Here's the code that supposed to display the texture:

SDL_RenderCopy(renderer, mainRaster.getOutput,null,null);

The code that runs the update process:

if(doubleBufferRegisters[0] == 0){
                doubleBufferRegisters[0] = 1;
                doubleBufferRegisters[1] = 0;
                doubleBufferRegisters[0] = 0;
                doubleBufferRegisters[1] = 1;
    SDL_LockTexture(frameBuffer[doubleBufferRegisters[0]], null, &fbData[doubleBufferRegisters[0]], &fbPitch[doubleBufferRegisters[0]]);

            for(int i ; i < layerList.length ; i++){
                layerList[i].updateRaster(fbData[doubleBufferRegisters[0]], fbPitch[doubleBufferRegisters[0]], palette);
            r = false;

My engine uses CPU based rendering for the most part. I had to remove some of the overhead, as the engine first had to create an SDL_Surface from the proprietary bitmap format, then an SDL_Texture, now the palette conversion is done on the fly, enabling the use of transparencies and non-paletted graphics. However, an old problem came back after this, namely I'm getting a black screen. Even back then it was almost impossible to fix and still don't know what I did to make it work (other that trial and error with small modifications). Bypassing the "double buffering" doesn't change anything.

I'm writing my engine in D, however any help for C++ also would be helpful.

  • \$\begingroup\$ Is there a member variable called getOutput, or are you actually missing parenthesis there? (Or is that some D stuff?) \$\endgroup\$ – Tyyppi_77 May 2 '16 at 5:31
  • \$\begingroup\$ Yes, there is, however I obscured it. That function returns the unlocked texture now, as it gives a bit of boost on the framerate. Changing to return the same exact texture only results in lower update rate, still no output. \$\endgroup\$ – László Szerémi May 2 '16 at 14:46
  • \$\begingroup\$ I think you would have to provide more code. As a side note, you should try coloring your buffers with a static red or something to see if the problem is the SDL rendering, your CPU rendering or potentially something else. \$\endgroup\$ – Tyyppi_77 May 2 '16 at 18:57

Well, it seems I had some problem with the alpha-blending algorithm I wrote in the in-line assembler, that uses MMX. While I could test that the code actually works and wrote actual data to the framebuffer, the image didn't show on the window.

Here's the assembly part to the alpha-blending:

            //mov       edi,dest;                         
            //mov       ebx,src;                          
            //mov       ecx,len;
            mov     EDX, alpha;

            //movzx   EAX,DL;
            movq    MM7,alphaMMXmul_const1;

            shl     EAX,16;
            add     EAX,EDX;
            mov     [alphaMMXmul_0],EAX;
            mov     [alphaMMXmul_0 + 4],EAX ;                     
            movq    MM6,alphaMMXmul_0; //mm6(X) = alpha (4 words)        
            pxor    MM5,MM5;
            psubusw MM7,MM6; //mm7(Y) = 256 – alpha (4 words)      

            //ALIGN 16                                    
            //  MixAlphaMMX32_MainLoop:                               

            //movd      MM0,[EDI]; //mm0(A) = 0 0 0 0 | 0 Ra Ga Ba            
            movd        MM0,dest;                     
            //movd      mm1,[ebx]; //mm1(B) = 0 0 0 0 | 0 Rb Gb Bb            
            movd        MM1,src;

            punpcklbw   MM0,MM5; //mm0 = 0 0 0 Ra | 0 Ga 0 Ba             
            punpcklbw   MM1,MM5; //mm1 = 0 0 0 Rb | 0 Gb 0 Bb             
            pmullw      MM0,MM6; //mm0 = 0 Ra*X | Ga*X Ba*X           
            pmullw      MM1,MM7; //mm1 = 0 Rb*Y | Gb*Y Bb*Y               
            paddusw     MM0,MM1; //mm0 = 0 Ra*X+Rb*y | Ga*X+Gb*y Ba*X+Bb*Y     
            psrlw       MM0,8; //mm0 = 0 0 0 Rc | 0 Gc 0 Bc           
            packuswb    MM0,MM0; //mm0 = 0 0 0 0 | 0 Rc Gc Bc             
            //movd      [edi],mm0;
            movd        dest,MM0;

            //dec       ecx ;             

            emms        ;

Going to use some non-scalar method until I come up with something better. The assembly code had built-in type conversion, most samples I could find on the internet even used floating point operations. Then I can focus on fixing the framerate issues either by reducing the redraw calls (by using a different framebuffer for all layers), or finding a faster way to update the framebuffer.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.