#Gosync map how to#
#Gosync map update#
#Gosync map registration#

When visited many times, Find out dirty map in ,read map Does not exist in the, Lead to misses The quantity is greater than or equal to dirty map The number of elements of, It will be copied as a whole dirty map To read map.

stay dirty map Exists in, stay read map There may also be.

stay read map Exists in, stay dirty map There may not be.

SyncMap.Range(func(key, value interface read map And dirty map The relationship between ？ Whatever the reason, When you read this article, In certain concurrency scenarios, It is recommended to use sync.Map Instead of Map + Mutex Of. Let a simple Map It has a higher learning cost to use. lost map Due privilege grammar, Such as ：make, map etc.

sync.Map It's hard to use in itself, It doesn't look like a Map.

I've roughly analyzed that you'd rather use Mutex + Map, I don't want to use sync.Map Why ： Maybe write Java My classmates are familiar with the above code, But it was a mistake, About why it was used wrong and what the impact would be, Please pay attention to the following articles. Here sync.Map It would be more reasonable, because sync.Map The bottom layer completely contains this logic.

#Gosync map registration#

Here's the code, That seems reasonable, Actually, it's a mistake （ background ： Get registration information in concurrent scenarios ）. But whether it's daily development or open source projects, As if sync.Map It's not being used very well, People are still used to using Mutex + Map To use.

#Gosync map update#

If it is found that the element does not exist in read, but exists in dirty, the point to update entry is written directly.This article will analyze from the perspective of usage and source code sync.Map.

Change the element state from deleted to nil.

It will perform the following operations.

If it is found that the element exists in read but has been marked as deleted, dirty is not equal to nil (the element certainly does not exist in dirty).

It is divided into the following three processing branches: Since the dirty process has been reached, it is called directly at the beginning Lockmethod Upper mutexIt is also important to ensure data security The first act of poor performance。 Let’s verify it.įirst, we define the basic data structure: //Represents a mutexįooMap = &FooMap) performance testingĪfter listening to the official documents introducing a bunch of benefits, he did not talk about the disadvantages, and whether the advantages after performance optimization are true and credible.

When multiple goroutines read, write, and overwrite entries of unrelated key sets.Ĭompared with go map with separate mutex or rwmutex, using map type can greatly reduce lock contention.

For example, in a cache that will only grow, there will be such a business scenario. When an entry for a given key is written only once but read multiple times.Most code should use native maps instead of individual locking or coordination controls for better type security and maintainability.Īt the same time, the map type has also been optimized for the following scenarios:.The concurrent use of multiple goroutines is secure and does not require additional locking or coordination control.Some suggestions on map types are clearly pointed out in the official go documents: Happily start the road of sucking fish together. Then, according to the results of map performance analysis, sync Map source code anatomy to understand why.

#Gosync map how to#

Today’s fried fish will reveal your secret go sync Map, we will first understand what scenarios, how to use various types of go map, and who has the best performance! With a choice, there is always a choice difficulty, which is difficult How to choose the two? Whose performance is better？ A friend of mine said that the standard library sync Map performance is very good. Standard library sync Map (go1.9 and later).Native map + mutex or read / write lock.Before《 Why are go map and slice non thread safe?》In this paper, we discuss the non thread safety of map and slice in go language, and based on this, we extend the two most used concurrency support modes of map in the industry.