The online community for MiSTer FPGA enthusiasts
https://misterfpga.org/
Hi, I'm having severe scrolling issues on the shmup-sidescrolling levels on the mister fpga. The first and third shmup levels have these issues, the second one (the slow one) does not.
Game: https://sonicslothgames.itch.io/turrican2aga
I posted videos of the bugs:
I just tried the PAL floppy version, and can see no scrolling issues in the first level. My settings for minimig are A1200 with 2+8MB, no hdd and two floppy drives. Kickstart 3.1.
No.... It's not in the first level. Like I said, it's the shmup-levels.World. 3. Watch this video I recorded...
Like I said, it's a problem in level 3-1 and 3-3. 3-2 (the slow scrolling one) does not have it.
Thanks for the video and the explanation, I will check later if I see the same in 3-1.
hmmm heard about this version of turrican earlier this week and just looked up a youtube video of it. just me, or does the original OCS version look better? main sprite and the enemy sprites look... off.
There's something strange when running Turrican II Aga on Mister, as if the right part of the screen is slightly cut (the graphic border next to the life bar e.g.). Gotta check the 3rd level later.
You can choose the classic Amiga main sprite or the bulky DOS one before playing.
Didn't make this thread to discuss opinions on the game, just want to squeeze out the bugs 
@moroder,
I haven't tried yet this new game (and if I'll do, I'm not sure that I'll reach the world 3 
), but I'm just thinking to something which could potentially create this issue. Have you tried to removed the D-Cache in the A1200 system settings ?
Just reached the shooter section, same glitchy scrolling. Also, my D-Cache was set to OFF from the start. Sorry!
@moroder,
I can confirm the issue. The new turrican has a built-in trainer so it is easy to reach world 3. D-Cache was off in my case.
It would be nice to open an issue on Github.
aww, i was looking forward to trying this version out.
i have the PAL version downloaded on the MiSTer, just havent loaded it up yet
fingers crossed its a simple fix!
Yeah, I used your video too, hope it is not a problem.
Hi - just to post a quick follow up. There's some info from the dev on this topic explaining that the issue lies with MiSTers AGA core's copper timing. There's a more in depth explanation here:
http://eab.abime.net/showthread.php?t=112916&page=17
Great to know that it's not an issue with the game, hopefully identifying this helps lead to developments with the core that may eventually iron out these issues. I'll loop back on the Github page and add to the thread there. It sounds like a lot of people want to enjoy this game on MiSTer!
Dunno if it's the copper itself or that the 68020 is like... 10x faster than it should be! It's not even in the ballpark of a correct CPU speed, never mind cycle accuracy! Sounds like a lot of work but it would be nice from multiple angles for the 68020 to at least roughly approximate the OG speed. (and maybe add a 68020+ option for the current all-out speed demon)
If this gets fixed, I hope it also fixes the problems with Hybris (outlines of score and other slight problems), as those could very well be related to copper timing.
Yep, inaccurate AGA chipset timings cause a lot of minor issues here and there, while you’d be hard pressed to find a reason to have more accurate 020 timings.
FoxbatStargazer wrote: ↑Fri Feb 10, 2023 8:04 amDunno if it's the copper itself or that the 68020 is like... 10x faster than it should be!
SysInfo isn’t a good benchmarking tool, the calculations it does are not very representative of real world app performance (hence why it is named “SysInfo”) — it’s a tool to get a quick overview of what system you are dealing with and maybe get a ballpark performance number.
If you use a recognized “proper” benchmarking tool with a suite of tests like AIBB (Amiga Intuition-Based Benchmarks), it comes out roughly like this when you ask it for a summary:
MiSTer Minimig core is:
~5× the speed of an Amiga 1200
~2× the speed of an Amiga 3000/25
~1.5× the speed of an Amiga 4000/040 — but is slower than the A4000 on gfx and floating point
As stated in that thread, it’s not too far off the experience of using an A4000 or an A1200 with a 50MHz ’030 expansion. The disks are a lot faster than your typical setup at the time though, which makes it feel a lot snappier in use.
It's not even in the ballpark of a correct CPU speed, never mind cycle accuracy!
To my knowledge, there is no software, games or demos for the Amiga that requires or relies on 68020 cycle accuracy.
For the 68000, the case is of course different. :)
I also dont think its CPU speed with this as this game runs ok on the Vampire which is significantly faster..
Hi, any news about the scrolling issue?
Did you file a bug?
someone did awhile back
it was me, last december. No love for Turry
I don’t think there’s lack of love, but there are precious few people that would be able to fix the AGA chipset timings. Basically only a handful of people have intimate knowledge of undocumented AGA details (and it’s famously poorly documented), and another handful know their way around the minimig codebase. Not sure about how much these overlap.
one year later, still no fix
