Daimakaimura (Ghouls 'n Ghosts) for the Atari STE

[Anima]

Ok, it seems that the discussion is quite going to an expected direction. To calm down the audience I'll give you some facts:

Any MC68000 equipped machine with a clock lower than 12 MHz won't be able to get a "arcade perfect" port.

However, the reason to port this game to the Atari STE is to hope that the "best case" scenario in game is always less performance demanding than the "worst case". Keep in mind that the CPS1 CPU is almost completely busy handling the enemy "AI" even without having to care about drawing sprites and tiles. Especially cases with high sprite numbers you will notice a drop of frames per second on this port for sure. So in conclusion the only good target for any CPS1 port ist the Atari Falcon030 and I am not even talking about an "arcade perfect" but "gameplay perfect" port.

Some notes about the Amiga: no, I don't see any reason why the Amiga would have an advantage against the Atari STE. The main problem here is indeed the amount of RAM for graphics/sprites and simply the speed of the CPU. There's no reason to think that the custom chipset of the Amiga will give it an edge over the Atari STE since the Amiga is stuck in its limited design.

[AtariZoll]

I think the ST blitter acts like a 68000 on the bus. It has all the same pin outs. It might be possible to just fit it on a CPU breakout card on the Amiga. It was possible to retrofit to earliest ST models. That would give you full blitter access to fast ram for storing sprites. No idea if that would be feasible but it would be a really good hack!

Well, that may be good experiment, but I'm not sure that it will be really better than Amiga's blitter. And question is how much + logic it would require to work well.
Better would be to make new blitter chip in some CPLD logic, with new features, more address space and extra 3D help

[AtariZoll]

...
Any MC68000 equipped machine with a clock lower than 12 MHz won't be able to get a "arcade perfect" port.
However, the reason to port this game to the Atari STE is to hope that the "best case" scenario in game is always less performance demanding than the "worst case". Keep in mind that the CPS1 CPU is almost completely busy handling the enemy "AI" even without having to care about drawing sprites and tiles. Especially cases with high sprite numbers you will notice a drop of frames per second on this port for sure. So in conclusion the only good target for any CPS1 port ist the Atari Falcon030 and I am not even talking about an "arcade perfect" but "gameplay perfect" port. ...

I looked specs of CPS1 - never heard before about it. I don't think that CPU clock is main problem here. 50% more is not little indeed, but that console has a lot of graphic HW support, what does certainly much more and faster than our blitter can. And there is 192K of VRAM - what is for sure faster than STE RAM.
Console has only 64KB of "work" RAM for CPU, and that's actually enough for ROM based system.
Advantages of STE would be: more RAM, so can do some preparations of data for faster screen processing. Disk access - but that needs understanding of whole level data structure of original. And yes, less colors and lower res. are advantage too - that makes screen operations relative faster

Arcade perfect port would mean much more colors, bigger res. too. I doubt that even Falcon can do it. It has enough res. and colors, but is too slow in that mode for the task.

P.S. talking about running it on 68030 - I have pretty much experience in fixing 68000 code for - that includes stackframe related, and so called pipeline related ones. Later means that code will not work properly on 68030 if write only few bytes ahead in RAM - CPU will get old content, because it is already in CPU pipeline.

[rj1]

However, the reason to port this game to the Atari STE is to hope that the "best case" scenario in game is always less performance demanding than the "worst case". Keep in mind that the CPS1 CPU is almost completely busy handling the enemy "AI" even without having to care about drawing sprites and tiles. Especially cases with high sprite numbers you will notice a drop of frames per second on this port for sure. So in conclusion the only good target for any CPS1 port ist the Atari Falcon030 and I am not even talking about an "arcade perfect" but "gameplay perfect" port.

Very interesting project.
Will faster machines take advantage and give more FPS in your port?
Is it possible to add some kind of benchmarking attract mode with measured FPS rate at the end - I would like to try on our in-progress 030 STE card.

[dlfrsilver]

Ok, it seems that the discussion is quite going to an expected direction. To calm down the audience I'll give you some facts:

Any MC68000 equipped machine with a clock lower than 12 MHz won't be able to get a "arcade perfect" port.

Good to know.

Keep in mind that the CPS1 CPU is almost completely busy handling the enemy "AI" even without having to care about drawing sprites and tiles.

this is what one of the SF2 programmers said. He explains that basically, whenthe 68000 has finished to send the sprite list / tile list in dedicated ram, once the graphic chipset take the hand, the 68000 is working out the game logic during the VBL !

So it's not one task after the other, it's both tasks in parallel !

Especially cases with high sprite numbers you will notice a drop of frames per second on this port for sure. So in conclusion the only good target for any CPS1 port ist the Atari Falcon030 and I am not even talking about an "arcade perfect" but "gameplay perfect" port.

Well don't be too modest. The result you got is very good on a single 8mhz Atari STe. 4mb of ram, but great version !

Some notes about the Amiga: no, I don't see any reason why the Amiga would have an advantage against the Atari STE. The main problem here is indeed the amount of RAM for graphics/sprites and simply the speed of the CPU. There's no reason to think that the custom chipset of the Amiga will give it an edge over the Atari STE since the Amiga is stuck in its limited design.

This kind of game, for the Amiga would be better suited on 68020 or 68030 than regular Amiga 500 anyway. Easier to do.

[dhedberg]

Yes! Would be great if the port would take advantage of the Falcon030, at least the 030 at 16Mhz? Pretty please?

[Anima]

P.S. talking about running it on 68030 - I have pretty much experience in fixing 68000 code for - that includes stackframe related, and so called pipeline related ones. Later means that code will not work properly on 68030 if write only few bytes ahead in RAM - CPU will get old content, because it is already in CPU pipeline.

+

Very interesting project.
Will faster machines take advantage and give more FPS in your port?
Is it possible to add some kind of benchmarking attract mode with measured FPS rate at the end - I would like to try on our in-progress 030 STE card.

+

Yes! Would be great if the port would take advantage of the Falcon030, at least the 030 at 16Mhz? Pretty please?

From what I have seen so far there are only some stack frame manipulations (RTE return address) which prevents the original program from running on a MC68030+ machine. Hopefully this will be the only thing to fix to get it working on faster machines. Also the sprite routine doesn't work due to the optimized Blitter code. However, this is quite easy to change and would result in a lower memory requirement too. The definite goal is to have an appropriate Atari Falcon030 version as well.

[dlfrsilver]

yes i guess versions most possibly complete would be on Falcon and 1200 anyway.

It would be the same if you wanted to port Art Of Fighting / neogeo on Atari STe or Falcon (i have the full source code 68000 for this one ahaha).

[calimero]

yes i guess versions most possibly complete would be on Falcon and 1200 anyway.

It would be the same if you wanted to port Art Of Fighting / neogeo on Atari STe or Falcon (i have the full source code 68000 for this one ahaha).

Fatal Fury 2 neogeo on Falcon:
https://www.youtube.com/watch?v=Gataorz ... 5kZ6d9YzUA

[Estrayk]

Anima, BTW, Seems the Amiga community is a bit confused and surprised with your work.

http://eab.abime.net/showthread.php?t=87502

[Cyprian]

Estrayk, check user "AnimaInCorpore"

[Estrayk]

Estrayk, check user "AnimaInCorpore"

Ops! Owned!

[dlfrsilver]

Anima is with no doubt a great coder.

[Estrayk]

Anima is with no doubt a great coder.

Yes, they are still discussing if it would be possible to do what it does in an a500 and apparently everyone thinks they would need a 1200. it's fun to see how they look for a solution. Seems the 512k chip RAM is a big problem for them.

P.D. I love both platforms, don't judge me.

[dlfrsilver]

Anima is with no doubt a great coder.

Yes, they are still discussing if it would be possible to do what it does in an a500 and apparently everyone thinks they would need a 1200. it's fun to see how they look for a solution. Seems the 512k chip RAM is a big problem for them.

P.D. I love both platforms, don't judge me.

512 k of ram is also out of question on Atari ST or STe. Those coin-op games are the superior league, in size, complexity, program, logic, graphics, and so on.

to me 4mb of ram is a good start point. but yes, considering the difficulty to make a correct port on the A500 due to the graphics size and the need to blit the graphics, the 1200 is the choice machine.

[keops]

Anima: excellent. Do you have a page where you sum up the technical aspect of the project or your work in general ?

[Anima]

Anima: excellent. Do you have a page where you sum up the technical aspect of the project or your work in general ?

Sorry for the delay. The answer is: not yet but I'll give you details later on.
There are still many things to do in real life so everthing here is slower than expected.

[Anima]

WIP update: (vertical) scrolling.

https://www.youtube.com/watch?v=yDFkg9p ... e=youtu.be

[chlu600]

Amazing work! Looks better than expected, and of cource even better on real hardware not a compressed Youtube clip. Thumbs up!

[shoggoth]

W00t! Description says it's running on a plain 8MHz STE? Amazing

[Marakatti]

Wow! Fantastic job!

[wongck]

Wow... looks great... smooth... great great job

[jury]

Do I see what I see? Damn, it looks amazing so far.

[calimero]

It looks great! You use same palate for entire map (everything we saw in video)?

[Anima]

It looks great! You use same palate for entire map (everything we saw in video)?

Yes, so far only one palette (Dawnbringer) is being used for the whole game.