SpritesMind.Net

Hello to you all.
I've been toying around with the VDP FILL function, and especially, trying to trigger it again while already filling.
Here's my code: And here is the result : The first red line is a template, 40 pixels long.
The second line is our test. I first draw a blue line, to clearly see things,
then I trigger a FILL with red,
then, during this red FILL, I trigger a green FILL.
What does this picture tell us?

1. You can trigger a FILL while being already FILLing. Upon that, the current FILL is discarded while the new one starts.
2. When you trigger a new FILL, the value of length register is read again. I deduce this from the fact that the green line is ~40 pixels long
3. When you trigger a new FILL, the value of start register is NOT read again. I deduce this from the fact that the green line does NOT start from where the second red line started, that is, we can see some red pixels before the green ones.

It actually is pretty hard to have precise timing. Moreover, I usually see glitches on some frame.

Going further, we may imagine an image where the pixels would be put by FILLing, and changing length and data on the fly.
Pretty useless, but this reminds me of the oerg666 Blast Processing.

In my current 32x experiments, I'm using FILL (multiple times) to clear the background before drawing sprites. Are there any known issues to this?

Alternatively, do you know of faster methods of clearing the framebuffer?

That's exactly the stuff I want to try next.
Beforehand, I would say that the FILL is a bit faster than a well optimized loop, but I'll have us the numbers ^^

done some testing on a real PAL 32X, and FILL is ~36% faster than ASM loops.
Your mileage may vary.