性能比较

~~~ package main import ( "fmt" "math/rand" "sync" "sync/atomic" "time" ) //读写锁 var rwLock sync.RWMutex var lock sync.Mutex func testRWLock() { var a map[int]int a = make(map[int]int, 5) var count int32 a[8] = 10 a[3] = 10 a[2] = 10 a[1] = 10 a[18] = 10 for i := 0; i < 2; i++ { go func(b map[int]int) { //rwLock.Lock() //读写锁的代码 lock.Lock() //互斥锁的代码 b[8] = rand.Intn(100) time.Sleep(10 * time.Microsecond) //微妙 //rwLock.Unlock() lock.Unlock() }(a) } for i := 0; i < 100; i++ { go func(b map[int]int) { for { //rwLock.RLock() //读写锁的代码 lock.Lock() time.Sleep(time.Millisecond) //rwLock.RUnlock() lock.Unlock() atomic.AddInt32(&count, 1) } }(a) } time.Sleep(time.Second * 20) fmt.Println(atomic.LoadInt32(&count)) } func main() { //互斥锁 testRWLock() //读多写少的时候，用读写锁 } ~~~ 输出结果： ~~~ 15589 ~~~ ~~~ package main import ( "fmt" "math/rand" "sync" "sync/atomic" "time" ) //读写锁 var rwLock sync.RWMutex var lock sync.Mutex func testRWLock() { var a map[int]int a = make(map[int]int, 5) var count int32 a[8] = 10 a[3] = 10 a[2] = 10 a[1] = 10 a[18] = 10 for i := 0; i < 2; i++ { go func(b map[int]int) { rwLock.Lock() //读写锁的代码 // lock.Lock() //互斥锁的代码 b[8] = rand.Intn(100) time.Sleep(10 * time.Microsecond) //微妙 rwLock.Unlock() // lock.Unlock() }(a) } for i := 0; i < 100; i++ { go func(b map[int]int) { for { rwLock.RLock() //读写锁的代码 // lock.Lock() time.Sleep(time.Millisecond) rwLock.RUnlock() // lock.Unlock() atomic.AddInt32(&count, 1) } }(a) } time.Sleep(time.Second * 20) fmt.Println(atomic.LoadInt32(&count)) } func main() { //互斥锁 testRWLock() //读多写少的时候，用读写锁 } ~~~ 输出结果： ~~~ 1525002 ~~~