d74652373c
refactor(socket): 更新广播和退出逻辑中的类型引用 更新socket服务器中广播功能和退出功能的代码， 将player.Player类型替换为player.ClientData类型， 并相应调整方法调用以适应新的数据结构。 feat(map): 添加LoadOrStore方法支持 在并发安全的swiss map中新增LoadOrStore方法， 提供原子性的加载或存储功能，增强map的操作能力。 refactor(login): 优化登录逻辑中的玩家获取方式 重构登录控制器中获取玩家对象的方式， 直接从
Concurrent Swiss Map 
Concurrent Swiss Map is an open-source Go library that provides a high-performance, thread-safe generic concurrent hash map implementation designed to handle concurrent access efficiently. It's built with a focus on simplicity, speed, and reliability, making it a solid choice for scenarios where concurrent access to a hash map is crucial.
Uses dolthub/swiss map implementation under the hood.
Installation
Supports 1.18+ Go versions because of Go Generics
go get github.com/mhmtszr/concurrent-swiss-map
Usage
New functions will be added soon...
package main
import (
"hash/fnv"
csmap "github.com/mhmtszr/concurrent-swiss-map"
)
func main() {
myMap := csmap.New[string, int](
// set the number of map shards. the default value is 32.
csmap.WithShardCount[string, int](32),
// if don't set custom hasher, use the built-in maphash.
csmap.WithCustomHasher[string, int](func(key string) uint64 {
hash := fnv.New64a()
hash.Write([]byte(key))
return hash.Sum64()
}),
// set the total capacity, every shard map has total capacity/shard count capacity. the default value is 0.
csmap.WithSize[string, int](1000),
)
key := "swiss-map"
myMap.Store(key, 10)
val, ok := myMap.Load(key)
println("load val:", val, "exists:", ok)
deleted := myMap.Delete(key)
println("deleted:", deleted)
ok = myMap.Has(key)
println("has:", ok)
empty := myMap.IsEmpty()
println("empty:", empty)
myMap.SetIfAbsent(key, 11)
myMap.Range(func(key string, value int) (stop bool) {
println("range:", key, value)
return true
})
count := myMap.Count()
println("count:", count)
// Output:
// load val: 10 exists: true
// deleted: true
// has: false
// empty: true
// range: swiss-map 11
// count: 1
}
Basic Architecture
Benchmark Test
Benchmark was made on:
- Apple M1 Max
- 32 GB memory
Benchmark test results can be obtained by running this file on local computers.
Benchmark Results
- Memory usage of the concurrent swiss map is better than other map implementations in all checked test scenarios.
- In high concurrent systems, the concurrent swiss map is faster, but in systems containing few concurrent operations, it works similarly to RWMutexMap.