SDL2内存RGB图像显示速度过慢，寻求高效渲染优化方案

问题说明 #

使用SDL2显示内存中的RGB值列表时，中等尺寸图像的屏幕更新耗时过长。测试代码中单帧更新耗时11.8毫秒，要实现60fps（每帧约16.7毫秒）的输出，留给其他任务的时间仅剩4.2毫秒，急需优化显示流程。

原测试代码 #

//gcc -Ofast -g "%f" `pkg-config --cflags --libs sdl2` -o SDL_test_3
#include <SDL2/SDL.h>
#include <stdio.h>
#include <time.h>

//====---The_General_Use_Variables---====
unsigned char uc1;

long l1;

float f1;
float f2;
float f3;
float t1;
float t2;

int main(){
    int sx=1000;
    int sy=700;
    int did_work;
    
    //初始化SDL
    did_work=SDL_Init(SDL_INIT_VIDEO);
    if (did_work!=0){printf("Issue_in_SDL_Init\n");}

    //创建窗口
    SDL_Window *window=NULL;
    window = SDL_CreateWindow( "win1", SDL_WINDOWPOS_UNDEFINED, SDL_WINDOWPOS_UNDEFINED, sx, sy, SDL_WINDOW_SHOWN );

    SDL_Event windowEvent;
    if (!window){printf("SDL_CreateWindow_failed\n");}

    //获取窗口Surface
    SDL_Surface *screenSurface;
    screenSurface=SDL_GetWindowSurface(window);

    //创建图像缓冲区
    unsigned char *im_buffer;im_buffer=malloc(sx*sy*3*sizeof(unsigned char));
    for(l1=0;l1<sx*sy*3;l1++){im_buffer[l1]=0;}

    //创建基于缓冲区的Surface
    SDL_Surface *im_surface = SDL_CreateRGBSurfaceFrom((void*)im_buffer,
            sx,
            sy,
            24,          // bits per pixel = 24
            sx*3,  // pitch
            0x0000FF,              // red mask
            0x00FF00,              // green mask
            0xFF0000,              // blue mask
            0);

    
    //初始绘制并更新窗口
    SDL_BlitSurface(im_surface, NULL, screenSurface, NULL);
    SDL_UpdateWindowSurface(window);
    
    //循环设置
    int seconds=5;
    t1=clock();
    t1+=CLOCKS_PER_SEC*seconds;
    t2=clock();
    long counter=0;
    
    uc1=0;
    while(t2<t1){
        uc1++;
        counter++;
        SDL_PollEvent(&windowEvent);
        
        //更新红色通道
        for(l1=0;l1<sx*sy*3;l1+=3){im_buffer[l1]=uc1;}
        //错误：每帧重新创建Surface
        im_surface=SDL_CreateRGBSurfaceFrom((void*)im_buffer,
            sx,
            sy,
            24,          // bits per pixel = 24
            sx*3,  // pitch
            0x0000FF,              // red mask
            0x00FF00,              // green mask
            0xFF0000,              // blue mask
            0);
        
        SDL_BlitSurface(im_surface, NULL, screenSurface, NULL);
        SDL_UpdateWindowSurface(window);
        
        t2=clock();
        }
    f1=counter;
    f2=seconds;
    printf("average_fps=%f \n",f1/f2);
    printf("frametime=%f \n",1000*f2/f1);
}

核心优化点 #

1. 避免每帧重复创建SDL_Surface #

SDL_CreateRGBSurfaceFrom是轻量包装内存的操作，但重复调用仍会产生不必要的资源分配/销毁开销。只需在初始化时创建一次，后续直接修改im_buffer即可，Surface会自动关联最新的缓冲区数据。

2. 切换到SDL_Renderer硬件加速（推荐） #

SDL_Surface是软件渲染，依赖CPU计算；SDL_Renderer利用GPU硬件加速，能大幅提升渲染效率，尤其适合大尺寸图像。

3. 优化内存访问与更新区域 #

• 尽量保证内存访问连续，减少缓存 miss；
• 若仅更新部分区域，使用SDL_UpdateWindowSurfaceRects替代SDL_UpdateWindowSurface，只传输变化区域的数据。

优化后的代码示例 #

方案一：Surface版本优化（最小改动） #

//gcc -Ofast -g "%f" `pkg-config --cflags --libs sdl2` -o SDL_test_3_opt
#include <SDL2/SDL.h>
#include <stdio.h>
#include <time.h>
#include <string.h>

unsigned char uc1;
long l1;
float f1, f2, t1, t2;

int main(){
    int sx=1000;
    int sy=700;
    int did_work;
    
    did_work=SDL_Init(SDL_INIT_VIDEO);
    if (did_work!=0){printf("Issue_in_SDL_Init\n");}

    SDL_Window *window=SDL_CreateWindow( "win1", SDL_WINDOWPOS_UNDEFINED, SDL_WINDOWPOS_UNDEFINED, sx, sy, SDL_WINDOW_SHOWN );
    if (!window){printf("SDL_CreateWindow_failed\n");}

    SDL_Surface *screenSurface=SDL_GetWindowSurface(window);
    unsigned char *im_buffer=malloc(sx*sy*3*sizeof(unsigned char));
    memset(im_buffer, 0, sx*sy*3); // 用memset替代循环初始化，更高效

    // 仅创建一次Surface
    SDL_Surface *im_surface = SDL_CreateRGBSurfaceFrom((void*)im_buffer,
            sx, sy, 24, sx*3,
            0x0000FF, 0x00FF00, 0xFF0000, 0);

    SDL_BlitSurface(im_surface, NULL, screenSurface, NULL);
    SDL_UpdateWindowSurface(window);
    
    int seconds=5;
    t1=clock() + CLOCKS_PER_SEC*seconds;
    long counter=0;
    uc1=0;

    SDL_Event windowEvent;
    while(clock() < t1){
        uc1++;
        counter++;
        SDL_PollEvent(&windowEvent);
        
        // 直接修改缓冲区，无需重新创建Surface
        for(l1=0;l1<sx*sy*3;l1+=3){
            im_buffer[l1] = uc1;
        }
        
        SDL_BlitSurface(im_surface, NULL, screenSurface, NULL);
        SDL_UpdateWindowSurface(window);
    }

    f1=counter;
    f2=seconds;
    printf("average_fps=%f \n",f1/f2);
    printf("frametime=%f \n",1000*f2/f1);

    // 释放资源
    SDL_FreeSurface(im_surface);
    free(im_buffer);
    SDL_DestroyWindow(window);
    SDL_Quit();
    return 0;
}

方案二：Renderer硬件加速版本（性能最优） #

//gcc -Ofast -g "%f" `pkg-config --cflags --libs sdl2` -o SDL_test_3_renderer
#include <SDL2/SDL.h>
#include <stdio.h>
#include <time.h>
#include <string.h>

unsigned char uc1;
long l1;
float f1, f2, t1, t2;

int main(){
    int sx=1000;
    int sy=700;
    int did_work;
    
    did_work=SDL_Init(SDL_INIT_VIDEO);
    if (did_work!=0){printf("Issue_in_SDL_Init\n");}

    // 创建窗口和硬件加速Renderer，开启VSYNC避免画面撕裂
    SDL_Window *window=SDL_CreateWindow( "win1", SDL_WINDOWPOS_UNDEFINED, SDL_WINDOWPOS_UNDEFINED, sx, sy, SDL_WINDOW_SHOWN );
    SDL_Renderer *renderer=SDL_CreateRenderer(window, -1, SDL_RENDERER_ACCELERATED | SDL_RENDERER_PRESENTVSYNC);
    if (!window || !renderer){printf("Create window/renderer failed\n");}

    unsigned char *im_buffer=malloc(sx*sy*3*sizeof(unsigned char));
    memset(im_buffer, 0, sx*sy*3);

    // 创建可流式更新的Texture
    SDL_Texture *texture=SDL_CreateTexture(renderer, SDL_PIXELFORMAT_RGB24, SDL_TEXTUREACCESS_STREAMING, sx, sy);

    int seconds=5;
    t1=clock() + CLOCKS_PER_SEC*seconds;
    long counter=0;
    uc1=0;

    SDL_Event windowEvent;
    while(clock() < t1){
        uc1++;
        counter++;
        SDL_PollEvent(&windowEvent);
        
        // 更新缓冲区数据
        for(l1=0;l1<sx*sy*3;l1+=3){
            im_buffer[l1] = uc1;
        }

        // 将缓冲区数据更新到Texture
        SDL_UpdateTexture(texture, NULL, im_buffer, sx*3);
        
        // 清空渲染目标并绘制Texture
        SDL_RenderClear(renderer);
        SDL_RenderCopy(renderer, texture, NULL, NULL);
        // 提交渲染结果到屏幕
        SDL_RenderPresent(renderer);
    }

    f1=counter;
    f2=seconds;
    printf("average_fps=%f \n",f1/f2);
    printf("frametime=%f \n",1000*f2/f1);

    // 释放资源
    SDL_DestroyTexture(texture);
    free(im_buffer);
    SDL_DestroyRenderer(renderer);
    SDL_DestroyWindow(window);
    SDL_Quit();
    return 0;
}

内容的提问来源于stack exchange，提问作者Robin