GPUEAExecutionContext.java

1 
package net.bmahe.genetics4j.gpu.spec;
2
3 
import java.util.function.Predicate;
4
5 
import org.immutables.value.Value;
6
7 
import net.bmahe.genetics4j.core.spec.AbstractEAExecutionContext;
8 
import net.bmahe.genetics4j.gpu.opencl.model.Device;
9 
import net.bmahe.genetics4j.gpu.opencl.model.Platform;
10
11 
/**
12 
 * GPU-specific execution context that extends the core EA framework with OpenCL device selection capabilities.
13 
 *
14 
 * <p>GPUEAExecutionContext extends {@link AbstractEAExecutionContext} to include GPU-specific execution parameters
15 
 * required for OpenCL device discovery and selection. This context combines traditional EA execution settings
16 
 * (population size, termination criteria) with GPU-specific device filtering capabilities.
17 
 *
18 
 * <p>Key GPU-specific additions:
19 
 * <ul>
20 
 * <li><strong>Platform filtering</strong>: Predicates for selecting appropriate OpenCL platforms</li>
21 
 * <li><strong>Device filtering</strong>: Predicates for selecting compatible OpenCL devices</li>
22 
 * <li><strong>Multi-device support</strong>: Automatic discovery and utilization of multiple GPU devices</li>
23 
 * <li><strong>Hardware abstraction</strong>: Device-agnostic configuration with runtime device selection</li>
24 
 * </ul>
25 
 *
26 
 * <p>Device selection workflow:
27 
 * <ol>
28 
 * <li><strong>Platform discovery</strong>: Enumerate all available OpenCL platforms</li>
29 
 * <li><strong>Platform filtering</strong>: Apply platform predicates to select compatible platforms</li>
30 
 * <li><strong>Device enumeration</strong>: Discover devices for each selected platform</li>
31 
 * <li><strong>Device filtering</strong>: Apply device predicates to select suitable devices</li>
32 
 * <li><strong>Context creation</strong>: Create OpenCL contexts for selected devices</li>
33 
 * </ol>
34 
 *
35 
 * <p>Common filtering patterns:
36 
 *
37 
 * <pre>{@code
38 
 * // Select only GPU devices with sufficient memory
39 
 * GPUEAExecutionContext<Double> context = GPUEAExecutionContext.<Double>builder()
40 
 * 		.populationSize(2000)
41 
 * 		.termination(Generations.of(100))
42 
 *
43 
 * 		// Platform filtering - prefer full OpenCL profiles
44 
 * 		.platformFilter(platform -> platform.profile() == PlatformProfile.FULL_PROFILE)
45 
 *
46 
 * 		// Device filtering - GPU devices with at least 2GB memory
47 
 * 		.deviceFilter(device -> device.type() == DeviceType.GPU && device.globalMemSize() >= 2L * 1024 * 1024 * 1024)
48 
 *
49 
 * 		.build();
50 
 *
51 
 * // Select any available compute device (GPUs or CPUs)
52 
 * GPUEAExecutionContext<Double> flexibleContext = GPUEAExecutionContext.<Double>builder()
53 
 * 		.populationSize(1000)
54 
 * 		.termination(FitnessTarget.of(0.95))
55 
 *
56 
 * 		// Accept any platform
57 
 * 		.platformFilter(platform -> true)
58 
 *
59 
 * 		// Prefer GPUs but accept CPUs as fallback
60 
 * 		.deviceFilter(device -> device.type() == DeviceType.GPU || device.type() == DeviceType.CPU)
61 
 *
62 
 * 		.build();
63 
 * }</pre>
64 
 *
65 
 * <p>Performance optimization through device selection:
66 
 * <ul>
67 
 * <li><strong>Compute capability</strong>: Filter devices by OpenCL version and feature support</li>
68 
 * <li><strong>Memory capacity</strong>: Ensure devices have sufficient memory for population size</li>
69 
 * <li><strong>Compute units</strong>: Prefer devices with more parallel processing units</li>
70 
 * <li><strong>Memory bandwidth</strong>: Select devices optimized for data-intensive operations</li>
71 
 * </ul>
72 
 *
73 
 * <p>Multi-device strategies:
74 
 * <ul>
75 
 * <li><strong>Load balancing</strong>: Automatic population distribution across selected devices</li>
76 
 * <li><strong>Heterogeneous computing</strong>: Utilize both GPU and CPU devices simultaneously</li>
77 
 * <li><strong>Fault tolerance</strong>: Graceful degradation when devices become unavailable</li>
78 
 * <li><strong>Resource optimization</strong>: Efficient utilization of available compute resources</li>
79 
 * </ul>
80 
 *
81 
 * <p>Default behavior:
82 
 * <ul>
83 
 * <li><strong>Platform acceptance</strong>: All platforms accepted by default</li>
84 
 * <li><strong>Device acceptance</strong>: All devices accepted by default</li>
85 
 * <li><strong>Discovery process</strong>: Automatic enumeration of available hardware</li>
86 
 * <li><strong>Validation</strong>: Runtime validation ensures at least one device is selected</li>
87 
 * </ul>
88 
 *
89 
 * @param <T> the type of fitness values used in the evolutionary algorithm
90 
 * @see AbstractEAExecutionContext
91 
 * @see Platform
92 
 * @see Device
93 
 * @see net.bmahe.genetics4j.gpu.GPUFitnessEvaluator
94 
 */
95 
@Value.Immutable
96 
public abstract class GPUEAExecutionContext<T extends Comparable<T>> extends AbstractEAExecutionContext<T> {
97
98 
	/**
99 
	 * Returns the predicate used to filter OpenCL platforms during device discovery.
100 
	 *
101 
	 * <p>Platform filtering allows selective use of OpenCL platforms based on vendor, version, profile, or other
102 
	 * platform characteristics. This enables optimization for specific hardware configurations or requirements.
103 
	 *
104 
	 * <p>Common filtering criteria:
105 
	 * <ul>
106 
	 * <li><strong>Profile support</strong>: Filter by FULL_PROFILE vs EMBEDDED_PROFILE</li>
107 
	 * <li><strong>Vendor preference</strong>: Select platforms from specific vendors</li>
108 
	 * <li><strong>Version requirements</strong>: Ensure minimum OpenCL version support</li>
109 
	 * <li><strong>Extension support</strong>: Filter platforms with required extensions</li>
110 
	 * </ul>
111 
	 *
112 
	 * @return the platform filtering predicate (default accepts all platforms)
113 
	 */
114 
	@Value.Default
115 
	public Predicate<Platform> platformFilters() {
116 3 1. lambda$platformFilters$0 : replaced boolean return with false for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::lambda$platformFilters$0 → NO_COVERAGE
2. platformFilters : replaced return value with null for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::platformFilters → NO_COVERAGE
3. lambda$platformFilters$0 : Substituted 1 with 0 → NO_COVERAGE
 
		return (platform) -> true;
117 
	}
118
119 
	/**
120 
	 * Returns the predicate used to filter OpenCL devices during device discovery.
121 
	 *
122 
	 * <p>Device filtering enables selection of appropriate compute devices based on type, capabilities, memory, and
123 
	 * performance characteristics. This allows optimization for specific workload requirements and hardware constraints.
124 
	 *
125 
	 * <p>Common filtering criteria:
126 
	 * <ul>
127 
	 * <li><strong>Device type</strong>: GPU, CPU, ACCELERATOR, or combinations</li>
128 
	 * <li><strong>Memory capacity</strong>: Minimum global or local memory requirements</li>
129 
	 * <li><strong>Compute units</strong>: Minimum parallel processing capability</li>
130 
	 * <li><strong>OpenCL version</strong>: Required feature support level</li>
131 
	 * <li><strong>Extensions</strong>: Specific OpenCL extension requirements</li>
132 
	 * </ul>
133 
	 *
134 
	 * @return the device filtering predicate (default accepts all devices)
135 
	 */
136 
	@Value.Default
137 
	public Predicate<Device> deviceFilters() {
138 3 1. lambda$deviceFilters$1 : replaced boolean return with false for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::lambda$deviceFilters$1 → NO_COVERAGE
2. deviceFilters : replaced return value with null for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::deviceFilters → NO_COVERAGE
3. lambda$deviceFilters$1 : Substituted 1 with 0 → NO_COVERAGE
 
		return (device) -> true;
139 
	}
140
141 
	/**
142 
	 * Creates a new builder for constructing GPU EA execution contexts.
143 
	 *
144 
	 * <p>The builder provides a fluent interface for specifying both core EA execution parameters and GPU-specific
145 
	 * device selection criteria. Type safety is ensured through generic parameterization.
146 
	 *
147 
	 * @param <U> the type of fitness values for the execution context
148 
	 * @return a new builder instance for creating GPU EA execution contexts
149 
	 */
150 
	public static <U extends Comparable<U>> ImmutableGPUEAExecutionContext.Builder<U> builder() {
151 2 1. builder : removed call to net/bmahe/genetics4j/gpu/spec/ImmutableGPUEAExecutionContext::builder → NO_COVERAGE
2. builder : replaced return value with null for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::builder → NO_COVERAGE
 
		return ImmutableGPUEAExecutionContext.builder();
152 
	}
153 
}

Mutations

116

1.1
Location : lambda$platformFilters$0
Killed by : none replaced boolean return with false for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::lambda$platformFilters$0 → NO_COVERAGE

2.2
Location : platformFilters
Killed by : none replaced return value with null for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::platformFilters → NO_COVERAGE

3.3
Location : lambda$platformFilters$0
Killed by : none Substituted 1 with 0 → NO_COVERAGE

138

1.1
Location : lambda$deviceFilters$1
Killed by : none replaced boolean return with false for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::lambda$deviceFilters$1 → NO_COVERAGE

2.2
Location : deviceFilters
Killed by : none replaced return value with null for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::deviceFilters → NO_COVERAGE

3.3
Location : lambda$deviceFilters$1
Killed by : none Substituted 1 with 0 → NO_COVERAGE

151

1.1
Location : builder
Killed by : none removed call to net/bmahe/genetics4j/gpu/spec/ImmutableGPUEAExecutionContext::builder → NO_COVERAGE

2.2
Location : builder
Killed by : none replaced return value with null for net/bmahe/genetics4j/gpu/spec/GPUEAExecutionContext::builder → NO_COVERAGE

Active mutators

Tests examined


Report generated by PIT 1.20.3