»

[balrogsoft]

Hi, I'm working on two projects that i want to make for mobile phones, all related with Amiga, first i want to explain that i work on a videogame company programming games for mobile phones, i know j2me very well, and i made some bug corrections on symbian projects on the company i work.

One project is a Xtreme racing clone for mobile phones with java, and with bluetooth options to play with other users, i made the raycasting engine to drawing floor, and it is working now, on my Nokia 6600 it gets 23-24 fps on a small screen of 128x128 (this mobile have a bigger screen, 176x208).

The other project is an Amiga 500 Emulator for Symbian, it will be hard, i'm new to C++, but as i said i made some bugs corrections on symbian projects, i will not port any existing emulator to Symbian, i want to make optimized for mobiles, and i'm not soo good with C++ to understand a big project like uae to port, i prefer to start from 0. I'm searching any kind of info, i have three books: The amiga hardware reference, one of 68000 processors (i don't remember the name), and a Spanish book of Amiga hardware. With this books i think that i have most of the information i need, but if someone have other documents like 68000 processor opcodes, unknow custom chips functions, and any other info, please send me anything interesing ...

Bye!

[InTheSand]

Hi,

A great idea, but would a mobile phone CPU have enough grunt to emulate an Amiga at a reasonable speed?

As far as I'm aware, Symbian-based phones use ARM-based CPUs in the 150-200Mhz range and I'm not sure if this would be enough, even without the overhead of the screen scaling to fit the low-res phone screens, plus mixing the Amiga's audio down to suit the phone hardware...

I wish you luck anyway!

- Ali

[bloodline]

My XScale based PDA at 400Mhz just about emulates an A500... I would dread to think what any slower CPU would be like

[balrogsoft]

I think that some tips could be used on a A500 emulation for Symbian, for slow devices it could be emulated without sound, sound usually takes a lot of cpu use on emulation, i have some emulators of 16 bits machines that work really nice without sound, i have a Megadrive emulator for my 6600 that have a arm 104 mhz processor, i know also that exist an AtariST on Symbian UIQ phones, Amiga hardware is more complex than a Megadrive, but i think that it could work...

[Piru]

@balrogsoft

UAE is C.

To be frank, you're facing an impossible task if you're going to figure out Amiga emulation all by yourself.

[balrogsoft]

Piru wrote:
@balrogsoft

UAE is C.

To be frank, you're facing an impossible task if you're going to figure out Amiga emulation all by yourself.

Yes, UAE is C and WinUAE is C++, but i must to code in C++ with Visual Studio and Symbian SDK. And i don't need to figure how Amiga emulation work, as i said, i have two Amiga hardware books, and a 68000 reference book. And i can use UAE code to find and solve specific problems, but i don't want to port directly UAE, because as i said my C++ knowledge is limited, i used more Java and other languages.

[bloodline]

Amiga emulation requires some really complex timing stuff... you would do well to start with the UAE source and work from there.

Let me introduce you to my new invention... I call it "The Wheel" ;-)

[Piru]

@balrogsoft

Yes, UAE is C and WinUAE is C++, but i must to code in C++ with Visual Studio and Symbian SDK.

I just won't believe that. Not being able to use C would be just insane and suicidal.

And i don't need to figure how Amiga emulation work, as i said, i have two Amiga hardware books, and a 68000 reference book.

I don't want to be rude, but is is just insane.

You will NEVER be able to write amiga emulation just with "two Amiga hardware books, and a 68000 reference book".

And i can use UAE code to find and solve specific problems

Even when using UAE for reference (unless if you end up copying UAE pretty much as-is, in which case your app is GPL) you're into several years of work at least.

but i don't want to port directly UAE, because as i said my C++ knowledge is limited, i used more Java and other languages.

But UAE is not C++, it's C. Even WinUAE is mostly C.

[balrogsoft]

Denis (who made a FPGA implementation of an A500) said that he had three Amiga books that have 90% of information need to implement Custom chips, i think that is a high percent of information to start the project, but i will not able to do, why?

I never said that i will finish in weeks, i said that it will be a hard task, it will me take months to have a first version running something... UAE was the only one Amiga emulator, then appeared Fellow, why to develop another Amiga emulator if exist UAE? maybe is funny, and you learn a lot more that porting an existing code.

I post this topic to request information and help, the only one thing that i got are critics, you will not be able, this is insane and suicidal.

Forget all i said... too many negative people here...

[Piru]

Denis (who made a FPGA implementation of an A500) said that he had three Amiga books that have 90% of information need to implement Custom chips, i think that is a high percent of information to start the project, but i will not able to do, why?

IMO that 90% is hugely overestimated (or misquoted). I'd say it's 10% or less of the information needed. The manuals explain how the hardware works, not how it's implemented. Even though you'd replicate every documented functionality 100%, it would still not work correctly. How everything relates to each other, what is undocumented or abused in various games, demos and apps, and figuring out how to get everything work right is the 90+%.

I post this topic to request information and help

The best you can get is UAE source.

the only one thing that i got are critics

And rightly so. Amiga emulation is just too much work to do from 0.

suicidal.

That comment was about Symbian, not you.

[balrogsoft]

Piru wrote:

IMO that 90% is hugely overestimated (or misquoted). I'd say it's 10% or less of the information needed. The manuals explain how the hardware works, not how it's implemented. Even though you'd replicate every documented functionality 100%, it would still not work correctly. How everything relates to each other, what is undocumented or abused in various games, demos and apps, and figuring out how to get everything work right is the 90+%.

You are wrong here, i wrote two emulators, one was very very simple, but i started also a GameBoy emulator, it start to show some PD roms, for example, you want to implement a MOVE instruction of a CPU, you don't need to know how is implemented, you must make it work in your emulated registers and emulate the behaviour not the implementation. And you can implement in a totally different way, if it finally makes the same work, your emulator will work. And why you think that only will be 10% if Denis have his FPGA working. Sorry if i can't explain better, but i'm Spanish and my English is not very good.

[Piru]

@balrogsoft
Amiga is a very complex system compared to many game consoles and other seemingly similar systems (say: Atari). The reason is the custom chips, and their complexity. The chips are interconnected with each other, and for example you can trigger various operations from within another (for example: copperlist can be used to control blitter, which in turn can generate new copperlist, etc).

Emulating all this in software is very hard, esp if you want the emulation to function properly even with the games, demos and apps that use the chipset's full potential (various tricks and hacks, undocumented and often unintended behaviour of the hardware, but found to be useful for generating various cool effects). Actual CPU emulation is trivial in comparision, you can write cpu emu in couple of days.

Proof of the complexity of the task is UAE: It took years and years to get it to even function relatively usable level, and still the emulation is improved and fixed in every release.

So, while I give full credit to Dennis for is great work on the Minimig project, I still have to wonder if the project would have ever existed without UAE/Fellow... My understanding is that these have been invaluable help for him. To quote Dennis:

A lot of the work I did was indeed some form of calculated guessing. Days of reading through the hardware reference manual and trying to find the out the logic behind things. Also, because Winuae and Winfellow already exist, I have the luxury of being able to look how they did it when I'm stuck. (which I did on some occasions)

Still, what Dennis has done is something I had very hard time believing at first. I was very sceptical.

So I guess I could be just plain wrong, maybe it's not that hard. Or maybe it is, and Dennis just kicks serious behind. I guess you should ask Dennis directly what he thinks about feasibility of your project and ignore me. :-)