The X68000 games porting experiment

[Anima]

OK bad news I'm afraid Anima.

Damn...

Anyway, thanks for testing. I was expecting a different result but I am not really surprised. Are you able to obtain some infos using a debugger? Probably I need to add a simple bus-/address-error handler.

[EvilFranky]

If you can give me instructions on how to do it then I'll help in any way I can.

[Xerus]

I can't get the game to run at all under 68060 mode. I tried like I did when I made that video a couple of weeks ago, so my method hasn't changed.

68060 + MINT = Detected as 060 with FastRAM, hit space then just a black screen.
68060 + TOS = Detected as 060 with FastRAM, hit space then just the background image, nothing else happens.
68030 + TOS = Detected as 030, hit space and the game runs as expected.

Same tests, same results with my Falcon.
I deactivated the cache and no change either.
Keep up the fight Anima

[Anima]

A simple way to accelerate this on the SuperVidel would be to just add 0xA000000 to your framebuffer pointer. This means you'll be rendering graphics directly to the snooped STRAM area of the SuperVidel VRAM. Make sure to check for the SupV-cookie first.

Thanks for the info. Using the SuperVidel seems to be a good idea if it's that easy.
So far there should be no caching problems at all since the sprites are not compiled as code in this version. Still wondering why it fails.

[Anima]

I can't get the game to run at all under 68060 mode. I tried like I did when I made that video a couple of weeks ago, so my method hasn't changed.

68060 + MINT = Detected as 060 with FastRAM, hit space then just a black screen.
68060 + TOS = Detected as 060 with FastRAM, hit space then just the background image, nothing else happens.
68030 + TOS = Detected as 030, hit space and the game runs as expected.

Same tests, same results with my Falcon.
I deactivated the cache and no change either.
Keep up the fight Anima

Thanks. Could someone please check what will happen when the program flags are changed, i.e. that the program will not be loaded into the Fast-RAM (e.g. TOS program flags = 0)!? The whole sprite routine is not different to the previous version but the screen memory is now being allocated in ST-RAM (due to the Videl DMA access). This behavior depends of course on the Fast-RAM detection. Probably I should use the appropriate XBIOS functions for this.

[EvilFranky]

Can't believe I didn't think about trying that earlier. I'll have a bash tomorrow mate.

[Xerus]

Thanks. Could someone please check what will happen when the program flags are changed, i.e. that the program will not be loaded into the Fast-RAM (e.g. TOS program flags = 0)!? The whole sprite routine is not different to the previous version but the screen memory is now being allocated in ST-RAM (due to the Videl DMA access). This behavior depends of course on the Fast-RAM detection. Probably I should use the appropriate XBIOS functions for this.

Sorry but it changes nothing.

[Anima]

Thanks. Could someone please check what will happen when the program flags are changed, i.e. that the program will not be loaded into the Fast-RAM (e.g. TOS program flags = 0)!?

Sorry but it changes nothing.

Thanks to all for your reports. I think I'll prepare a special version with disabled sprites. I would like to see if that will run on your machine (if that's ok for you). That should narrow down the problem.

I also checked the binary on Mint using ARAnyM MMU with activated memory protection and I was so happy to see it running but that's worth nothing.

[Xerus]

No problem, as EvilFranky if I can give a little help it will be a pleasure.

[Anima]

I have prepared a test version which has the sprite drawing disabled. Can you please try this version just to see if it will run? You should see the background and seconds later the main menu.

[Xerus]

Always the same problem, there is a freeze on the background.
I can't see the main menu.

[ctirad]

I think the problem is the MMU. Aranym emulates the 030 one, so it is no surprise it works here. The 060 is quite different, AFAIK.

[Anima]

I think I've found the cause of the freezing problem. The program simply chokes on the graphics drawing. So without any appropriate drawing optimizations this will never work (at least not trying to achieve 50/60 frames per second). The ST-RAM is definitely to slow for the current simple copy approach.

However, I've prepared another version which should work on SuperVidel equipped machines. Can someone please try it and report the result? Also is the SuperVidel detected properly (the program looks for the SupV cookie)?

Edit: please use the game data from the lastest test release.

[shoggoth]

If it's still not fast enough, you'd have to use the SuperBlitter, but in such case you'll have to move all your sprite data to VRAM.

[jury]

However, I've prepared another version which should work on SuperVidel equipped machines. Can someone please try it and report the result? Also is the SuperVidel detected properly (the program looks for the SupV cookie)?

I confirm that this version does not freeze at the beginning, so its possible to start it now, but it freezes quite often later on.
If it goes for performance, it's really fast.
And if it goes for the last question ( regarding SV cookie ) I do not understand what to check.
Test machine: F030 + CT63 ( @93MHz ) + SV

[shoggoth]

Also is the SuperVidel detected properly (the program looks for the SupV cookie)?

I expect this to work. It's the official way of detecting the card. No other probing should be made.

[Anima]

I confirm that this version does not freeze at the beginning, so its possible to start it now, but it freezes quite often later on.
If it goes for performance, it's really fast.
And if it goes for the last question ( regarding SV cookie ) I do not understand what to check.
Test machine: F030 + CT63 ( @93MHz ) + SV

Thanks for this report.

The sprite drawing code is really really slow so there's much room for improving the speed. At least the initialization is correct. Good to know.

Also is the SuperVidel detected properly (the program looks for the SupV cookie)?

I expect this to work. It's the official way of detecting the card. No other probing should be made.

My question was misleading, so sorry about that. I just wanted to know if the program tells you by displaying an appropriate text on the "detected machine" output at start when a SuperVidel is installed.

The next step is to implement the sprite compiling and an improved display routine for non SuperVidel users. Besides some glitches and the missing music the game considered complete then.

Damn, it's so difficult to post a message here... definitely harder than landing a probe on a comet.

[EvilFranky]

The sprite drawing code is really really slow so there's much room for improving the speed. At least the initialization is correct. Good to know.

The next step is to implement the sprite compiling and an improved display routine for non SuperVidel users. Besides some glitches and the missing music the game considered complete then.

Damn, it's so difficult to post a message here... definitely harder than landing a probe on a comet.

Hahaha

Hopefully some talented musician may come along and re-compose the music using AceTracker

[Anima]

Just a short note: this project is not dead.

Still no new improvements about ChoRenSha running on real machines but I have some ideas how to speed up the sprite drawing on the stock Falcon030. The main problem is that the current "simple" sprite engine (clear screen and draw sprites on ST-RAM) will choke any machine and so it's not playable on them. That's the reason why it works best on Aranym yet.

All I can say so far is: DSP to the rescue! Doug has had some ideas about utilizing the DSP for the NeoGeo project and my idea is going in that direction. So please stay tuned...

[calimero]

plz can you share technical side?

or, at least, please share details after you implement them!

[qq1975b]

cool!

[dml]

Awesome.

I was literally ready to post asking where you had disappeared to This thing is way too good to fade.

[EvilFranky]

Oooo, exciting stuff!! Looking forward to more details.

[Scarlettkitten]

Fantastic news

[nemodhs]

Great news. Keep us posted. Following the journey is part of the fun.