Tiles & open gL?

Graphics and GPU Programming Programming

Started by DevLiquidKnight June 02, 2003 11:14 PM

5 comments, last by DevLiquidKnight 20 years, 10 months ago

834

Author

June 02, 2003 11:14 PM

I making a 2D tile based game and I am finding it is extremely slow when I load the tiles I dont know what im doing wrong ot load the tiles i use a for loop which creates a bunch of square planes using the glVertex3f and such. After its done and i load it the FPS drop to like really low. What am i doing wrong is there some other way of loading 2d tiles into open gl I dont understand what im doing wrong :/ lol any help would be nice! Thanks! coder requires 0xf00d before continue().
Killer Eagle Software

Aldacron

4,545

June 03, 2003 02:03 AM

You say loading, but you mean rendering, right? Without any code to see, it will be hard to know what the problem is. How about posting the problem code so someone can try to help you?

--- Official D Blog | Learning D | The One With D | D Bits

DevLiquidKnight

834

Author

June 03, 2003 02:12 AM

This is an example of how im doing the tiles :/for(int x=0; x<=100; x++){for(int y=0; y<=100; y++){x1=x/x+5;y1=y/y+5;glBegin(GL_QUADS);glTexCoord2f(0.0f, 1.0f); glVertex3f(-x1, y1, 0.0f);glTexCoord2f(1.0f, 1.0f); glVertex3f( x1, y1, 0.0f);		glTexCoord2f(1.0f, 0.0f); glVertex3f( x1,-y1, 0.0f);glTexCoord2f(0.0f, 0.0f); glVertex3f(-x1,-y1, 0.0f);glEnd();}}

coder requires 0xf00d before continue().

Killer Eagle Software

Null and Void

1,088

June 03, 2003 02:50 AM

Without really changing anything, it''d be better like this:

glBegin(GL_QUADS)for(unsigned int x=0; x<=100; ++x) {  x1=x/x+5; // I''ll assume your x1 & y1 stuff is correct  for(unsigned int y=0; y<=100; ++y) {      y1=y/y+5;    glTexCoord2f(0.0f, 1.0f); glVertex2f(-x1, y1);    glTexCoord2f(1.0f, 1.0f); glVertex2f( x1, y1);		    glTexCoord2f(1.0f, 0.0f); glVertex2f( x1,-y1);    glTexCoord2f(0.0f, 0.0f); glVertex2f(-x1,-y1);  }}glEnd();

That should be enormously faster if you''re not fillrate limited. However, that still is far from optimized. Here it is with client side vertex arrays:

// NOTE: DO THIS PART ONLY ONCE, STORE IT SOMEWHEREconst unsigned int x_max = 100; // I dislike your <= stuff,const unsigned int y_max = 100; // but, whatever =P...GLfloat vert_array[(x_max + 1) * (y_max + 1) * 4][2];GLfloat texc_array[(x_max + 1) * (y_max + 1) * 4][2];for(int x=0; x<=x_max; ++x) {  const unsigned int offset = y_max * x;  x1=x/x+5; // Same assumption as above  for(int y=0; y<=y_max; ++y) {      y1=y/y+5;    // Hopefully I didn''t screw this part up;    // you''ll get the idea either way...    texc_array[(offset + y) * 4 + 0][0] = 0.0f;    texc_array[(offset + y) * 4 + 0][1] = 1.0f;    vert_array[(offset + y) * 4 + 0][0] = -x1;    vert_array[(offset + y) * 4 + 0][1] =  y1;    texc_array[(offset + y) * 4 + 1][0] = 1.0f;    texc_array[(offset + y) * 4 + 1][1] = 1.0f;    vert_array[(offset + y) * 4 + 1][0] = -x1;    vert_array[(offset + y) * 4 + 1][1] =  y1;    texc_array[(offset + y) * 4 + 2][0] = 1.0f;    texc_array[(offset + y) * 4 + 2][1] = 0.0f;    vert_array[(offset + y) * 4 + 2][0] = -x1;    vert_array[(offset + y) * 4 + 2][1] =  y1;    texc_array[(offset + y) * 4 + 3][0] = 0.0f;    texc_array[(offset + y) * 4 + 3][1] = 0.0f;    vert_array[(offset + y) * 4 + 3][0] = -x1;    vert_array[(offset + y) * 4 + 3][1] =  y1;  }}// NOTE: DO THIS TO RENDER THAT// (Push states if you want, or whatever)glVertexPointer(2, GL_FLOAT, 0, vert_array);glTexCoordPointer(2, GL_FLOAT, 0, texc_array);glEnableClientState(GL_VERTEX_ARRAY);glEnableClientState(GL_TEXTURE_COORD_ARRAY);glDrawArrays(GL_QUADS, 0, (x_max + 1) * (y_max + 1) * 4);

From there, if you''re geometry limited, you can continue to improve on this with VBO or something that enables server side caching of the data (or VAR or VAO if you want to be vendor specific: you don''t).