Copper SKIP/WAIT timing

[roondar]

I've run into a bit of an oddity with my latest experiment.
The HRM states that WAIT & SKIP have different timing:

Originally Posted by HRM

The MOVE and SKIP . instructions require two memory cycles and two instruction words each. Because only the odd memory cycles are requested by the Copper, four memory cycle times are required per instruction. The WAIT instruction requires three memory cycles and six memory cycle times; it takes one extra memory cycle to wake up.

But some of my tests seem to suggest that SKIP & WAIT are actually identical in timing. Am I doing something wrong here or have I misunderstood the SKIP timing?

[roondar]

To answer my own question: the timing of WAIT & SKIP in the HRM are in fact accurate.

However, PAL Amiga's have scanlines that aren't all the same length. This can impact timing of certain effects and is something I didn't know until very recently.

[ross]

Originally Posted by roondar

However, PAL Amiga's have scanlines that aren't all the same length. This can impact timing of certain effects and is something I didn't know until very recently.

Hi roondar, what do you mean by this?
NTSC lines can be long or short but PAL lines are only short...

[roondar]

Yeah, that wasn't very clear. Mostly because I wrote something that's wrong. It is indeed NTSC where lines alternate in length, not PAL. Sorry, my mistake.
Edit: it's because of this bit in the HRM. It always confuses me:

Originally Posted by HRM

All lines are not the same length in NTSC. Every other line is a long line (228 color clocks , 0-$E3), with the others being 227 color clocks long. In PAL, they are all 227 long. The display sees all these lines as 227 1/2 color clocks long, while the Copper sees alternating long and short lines.

Somehow I always read that as saying that PAL lines alternate.

Believe it or not, but what I was getting at was that scanlines get 113 slots for Copper instructions and that this is not a very useful amount if you want to divide it by 2. But because I had in my head that PAL alternated in length, I remembered that when I wrote this. Busy day, mistakes happen.

[ross]

Originally Posted by roondar

it's because of this bit in the HRM. It always confuses me:

That HRM sentence is badly written and can only be confusing

Believe it or not, but what I was getting at was that scanlines get 113 slots for Copper instructions and that this is not a very useful amount if you want to divide it by 2.

Two possibility here:
- spread 113 CMOVE over two lines (but say goodbye to COLOR00 sync) [useful for BPLCON4 refill on AGA]
- use 55 CMOVE + 1 CWAIT per line
(as long as you don't have copper cycles stolen elsewhere)

[Photon]

It seems you have found the answer, but if someone ends up here looking for copper instruction timings, I wrote an article on Coppershade about MOVE and WAIT a while ago.