org.scijava.concurrent.Parallelization Maven / Gradle / Ivy
Show all versions of scijava-concurrent Show documentation
/*
* #%L
* SciJava library facilitating consistent parallel processing.
* %%
* Copyright (C) 2021 - 2024 SciJava developers.
* %%
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
* #L%
*/
package org.scijava.concurrent;
import java.util.concurrent.Callable;
import java.util.concurrent.ExecutorService;
/**
* This class allows to configure an algorithm for parallelization.
*
* The algorithm needs to use the {@link TaskExecutor} returned by
* {@link Parallelization#getTaskExecutor()} to implement the parallelization.
* Alternatively it can use {@link Parallelization#getExecutorService()}. But
* {@link TaskExecutor} is simpler and better suited for image processing
* algorithms.
*
* The algorithm can be executed singleThreaded, multiThreaded or by using a
* specified {@link ExecutorService} or {@link TaskExecutor}:
*
*
* {@code
*
* // Single-threaded call
* Parallelization.runSingleThreaded( () -> myAlgorithm( image ) );
*
* // Multi-threaded call
* Parallelization.runMultiThreaded( () -> myAlgorithm( image ) );
*
* // ExecutorService
* Parallelization.withExecutor( executorService, () -> myAlgorithm( image ) );
*
* // Multi-threaded is the default.
* // A normal function call, that's not somehow wrapped by
* // Parallelization.runSingleThreaded( ... ) runs multi-threaded.
* myAlgorithm( image );
*
* // Example Algorithm, that fills an image with ones.
* public void myAlgorithm( RandomAccessibleInterval< IntType > image )
* {
* TaskExecutor taskExecutor = Parallelization.getTaskExecutor();
* int numTasks = taskExecutor.suggestNumberOfTasks();
* List< Interval > chunks = IntervalChunks.chunkInterval( image, numTasks );
*
* // The TaskExecutor executes the forEach method in multiple threads, if requested.
* taskExecutor.forEach( chunks, chunk -> {
* for ( IntType pixel : Views.interval( image, chunk ) )
* pixel.setOne();
* } );
* }
* }
*
*/
public final class Parallelization {
private Parallelization() {
// prevent from instantiation
}
private static final ThreadLocal executor = ThreadLocal
.withInitial(TaskExecutors::multiThreaded);
// Methods to support the implementation of an multi-threaded algorithm
/**
* Returns the {@link TaskExecutor} that was set for this thread.
*/
public static TaskExecutor getTaskExecutor() {
return executor.get();
}
public static ExecutorService getExecutorService() {
return getTaskExecutor().getExecutorService();
}
// Method to call a multi-threaded algorithm
/**
* To run an algorithm single-threaded use:
*
* {@code Parallelization.runSingleThreaded( () -> myAlgorithm( input ) );}
*/
public static void runSingleThreaded(Runnable action) {
runWithExecutor(TaskExecutors.singleThreaded(), action);
}
/**
* To run an algorithm single-threaded use:
*
* {@code output = Parallelization.runSingleThreaded( () -> myAlgorithm( input ) );}
*/
public static T runSingleThreaded(Callable action) {
return runWithExecutor(TaskExecutors.singleThreaded(), action);
}
/**
* To run an algorithm multi-threaded use:
*
* {@code Parallelization.runMultiThreaded( () -> myAlgorithm( input ) );}
*/
public static void runMultiThreaded(Runnable action) {
runWithExecutor(TaskExecutors.multiThreaded(), action);
}
/**
* To run an algorithm multi-threaded use:
*
* {@code output = Parallelization.runMultiThreaded( () -> myAlgorithm( input ) );}
*/
public static T runMultiThreaded(Callable action) {
return runWithExecutor(TaskExecutors.multiThreaded(), action);
}
/**
* To run an algorithm a given number of threads use:
*
* {@code Parallelization.runWithNumThreads( numThreads, () -> myAlgorithm( input ) );}
*/
public static void runWithNumThreads(int numThreads, Runnable action) {
try (TaskExecutor taskExecutor = TaskExecutors.numThreads(numThreads)) {
runWithExecutor(taskExecutor, action);
}
}
/**
* To run an algorithm a given number of threads use:
*
* {@code output = Parallelization.runWithNumThreads( numThreads, () -> myAlgorithm( input ) );}
*/
public static R runWithNumThreads(int numThreads, Callable action) {
try (TaskExecutor taskExecutor = TaskExecutors.numThreads(numThreads)) {
return runWithExecutor(taskExecutor, action);
}
}
/**
* Executes the given {@link Runnable} with the given {@link ExecutorService},
* and waits for the execution to finish.
*/
public static void runWithExecutor(ExecutorService executorService,
Runnable action)
{
runWithExecutor(TaskExecutors.forExecutorService(executorService), action);
}
/**
* Executes the given {@link Callable} with the given {@link ExecutorService},
* waits for the execution to finish and returns the result.
*/
public static R runWithExecutor(ExecutorService executorService,
Callable action)
{
return runWithExecutor(TaskExecutors.forExecutorService(executorService),
action);
}
/**
* Executes the given {@link Runnable} with the given {@link TaskExecutor},
* and waits for the execution to finish.
*/
public static void runWithExecutor(TaskExecutor taskExecutor,
Runnable action)
{
try (Frame frame = setExecutorRequiresReset(taskExecutor)) {
action.run();
}
}
/**
* Executes the given {@link Callable} with the given {@link TaskExecutor},
* waits for the execution to finish and returns the result.
*/
public static T runWithExecutor(TaskExecutor taskExecutor,
Callable action)
{
try (Frame frame = setExecutorRequiresReset(taskExecutor)) {
return action.call();
}
catch (Exception e) {
throw new RuntimeException(e);
}
}
/**
* This method can be used to execute an algorithm with a given
* {@link TaskExecutor}. But it's easier to use {@link #runWithExecutor}.
*
* This method sets the {@link TaskExecutor} for the current thread. This can
* be used to execute a certain part of your code with the given
* {@link TaskExecutor}. It's mandatory to call the {@code close()} of the
* return {@link Frame frame} afterwards. This could be done using an
* try-with-resources statement.
*
*
* {@code
*
* try (Parallelization.Frame frame = Parallelization.setExecutorRequiresReset(
* taskExecutor))
* {
* myAlgorithm(input);
* }
* }
*
*
* Or by explicitly calling {@code frame.close()} in the finally block.
*
*
*
* {
* @code
*
* Parallelization.Frame frame = Parallelization.setExecutorRequiresReset(
* taskExecutor);
* try {
* myAlgorithm(input);
* }
* finally {
* frame.close();
* }
* }
*
*/
// NB: package-private to allow testing
static Frame setExecutorRequiresReset(TaskExecutor taskExecutor) {
final TaskExecutor old = executor.get();
executor.set(taskExecutor);
return () -> executor.set(old);
}
// NB: package-private to allow testing
interface Frame extends AutoCloseable {
@Override
void close(); // NB: Throws no exceptions
}
}