ash, how does this go about rendering its maps? it seemed to take a while for the PRIZM's fancy pants hardware, so i thought you must be doing something fancypants yourself and wanted to ask before looking at the source.
Well, it really isn't much. I just realized no matter how much I optimize the rendering process, it'll still be slow, since it's pretty simplistic:
void drawtilemap(city_cell*map, int xpos, int ypos) {
city_cell*cur_element = 0x00000000;
const char*tile = 0x00000000;
for(int j = 0; j < 10 ; j++) {
for(int i = 0; i < 20; i++) {
cur_element = access_map_element(map, xpos+i, ypos+j);
if(cur_element->occupied) {
switch(cur_element->occupied_type) {
case utility:
break;
case residential:
case commercial:
case industrial:
tile = ((cur_element->occupied_type-1)*512)+zoning_tiles;
break;
case school:
break;
case police:
break;
case firestation:
break;
case hospital:
break;
case road:
break;
case political:
break;
case park:
break;
case landmark:
break;
case ruins:
break;
}
} else {
switch(cur_element->natural_cell){
case grass:
tile = grass_tile;
break;
case water:
tile = water_tile;
break;
case dirt:
tile = dirt_tile;
break;
case sand:
tile = sand_tile;
break;
case tree:
tile = tree_tile;
break;
case mountain:
tile = mountain_tile;
break;
}
}
CopySprite(tile, i*16, j*16+56, 16, 16);
tile = 0x00000000;
}
}
}What I did to make sure it runs fast is create an organized system of Renderloops controlled by a single one, and within the renderloops check to see if anything had to be done or changed; if not, do nothing. Also made sure that you couldn't do much at one single time within the renderloop, i.e. only one process at a time. The main helper here isn't that the rendering process is fancypants, but rather I just used good coding regimes
