Should game devs have to support new hardware?

[grond]

Originally Posted by mcgeezer

I'm talking about the OS routines that Commodore supply to access the hardware, I'm pretty sure you (and others) know what I meant

Yes, and I pointed out that coding in compliance with Commodore guidelines does NOT mean that you have to use the OS for everything.

[grond]

Originally Posted by alpine9000

The "OS off" version drops no frames during the test.

The "system friendly" version dropped 5% of frames. Note: many scenes in the test sequence are not "busy" so there is little chance of a frame being dropped in these cases. So during busy times in the game, the frames are being dropped much more than 5% of the time. It was quite noticeable slowdown.

So based on these results for me, it's seems that on an A500, the OS running in the background has a significant negative performance impact.

If we run the same test on a standard A1200 with no fast ram, we are back to no dropped frames.

Very well done!

[Tigerskunk]

Reading through the thread up to now, and these are my 5 eurocents (and I hope I dont sound bitter, because it's a fast way to go there with this topic )...

* The Amiga is a super interesting system, but a chore to code for compared to any other system I worked on otherwise (which includes the C64, NeoGeo, Atari VCS, Vectrex, and a bit of Spectrum Next now). This already prevents newcomers to the platform to have a dip into the cold water that is Amiga coding with its Blitter save/restore shit, managing scrolling through double and triple buffer bitplanes, while synchronizing copper lists for effects that other platforms do without this added complexity.

* "System Setup" does not only include different CPUs, OS versions and if OCS and AGA, but also the medium how to deliver. Floppy (OS or trackloader?), hardware install or whdLoad, CD32, .adf. I had someone complaining he was not able to run the .adf in some of weird launcher that's running on a native Amiga, so he won't buy the game. Yeah, cool. People really expect you to support ALL of these.

* If our Ross@EAB here didn't have a look into the whole affair, my game would certainly not run on as many configs as it does now. How many people do we have here with that kind of knowledge? I am certainly not interested in learning this stuff, sorry. I wanna create games with the least amount of system knowledge needed. And a huge thank you to Ross again for doing these things for us lamers here.

* that said, in future I won't be doing this again if I create another Amiga game. Inviyya was my first endeavor on the Amiga, and I kind of wanted to make everyone as happy as possible. Which I guess took me a LOT time and motivation to make that happen. I am sure the game would have been finished a lot earlier without this.

In the future, I won't be creating a game for THE AMIGA anymore.
I will create a game for the 512k/512k Amiga 500 on a floppy disk, and people can try out if it runs on their system by checking out the demo.

And that's not because I hate those other machines, but because my nostalgy and the reason why I am doing this is about grinding the bare metal on that magical A500 I owned in my childhood that meant so much to me. I am a huge OCS Amiga fan foremost, and the Amiga 500 and also the A1000 will always be a magical experience for me.

I never experienced an AGA machine nor accelerator cards back then, and when you played a game, it was from floppy disk.

And this love for the A500 and A1000 is why I am coding games, and that's where I will be doing this for in the future..

Your mileage may vary, but so does mine...
Peace out, mates...

[grond]

Originally Posted by dreadnought

I can't really fault a bedroom coder for not knowing about, or adhering to, all the standards. They are, by definition, unpaid amateurs who do things for fun.

Yes, of course. But should the bedroom coder just decide to remain ignorant and even leave it to other people to clean up his work ("leave it to the WHDload team") or should he strive to learn how to do things properly (even if he fails)?

[Tigerskunk]

Originally Posted by grond

Yes, of course. But should the bedroom coder just decide to remain ignorant and even leave it to other people to clean up his work ("leave it to the WHDload team") or should he strive to learn how to do things properly (even if he fails)?

yes. he should.
(if the decision is either that the game is running on one machine perfectly and maybe not on the others, or leaving the Amiga as a viable platform completely)

[phx]

Originally Posted by dreadnought

I can't really fault a bedroom coder for not knowing about, or adhering to, all the standards.

Depends on how long this bedroom coder is working with Amigas.

We are all happy about new coders who are interested in programming the Amiga. Especially me, who could only dream about this positive development over the last years, after releasing all source codes and assets of my new game in 2013 in the hope to encourage more productions.

But we do not help new coders when we say: "it's ok when you do your code like in 1987 and as long as it works sometimes on some A500 configurations it will be great!".

The situation is different than 30 years ago. Today you have lots of information everywhere on the net. And you have forums, like EAB, with many experienced coders who are just happy to share their knowledge, and are literally waiting for questions every day.

In fact it will be difficult for new coders to miss the methods how to deal with different Amiga configurations and accelerators, when they are looking for information. Usually only few lines of code and avoiding common mistakes on timings and self-modifying code are sufficient to make it work on 90% of all models out there. At least you should try.

Nobody will blame a programmer for bad code, I hope. It usually has reasons. But many people will notice and remember it.

[dreadnought]

Originally Posted by grond

Yes, of course. But should the bedroom coder just decide to remain ignorant and even leave it to other people to clean up his work ("leave it to the WHDload team") or should he strive to learn how to do things properly (even if he fails)?

In an ideal scenario, of course he should. But our lives are far from ideal, and people and their circumstances vary wildly too. The bottom line is, it's a hobby, so I don't think we should be overtly critical of those who do things their own way.

I did not have that "WHDload team" line in mind", btw. Like I said in my first post, a coder can fix or adapt things later, should they want to, and there indeed is a big community which can help. But if he moves on to other things, that's fine too.

[mcgeezer]

Originally Posted by grond

Yes, of course. But should the bedroom coder just decide to remain ignorant and even leave it to other people to clean up his work ("leave it to the WHDload team") or should he strive to learn how to do things properly (even if he fails)?

Can I just ask, have you ever written an Amiga computer game in assembler?

[Tigerskunk]

I think people here don't realize that this is a question of "will I go on with Amiga game dev or not" for many of us people.

Creating, supporting and testing for every Amiga config and medium is simply not sustainable for bedroom coders.

Compare that to that one config and medium you have on almost any other retro platform.

[roondar]

Originally Posted by Steril707

I think people here don't realize that this is a question of "will I go on with Amiga game dev or not" for many of us people.

Creating, supporting and testing for every Amiga config and medium is simply not sustainable for bedroom coders.

Compare that to that one config and medium you have on almost any other retro platform.

I'm hopeful we'll be able to create some sort of basic framework that does the important things and refer people to that as needed. Many of the things that need to be done are not that complicated once you know them, but knowing them is not always a given. Support for new developers in the Amiga community isn't always great and it'd be nice if we as a community could try and fix that.

However, I'm 100% certain that it's functionally impossible to create a framework that is guaranteed to work on all Amiga configs (including the extremes you mentioned).

[Tigerskunk]

Originally Posted by roondar

However, I'm 100% certain that it's functionally impossible to create a framework that is guaranteed to work on all Amiga configs (including the extremes you mentioned).

You still need to test on vastly different configs.
I usually test on A500/A1000/1.3 and A1200/3.1, yet I have seen weird shit happening on other people's rigs all the time even with all the stuff Ross improved in my code. (damn, we actually have had a bug, where we (me and other testers) FEEL it was behaving differently on WinUAE and FS-UAE, as crazy as that sounds).

Plus then you still have the DREADED CD32 version to get right as well...

And still, with all these things running decently well, you will still have enough people telling you you need to go with the times and support their weirdo medium or config.

Nope, not sustainable if you are not a company.
And then we are entering "development needs to be profitable and not only fun" territory, where the whole thing goes completely bonkers, since that each and every game would cost a couple of hundred euros per box for those few copies sold.

My solution to this in the future is just doing what gives me fun.
And that is creating games for one platform and medium, without the need for ensuring the game runs on twenty other possible CPUs,RAM setups, Kickstart versions and hd-install/floppy/USBStick/CD /Laserdiscs, or your grandmom's gramophone...

Either people then follow and say fuck it, I have ten different Amigas anyway, one stock A500 is for gaming from now on, or it's NeoGeo, C64 or Megadrive for me as a game dev in the future, and I will just be happy being an Amiga fan as I was before and enjoy other people's works from now on.

I am sure Grond will deliver many amazing games for me to enjoy, if he has such a strong opinion on what bedroom coders should do or not do...

[grond]

Originally Posted by mcgeezer

Can I just ask, have you ever written an Amiga computer game in assembler?

Why would that be of interest? Because you that you have can then just easily discard my point of view?

As a matter of fact I have written quite a bit of asm code on the Amiga for demo effects and more. I also did an OS-friendly jump-and-run tech demo in the mid-90s completely in asm.

[DamienD]

Originally Posted by grond

As a matter of fact I have written quite a bit of asm code on the Amiga for demo effects and more. I also did an OS-friendly jump-and-run tech demo in the mid-90s completely in asm.

So, can we assume then that both of the above run on all of today's many different Amiga models / hardware setups; and your beloved Vampire?

[freehand]

Even thou there is a shortage of amiga games and developers apparently there is a certain criteria needed when hitting the hardware in 2021

[Tigerskunk]

Originally Posted by freehand

Even thou there is a shortage of amiga games and developers apparently there is a certain criteria needed when hitting the hardware in 2021

and that is?

[freehand]

Originally Posted by Steril707

and that is?

It needs to run on every combination under the sun.

[mcgeezer]

Originally Posted by grond

Why would that be of interest? Because you that you have can then just easily discard my point of view?

As a matter of fact I have written quite a bit of asm code on the Amiga for demo effects and more. I also did an OS-friendly jump-and-run tech demo in the mid-90s completely in asm.

I see, so you haven't written a game then on the Amiga.

Your point of view is to call other coders lazy and ignorant when it's pretty clear now that you don't have the credentials yourself to judge them with those kind of words. Therefore, I don't see anything to discard.

Lots of programmers can make stuff in assembler and have the sanctuary of the OS to shield them from making mistakes and make their stuff compatible with all Amiga's and the hybrids - that is up to them, totally their choice.

But I for one enjoy hitting the hardware directly because I like the overall control I have over the system and the performance reward I get for the effort I put in, this is one of the great things about programming on the Amiga. Yes, there are some Red lines that should not be crossed but overall I'm now of the opinion that dev's target what they want and if it doesn't work on higher spec machines then tough (for the player and his/her hybrid Amiga).

[Tigerskunk]

Originally Posted by freehand

It needs to run on every combination under the sun.

At the software company I work at we are telling our customers to use a certain hardware so use our software with, because we cannot be arsed with ensuring this works with every tablet or mobile phone under the sun. And we even have dedicated QA people testing full time.

But somehow, on the Amiga that's supposed to work out for hardware that spans 35 years now being tacked unto one another with software created by a dude doing ASM on 36 year old chipsets in his free time.

Yeah, perfectly viable, isn't it...

[grond]

Anyone is free to write bad code but at the same time cannot expect to only hear praise for it just because they did it for free and/or in their spare time. Bad code will be bad code.

I never said your code absolutely must run on any Amiga under the sun. I personally never code for anything below kickstart 3.0 (perhaps kick 2.0 if it doesn't need much). But my code graciously quits if you run it on anything below that.

If your code crashes just because it runs on a fast CPU (killing the OS means this is about the only possible source of the crash), then chances are that you made one of the mistakes that were common in the 80s, seemed to have died out by 1990 and now make a triumphant return in your software. Unless your code squeezes the last bit out of the OCS base configuration (I understand the attractiveness of targeting that configuration), you could have done better.

[Tigerskunk]

Originally Posted by grond

Anyone is free to write bad code but at the same time cannot expect to only hear praise for it just because they did it for free and/or in their spare time. Bad code will be bad code.

agreed

Originally Posted by grond

I never said your code absolutely must run on any Amiga under the sun. I personally never code for anything below kickstart 3.0 (perhaps kick 2.0 if it doesn't need much). But my code graciously quits if you run it on anything below that.

So you are already doing what we are talking about.

Originally Posted by grond

If your code crashes just because it runs on a fast CPU (killing the OS means this is about the only possible source of the crash), then chances are that you made one of the mistakes that were common in the 80s, seemed to have died out by 1990 and now make a triumphant return in your software.

Yeah, because you can talk about EVERY hardware config out there that is.

Originally Posted by grond

Unless your code squeezes the last bit out of the OCS base configuration (I understand the attractiveness of targeting that configuration), you could have done better.

That's what Graeme and I are doing with our games, I'd say.

On how many systems have you tested your "gracious" software?
Also, where can we download it?

Plus. if it's OS dependent, how are you even in a position to comment on this? Also furthered by the fact you never finished a project as it seems.