opencl3 0.12.3

A Rust implementation of the Khronos OpenCL 3.0 API and extensions.
Documentation 
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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297

// Copyright (c) 2021-2024 Via Technology Ltd. All Rights Reserved.
//
// 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.

#[cfg(any(feature = "CL_VERSION_2_0", feature = "dynamic"))]
extern crate opencl3;

use cl3::device::{
    CL_DEVICE_SVM_FINE_GRAIN_BUFFER, CL_DEVICE_SVM_FINE_GRAIN_SYSTEM, CL_DEVICE_TYPE_ALL,
    CL_DEVICE_TYPE_GPU,
};
use opencl3::Result;
use opencl3::command_queue::CommandQueue;
use opencl3::context::Context;
use opencl3::device::Device;
use opencl3::kernel::{ExecuteKernel, Kernel, create_program_kernels};
use opencl3::platform::get_platforms;
use opencl3::program::{CL_STD_2_0, Program};
use opencl3::svm::SvmVec;
use opencl3::types::cl_int;
use std::ptr;

// The OpenCL kernels in PROGRAM_SOURCE below use built-in work-group functions:
// work_group_reduce_add, work_group_scan_inclusive_add and work_group_broadcast
// which were introduced in OpenCL 2.0.
const PROGRAM_SOURCE: &str = r#"
kernel void sum_int (global int* sums,
                    global int const* values)
{
    int value = sub_group_reduce_add(values[get_global_id(0)]);

    if (0u == get_local_id(0))
        sums[get_group_id(0)] = value;
}

kernel void inclusive_scan_int (global int* output,
                                global int const* values)
{
    int sum = 0;
    size_t lid = get_local_id(0);
    size_t lsize = get_local_size(0);

    size_t num_groups = get_num_groups(0);
    for (size_t i = 0u; i < num_groups; ++i)
    {
        size_t lidx = i * lsize + lid;
        int value = sub_group_scan_inclusive_add(values[lidx]);
        output[lidx] = sum + value;

        sum += sub_group_broadcast(value, lsize - 1);
    }
}"#;

const SUM_KERNEL_NAME: &str = "sum_int";
const INCLUSIVE_SCAN_KERNEL_NAME: &str = "inclusive_scan_int";

#[test]
#[ignore]
fn test_opencl_2_kernel_example() -> Result<()> {
    let platforms = get_platforms()?;
    assert!(0 < platforms.len());

    /////////////////////////////////////////////////////////////////////
    // Query OpenCL compute environment
    let opencl_2: &str = "OpenCL 2";
    let opencl_3: &str = "OpenCL 3";

    // Find an OpenCL fine grained SVM, platform and device
    let mut device_id = ptr::null_mut();
    let mut is_fine_grained_svm: bool = false;
    for p in platforms {
        let platform_version = p.version()?;
        if platform_version.contains(&opencl_2) || platform_version.contains(&opencl_3) {
            let devices = p
                .get_devices(CL_DEVICE_TYPE_GPU)
                .expect("Platform::get_devices failed");

            for dev_id in devices {
                let device = Device::new(dev_id);
                let svm_mem_capability = device.svm_mem_capability();
                is_fine_grained_svm = 0 < svm_mem_capability & CL_DEVICE_SVM_FINE_GRAIN_BUFFER;
                if is_fine_grained_svm {
                    device_id = dev_id;
                    break;
                }
            }
        }
    }

    if is_fine_grained_svm {
        // Create OpenCL context from the OpenCL svm device
        let device = Device::new(device_id);
        let vendor = device.vendor()?;
        let vendor_id = device.vendor_id()?;
        println!("OpenCL device vendor name: {}", vendor);
        println!("OpenCL device vendor id: {:X}", vendor_id);

        /////////////////////////////////////////////////////////////////////
        // Initialise OpenCL compute environment

        // Create a Context on the OpenCL device
        let context = Context::from_device(&device).expect("Context::from_device failed");

        // Build the OpenCL program source.
        let program = Program::create_and_build_from_source(&context, PROGRAM_SOURCE, CL_STD_2_0)
            .expect("Program::create_and_build_from_source failed");

        // Create the kernels from the OpenCL program source.
        let kernels = create_program_kernels(&program)?;
        assert!(0 < kernels.len());

        let kernel_0_name = kernels[0].function_name()?;
        println!("OpenCL kernel_0_name: {}", kernel_0_name);

        let sum_kernel = if SUM_KERNEL_NAME == kernel_0_name {
            &kernels[0]
        } else {
            &kernels[1]
        };

        let inclusive_scan_kernel = if INCLUSIVE_SCAN_KERNEL_NAME == kernel_0_name {
            &kernels[0]
        } else {
            &kernels[1]
        };

        // Create a command_queue on the Context's device
        let queue = CommandQueue::create_default_with_properties(&context, 0, 0)
            .expect("CommandQueue::create_with_properties failed");

        // Get the svm capability of all the devices in the context.
        let svm_capability = context.get_svm_mem_capability();
        assert!(0 < svm_capability);

        // Create SVM vectors for the input and output data

        // The input data
        const ARRAY_SIZE: usize = 8;
        let value_array: [cl_int; ARRAY_SIZE] = [3, 2, 5, 9, 7, 1, 4, 2];

        // Copy into an OpenCL SVM vector
        let mut test_values =
            SvmVec::<cl_int>::allocate(&context, ARRAY_SIZE).expect("SVM allocation failed");
        test_values.copy_from_slice(&value_array);

        // Make test_values immutable
        let test_values = test_values;

        // The output data, an OpenCL SVM vector
        let mut results =
            SvmVec::<cl_int>::allocate_zeroed(&context, ARRAY_SIZE).expect("SVM allocation failed");

        // Run the sum kernel on the input data
        let sum_kernel_event = unsafe {
            ExecuteKernel::new(sum_kernel)
                .set_arg_svm(results.as_mut_ptr())
                .set_arg_svm(test_values.as_ptr())
                .set_global_work_size(ARRAY_SIZE)
                .enqueue_nd_range(&queue)?
        };

        // Wait for the kernel to complete execution on the device
        sum_kernel_event.wait()?;

        // Can access OpenCL SVM directly, no need to map or read the results
        println!("sum results: {:?}", results);
        assert_eq!(33, results[0]);
        assert_eq!(0, results[ARRAY_SIZE - 1]);

        // Run the inclusive scan kernel on the input data
        let kernel_event = unsafe {
            ExecuteKernel::new(inclusive_scan_kernel)
                .set_arg_svm(results.as_mut_ptr())
                .set_arg_svm(test_values.as_ptr())
                .set_global_work_size(ARRAY_SIZE)
                .enqueue_nd_range(&queue)?
        };

        kernel_event.wait()?;

        println!("inclusive_scan results: {:?}", results);
        assert_eq!(value_array[0], results[0]);
        assert_eq!(33, results[ARRAY_SIZE - 1]);
    } else {
        println!("OpenCL fine grained SVM capable device not found");
    }

    Ok(())
}

#[test]
#[ignore]
fn test_opencl_2_system_svm_example() -> Result<()> {
    let platforms = get_platforms()?;
    assert!(0 < platforms.len());

    /////////////////////////////////////////////////////////////////////
    // Query OpenCL compute environment
    let opencl_2: &str = "OpenCL 2";
    let opencl_3: &str = "OpenCL 3";

    // Find an OpenCL fine grained SVM, platform and device
    let mut device_id = ptr::null_mut();
    let mut is_fine_grained_system_svm: bool = false;
    for p in platforms {
        let platform_version = p.version()?;

        if platform_version.contains(&opencl_2) || platform_version.contains(&opencl_3) {
            let devices = p
                .get_devices(CL_DEVICE_TYPE_ALL)
                .expect("Platform::get_devices failed");

            for dev_id in devices {
                let device = Device::new(dev_id);
                let svm_mem_capability = device.svm_mem_capability();
                is_fine_grained_system_svm =
                    0 < svm_mem_capability & CL_DEVICE_SVM_FINE_GRAIN_SYSTEM;
                if is_fine_grained_system_svm {
                    device_id = dev_id;
                    break;
                }
            }
        }
    }

    if is_fine_grained_system_svm {
        // Create OpenCL context from the OpenCL svm device
        let device = Device::new(device_id);
        let vendor = device.vendor().expect("Device.vendor failed");
        let vendor_id = device.vendor_id().expect("Device.vendor_id failed");
        println!("OpenCL device vendor name: {}", vendor);
        println!("OpenCL device vendor id: {:X}", vendor_id);

        /////////////////////////////////////////////////////////////////////
        // Initialise OpenCL compute environment

        // Create a Context on the OpenCL svm device
        let context = Context::from_device(&device).expect("Context::from_device failed");

        // Build the OpenCL program source and create the kernel.
        let program = Program::create_and_build_from_source(&context, PROGRAM_SOURCE, "")
            .expect("Program::create_and_build_from_source failed");

        let kernel = Kernel::create(&program, SUM_KERNEL_NAME).expect("Kernel::create failed");

        // Create a command_queue on the Context's device
        let queue = CommandQueue::create_default_with_properties(&context, 0, 0)
            .expect("CommandQueue::create_default_with_properties failed");

        // The input data
        const ARRAY_SIZE: usize = 8;
        let value_array: [cl_int; ARRAY_SIZE] = [3, 2, 5, 9, 7, 1, 4, 2];

        // Copy into an OpenCL SVM vector
        let mut test_values =
            SvmVec::<cl_int>::allocate(&context, ARRAY_SIZE).expect("SVM allocation failed");
        test_values.copy_from_slice(&value_array);

        // Make test_values immutable
        let test_values = test_values;

        // The output data, an OpenCL SVM vector
        let mut results =
            SvmVec::<cl_int>::allocate_zeroed(&context, ARRAY_SIZE).expect("SVM allocation failed");

        // Run the sum kernel on the input data
        let sum_kernel_event = unsafe {
            ExecuteKernel::new(&kernel)
                .set_arg_svm(results.as_mut_ptr())
                .set_arg_svm(test_values.as_ptr())
                .set_global_work_size(ARRAY_SIZE)
                .enqueue_nd_range(&queue)?
        };

        // Wait for the kernel to complete execution on the device
        sum_kernel_event.wait()?;

        // Can access OpenCL SVM directly, no need to map or read the results
        println!("sum results: {:?}", results);
        assert_eq!(33, results[0]);
        assert_eq!(0, results[ARRAY_SIZE - 1]);
    } else {
        println!("OpenCL fine grained system SVM device not found")
    }

    Ok(())
}