c++ - How to update single pixels efficiently on SDL 1.2 surface? -

I am writing a small item that calculates RGB values ​​for each pixel of the image.

I want to display each after the calculation. The problem is that I did not know how to make it faster, it takes a few seconds to calculate the entire image and then display it, where each pixel takes a few minutes to appear.

It is important that I can see progress on the image count. Is there any way to solve this task efficiently in SDL 1.2?

My code now looks like this:

  SDL_Surface * screen = NULL; SDL_Surface * image = NULL; SDL_Init (SDL_INIT_EVERYTHING); // Set screen screen = SDL_SetVideoMode (img.width), img.height (), 32, SDL_SWSURFACE); // Create the surface with the image of the basic values ​​= SDL_CreateRGBSurface (SDL_SWSURFACE, img.width), img.height (), 32,0,0,0,0,0,0); If (image == faucet) {std :: cout & lt; & Lt; "SDL failed to form surface"; } // Create an array for pixels Uint32 * pixel = (Uint32 *) image- & gt; Pixels; // Apply the image to the SDL_ BlitSurface screen (image, zero, screen, tap); // Update Screen SDL_Flip (screen); // compute image (int i = 0; i  Format, (Uint8) getR (IMG (I),), (UINT8) getG (IMG (I, J)), (UIT 8) getB (img (i, j))); SDL_BlitSurface (image, tap, screen, tap); SDL_UpdateRect (Screen, I, J, I, J); }} // stop SDL Della (2000); // Free the loaded image SDL_FreeSurface (image); // Exit SDL SDL_Quit;  

You do not need an intermediate image, you can put pixels directly on the surface of the screen Are there. For each pixel, you can call SDL_Flip () or SDL_FupetRect () for each pixel, it will be very fast.

In addition, looping through lines rather than columns, will boost you a little speed, because the pixels are located in a physical memory on a single line, and the modern CPU's sequential memory access It's fast.

Create an array for pixels Uint32 * pixels = (Uint32 *) screen-> Pixels; // Clear screen SDL_FillRect (screen, tap, 0); // Update Screen SDL_Flip (screen); // for compute image (int j = 0; j & lt; img.height (); j ++) {for (int i = 0; i & lt; img.width (); i ++) {img (I, j) = computecolor (i, j, img, wthath), img. height ()); Pixel [i + j * img.width ()] = SDL_MapRGB (image-> Format, (Uint8) getR (IMG (I, J)), (Uint8) getG (IMG (I, J)), (UIT 8) getB (img (i, j))); } SDL_UpdateRect (screen, i, j, img.width (), 1); } // Display the final result SDL_Flip (screen); SDL_Delay (2000);

You can also create a different variable variable for img.height () and img.width (), so that you can not call each repetition function, but the compiler usually does anyway Able to customize