Thursday, July 12, 2007

Most speed issues tackled

The biggest part of my slowdown was the C code. I had to do the same thing for the SNES version, since the compiler didn't do any optimizations. In addition, I've optimized the blitting and erasing code a lot. I don't think I can get too much more speed at the assembly level. This brings to mind Michael Abrash (he wrote an old Dr. Dobbs Journal column on graphics programming) columns on programming and optimzation. He basically said that optimization to a routine are really only limited by the perceived limitations of the problem.

Another big problem was the sequence of waiting for the vertical blank, changing Plane start address, and starting the update. I got this working correctly and the flicker appears to be gone.

The next challenge is rewriting the input code, which should speed up the code a little bit more.

0 Comments:

Post a Comment

<< Home