWriteBuffer(queue, streams[0], CL_TRUE, 0, sizeof(cl_int)*num_elements, (void *)input_ptr, 0, NULL, NULL);
test_error(err, "clEnqueueWriteBuffer failed.");
err = clSetKernelArg(kernel, 0, sizeof streams[0], &streams[0]);
err |= clSetKernelArg(kernel, 1, sizeof num_elements, &num_elements);
err |= clSetKernelArg(kernel, 2, sizeof streams[2], &streams[2]);
err |= clSetKernelArg(kernel, 3, sizeof streams[1], &streams[1]);
test_error(err, "clSetKernelArg failed.");
global_threads[0] = max_threadgroup_size;
local_threads[0] = max_threadgroup_size;
err = clEnqueueNDRangeKernel( queue, kernel, 1, NULL, global_threads, local_threads, 0, NULL, NULL );
test_error(err, "clEnqueueNDRangeKernel failed.");
err = clEnqueueReadBuffer( queue, streams[1], true, 0, sizeof(cl_int), (void *)output_ptr, 0, NULL, NULL );
test_error(err, "clEnqueueReadBuffer failed.");
err = verify_sum(input_ptr, tmp_ptr, output_ptr, num_elements);
// cleanup
clReleaseMemObject(streams[0]);
clReleaseMemObject(streams[1]);
clReleaseMemObject(streams[2]);
clReleaseKernel(kernel);
clReleaseProgram(program);
free(input_ptr);
free(output_ptr);
return err;
}
|
