GCD提供两种Dispatch Queue, 分别为Concurrent Dispatch Queue 和 Serial Dispatch Queue.
Serial Dispatch Queue 使用一个thread。
Concurrent Dispatch Queue 使用多个thread。
获取Dispatch Queue:
0)dispatch_queue_create
dispatch_queue_t queue = dispatch_queue_create("name.of.your.queue",NULL);
第一个参数为queue的名字,第二个参数传NULL表明要创建一个Serial Dispatch Queue, 传入DISPATCH_QUEUE_CONCURRENT 创建一个
Concurrent Dispatch Queue。
如果是对数据库表更新或更新文件,最好为每一个表或文件建立一个Serial Dispatch Queue,这样能保证只有一个thread会对数据进行更新。
如果对那些不会引起数据不一致问题的任务,需要放到Concurrent Dispatch Queue执行。
1)从Main Dispatch Queue/Global Dispatch Queue获取
//main queue
dispatch_queue_t mainQueue = dispatch_get_main_queue();
//global queue of high priority
dispatch_queue_t globalHigh = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_HIGH,0);
//global queue of default priority
dispatch_queue_t globalDefault = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT,0);
//global queue of low priority
dispatch_queue_t globalLow = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_LOW,0);
//global queue of background priority
dispatch_queue_t globalBackground = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_BACKGROUND,0);
Main Dispatch Queue -> Serial dispatch queue
Global Dispatch Queue -> Concurrent dispatch queue
dispatch_set_target_queue: 主要用来给新建的queue设置优先级
dispatch_queue_t serialQueue = dispatch_queue_create("name.of.queue",NULL);
dispatch_queue_t globalQueue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_BACKGROUND,0);
dispatch_set_target_queue(serialQueue, globalQueue);
需要注意的是,第一个参数是自定义的queue,而不是系统的queue。因为你不能给系统的queue设置权限。通过上面设置,serialQueue 就有了与globalQueue一样的优先级。其实这个函数不仅可以设置queue的优先级,还可以设置queue之间的层级结构。
dispatch_after: 过一段时间执行queue中的task
dispatch_time_t time = dispatch_time(DISPATCH_TIME_NOW, 3ull * NSEC_PER_SEC);
dispatch_after(time, dispatch_get_main_queue(),^{
NSLog(@"after 3 seconds...");});
Dispatch Group
0)dispatch_group_notify
dispatch group 可以在group中的dispatch queue都执行完之后,通过一个dispatch_group_notify通知回调。
dispatch_queue_t queue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT,0);
dispatch_group_t group = dispatch_group_create();
dispatch_group_async(group, queue, ^{NSLog(@"task1");});
dispatch_group_async(group, queue, ^{NSLog(@"task2");});
dispatch_group_async(group, queue, ^{NSLog(@"task3");});
dispatch_group_notify(group, dispatch_get_main_queue(), ^{NSLog(@"finished...");});
输出总会保证"finished..."会在最后一句输出。
1)dispatch_group_wait
dispatch_queue_t queue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT,0);
dispatch_group_t group = dispatch_group_create();
dispatch_group_async(group, queue, ^{NSLog(@"task1");});
dispatch_group_async(group, queue, ^{NSLog(@"task2");});
dispatch_group_async(group, queue, ^{NSLog(@"task3");});
dispatch_group_wait(group, DISPATCH_TIME_FOREVER);
dispatch_group_wait提供了一种类似超时的机制,当然如果等待时间设置为DISPATCH_TIME_FOREVER,功能和dispatch_group_notify是一样的。
dispatch_barrier_async
dispatch_barrier_async是一个可以让当前任务等待queue其他任务完成再执行的函数。例如,从数据库读取数据是可以多线程并发读取的,因为这样提高效率。但是,更新数据库数据,就要保证只有一个线程进行更新操作。所以,当更新操作发生的时候,一定要保证读操作和其他更新操作等待。
dispatch_queue_t queue = dispatch_queue_create("barrier.queue", DISPATCH_QUEUE_CONCURRENT);
dispatch_async(queue, read_block_0);
dispatch_async(queue, read_block_1);
dispatch_async(queue, read_block_2);
dispatch_async(queue, read_block_3);
dispatch_async(queue, read_block_4);
dispatch_barrier_async(queue, update_block);
dispatch_async(queue, read_block_5);
dispatch_async(queue, read_block_6);
dispatch_async(queue, read_block_7);
dispatch_async(queue, read_block_8);
dispatch_async(queue, read_block_9);
这样就可以保证只有当更新操作结束后,才会进行其他数据库读操作,保证了数据的一致性。
dispatch_apply
dispatch_apply用于给一个block添加到dispatch queue若干次。 “for”循环
dispatch_queue_t queue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT,0);
dispatch_apply(10, queue, ^(size_t index) { NSLog(@"%zu",index);});
NSLog(@"done...");
dispatch_suspend/dispatch_resume
用于暂停和继续执行queue
dispatch semaphore
dispatch semaphore 适合处理比dispatch queue或者dispatch_barrier_async更小颗粒度的操作。
例如,
dispatch_queue_t queue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT,0);
NSMutableArray *array = [[NSMutableArray alloc] init];
for (int i = 0; i < 100000; i++)
{
dispatch_async(queue, ^{
[array addObject:[NSNumber numberWithInt:i]];
});
}
如果执行上段代码,因为NSMutableArray不是thread safe的,所以在global queue中对其进行添加对象,会导致程序crash掉。
使用dispatch semaphore我们可以实现多线程对NSMutableArray进行操作
dispatch_queue_t queue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT,0);
dispatch_semaphore_t semaphore = dispatch_semaphore_create(1);
NSMutableArray *array = [[NSMutableArray alloc] init];
for (int i = 0; i < 100000; i++)
{
dispatch_async(queue, ^{
dispatch_semaphore_wait(semaphore, DISPATCH_TIME_FOREVER);
[array addObject:[NSNumber numberWithInt:i]];
dispatch_semaphore_signal(semaphore);
});
}
dispatch semaphore 是一个信号量的counter,当counter为0,当前线程停止运行,当大于0,自减1,继续执行。
dispatch_semaphore_create
dispatch_semaphore_t semaphore = dispatch_semaphore_create(1);
初始化一个counter为1的semaphore.
dispatch_semaphore_wait
dispatch_semaphore_wait(semaphore, DISPATCH_TIME_FOREVER);
等待counter大于0才运行,第二个参数为等待时间。执行后counter减1.
dispatch_semaphore_signal
dispatch_semaphore_signal(semaphore);
给counter加1
所以在上面的代码中,就可以保证在wait和signal中的[array addObject:[NSNumber numberWithInt:i]]; 就只有一个thread进行操作,保证了线程安全。
dispatch_onece
保证了dispatch_once中的代码只会被执行一次,常用于单例。
+ (id)sharedManager {
static MyManager *sharedMyManager = nil;
static dispatch_once_t onceToken;
dispatch_once(&onceToken, ^{
sharedMyManager = [[self alloc] init];
});
return sharedMyManager;
}