search
HomeBackend DevelopmentGolangWhy is the performance of `moving_avg_concurrent2` not improving with increased concurrency, despite splitting the list into smaller chunks processed by individual goroutines?

Why is the performance of `moving_avg_concurrent2` not improving with increased concurrency, despite splitting the list into smaller chunks processed by individual goroutines?

Why does the performance of moving_avg_concurrent2 not improve with the increase of concurrent execution?

moving_avg_concurrent2 splits the list into smaller pieces and uses a single goroutine to handle each piece. For some reason (it's not clear why), this function using one goroutine is faster than moving_avg_serial4, but using multiple goroutines starts to perform worse than moving_avg_serial4.

Why moving_avg_concurrent3 is much slower than moving_avg_serial4?

The performance of moving_avg_concurrent3 is worse than moving_avg_serial4 when using a goroutine. Although increasing num_goroutines can improve performance, it is still worse than moving_avg_serial4.

Even though goroutines are lightweight, they are not completely free, is it possible that the overhead incurred is so large that it is even slower than moving_avg_serial4?

Yes, although goroutines are lightweight, they are not free. When using multiple goroutines, the overhead of launching, managing, and scheduling them may outweigh the benefits of increased parallelism.

Code

Function:

// 返回包含输入移动平均值的列表(已提供,即未优化)
func moving_avg_serial(input []float64, window_size int) []float64 {
    first_time := true
    var output = make([]float64, len(input))
    if len(input) > 0 {
        var buffer = make([]float64, window_size)
        // 初始化缓冲区为 NaN
        for i := range buffer {
            buffer[i] = math.NaN()
        }
        for i, val := range input {
            old_val := buffer[int((math.Mod(float64(i), float64(window_size))))]
            buffer[int((math.Mod(float64(i), float64(window_size))))] = val
            if !NaN_in_slice(buffer) && first_time {
                sum := 0.0
                for _, entry := range buffer {
                    sum += entry
                }
                output[i] = sum / float64(window_size)
                first_time = false
            } else if i > 0 && !math.IsNaN(output[i-1]) && !NaN_in_slice(buffer) {
                output[i] = output[i-1] + (val-old_val)/float64(window_size) // 无循环的解决方案
            } else {
                output[i] = math.NaN()
            }
        }
    } else { // 空输入
        fmt.Println("moving_avg is panicking!")
        panic(fmt.Sprintf("%v", input))
    }
    return output
}

// 返回包含输入移动平均值的列表
// 重新排列控制结构以利用短路求值
func moving_avg_serial4(input []float64, window_size int) []float64 {
    first_time := true
    var output = make([]float64, len(input))
    if len(input) > 0 {
        var buffer = make([]float64, window_size)
        // 初始化缓冲区为 NaN
        for i := range buffer {
            buffer[i] = math.NaN()
        }
        for i := range input {
            //            fmt.Printf("in mvg_avg4: i=%v\n", i)
            old_val := buffer[int((math.Mod(float64(i), float64(window_size))))]
            buffer[int((math.Mod(float64(i), float64(window_size))))] = input[i]
            if first_time && !NaN_in_slice(buffer) {
                sum := 0.0
                for j := range buffer {
                    sum += buffer[j]
                }
                output[i] = sum / float64(window_size)
                first_time = false
            } else if i > 0 && !math.IsNaN(output[i-1]) /* && !NaN_in_slice(buffer)*/ {
                output[i] = output[i-1] + (input[i]-old_val)/float64(window_size) // 无循环的解决方案
            } else {
                output[i] = math.NaN()
            }
        }
    } else { // 空输入
        fmt.Println("moving_avg is panicking!")
        panic(fmt.Sprintf("%v", input))
    }
    return output
}

// 返回包含输入移动平均值的列表
// 将列表拆分为较小的片段以使用 goroutine,但不使用串行版本,即我们仅在开头具有 NaN,因此希望减少一些开销
// 仍然不能扩展(随着大小和 num_goroutines 的增加,性能下降)
func moving_avg_concurrent2(input []float64, window_size, num_goroutines int) []float64 {
    var output = make([]float64, window_size-1, len(input))
    for i := 0; i  0 {
        num_items := len(input) - (window_size - 1)
        var barrier_wg sync.WaitGroup
        n := num_items / num_goroutines
        go_avg := make([][]float64, num_goroutines)
        for i := 0; i  0 {
        num_windows := len(input) - (window_size - 1)
        var output = make([]float64, len(input))
        for i := 0; i 

The above is the detailed content of Why is the performance of `moving_avg_concurrent2` not improving with increased concurrency, despite splitting the list into smaller chunks processed by individual goroutines?. For more information, please follow other related articles on the PHP Chinese website!

Statement
The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn
Go language pack import: What is the difference between underscore and without underscore?Go language pack import: What is the difference between underscore and without underscore?Mar 03, 2025 pm 05:17 PM

This article explains Go's package import mechanisms: named imports (e.g., import "fmt") and blank imports (e.g., import _ "fmt"). Named imports make package contents accessible, while blank imports only execute t

How to implement short-term information transfer between pages in the Beego framework?How to implement short-term information transfer between pages in the Beego framework?Mar 03, 2025 pm 05:22 PM

This article explains Beego's NewFlash() function for inter-page data transfer in web applications. It focuses on using NewFlash() to display temporary messages (success, error, warning) between controllers, leveraging the session mechanism. Limita

How to convert MySQL query result List into a custom structure slice in Go language?How to convert MySQL query result List into a custom structure slice in Go language?Mar 03, 2025 pm 05:18 PM

This article details efficient conversion of MySQL query results into Go struct slices. It emphasizes using database/sql's Scan method for optimal performance, avoiding manual parsing. Best practices for struct field mapping using db tags and robus

How do I write mock objects and stubs for testing in Go?How do I write mock objects and stubs for testing in Go?Mar 10, 2025 pm 05:38 PM

This article demonstrates creating mocks and stubs in Go for unit testing. It emphasizes using interfaces, provides examples of mock implementations, and discusses best practices like keeping mocks focused and using assertion libraries. The articl

How can I define custom type constraints for generics in Go?How can I define custom type constraints for generics in Go?Mar 10, 2025 pm 03:20 PM

This article explores Go's custom type constraints for generics. It details how interfaces define minimum type requirements for generic functions, improving type safety and code reusability. The article also discusses limitations and best practices

How to write files in Go language conveniently?How to write files in Go language conveniently?Mar 03, 2025 pm 05:15 PM

This article details efficient file writing in Go, comparing os.WriteFile (suitable for small files) with os.OpenFile and buffered writes (optimal for large files). It emphasizes robust error handling, using defer, and checking for specific errors.

How do you write unit tests in Go?How do you write unit tests in Go?Mar 21, 2025 pm 06:34 PM

The article discusses writing unit tests in Go, covering best practices, mocking techniques, and tools for efficient test management.

How can I use tracing tools to understand the execution flow of my Go applications?How can I use tracing tools to understand the execution flow of my Go applications?Mar 10, 2025 pm 05:36 PM

This article explores using tracing tools to analyze Go application execution flow. It discusses manual and automatic instrumentation techniques, comparing tools like Jaeger, Zipkin, and OpenTelemetry, and highlighting effective data visualization

See all articles

Hot AI Tools

Undresser.AI Undress

Undresser.AI Undress

AI-powered app for creating realistic nude photos

AI Clothes Remover

AI Clothes Remover

Online AI tool for removing clothes from photos.

Undress AI Tool

Undress AI Tool

Undress images for free

Clothoff.io

Clothoff.io

AI clothes remover

AI Hentai Generator

AI Hentai Generator

Generate AI Hentai for free.

Hot Article

R.E.P.O. Energy Crystals Explained and What They Do (Yellow Crystal)
2 weeks agoBy尊渡假赌尊渡假赌尊渡假赌
Repo: How To Revive Teammates
4 weeks agoBy尊渡假赌尊渡假赌尊渡假赌
Hello Kitty Island Adventure: How To Get Giant Seeds
4 weeks agoBy尊渡假赌尊渡假赌尊渡假赌

Hot Tools

Safe Exam Browser

Safe Exam Browser

Safe Exam Browser is a secure browser environment for taking online exams securely. This software turns any computer into a secure workstation. It controls access to any utility and prevents students from using unauthorized resources.

ZendStudio 13.5.1 Mac

ZendStudio 13.5.1 Mac

Powerful PHP integrated development environment

SublimeText3 English version

SublimeText3 English version

Recommended: Win version, supports code prompts!

Zend Studio 13.0.1

Zend Studio 13.0.1

Powerful PHP integrated development environment

Dreamweaver CS6

Dreamweaver CS6

Visual web development tools