我正在使用具有 8 個內(nèi)核的機器（具有“2,8 GHz Intel Core i7”處理器的 Mac），我可以看到正在運行fmt.Println(runtime.NumCPU()).我已經(jīng)實現(xiàn)了一個非常簡單的工作池模型來同時處理一些進入池的請求。進程類型是“CPU 密集型”，我想感受一下在給 GO 更多內(nèi)核時性能會提高多少。所以代碼如下func Run(poolSize int, workSize int, loopSize int, maxCores int) {    runtime.GOMAXPROCS(maxCores)    var wg sync.WaitGroup    wg.Add(poolSize)    defer wg.Wait()    // this is the channel where we write the requests for work to be performed by the pool    workStream := make(chan int)    // cpuIntensiveWork simulates an CPU intensive process    var cpuIntensiveWork = func(input int) {        res := input        for i := 0; i < loopSize; i++ {            res = res + i        }    }    // worker is the function that gets fired by the pool    worker := func(wg *sync.WaitGroup, workStream chan int, id int) {        defer wg.Done()        for req := range workStream {            cpuIntensiveWork(req)        }    }    // launch the goroutines of the pool    for i := 0; i < poolSize; i++ {        go worker(&wg, workStream, i)    }    // feed the workStream until the end and then close the channel    for workItemNo := 0; workItemNo < workSize; workItemNo++ {        workStream <- workItemNo    }    close(workStream)}基準是這些var numberOfWorkers = 100var numberOfRequests = 1000var loopSize = 100000func Benchmark_1Core(b *testing.B) {    for i := 0; i < b.N; i++ {        Run(numberOfWorkers, numberOfRequests, loopSize, 1)    }}func Benchmark_2Cores(b *testing.B) {    for i := 0; i < b.N; i++ {        Run(numberOfWorkers, numberOfRequests, loopSize, 2)    }}func Benchmark_4Cores(b *testing.B) {    for i := 0; i < b.N; i++ {        Run(numberOfWorkers, numberOfRequests, loopSize, 4)    }}func Benchmark_8Cores(b *testing.B) {    for i := 0; i < b.N; i++ {        Run(numberOfWorkers, numberOfRequests, loopSize, 8)    }}運行基準測試我注意到，從 1 核到 2 核再到 4 核，性能幾乎呈線性增長。但是我從 4 核到 8 核的性能差異非常有限。這是預期的行為嗎？如果是這樣，根本原因是什么？