Would you support a new Unified Opensource RTG standart?

[Gulliver]

Dumb question:

Is it technically feasable for this new rtg system to provide backwards compatibility for both Picasso96 and CybergraphX at the driver level? So that there is a chance we can still use old drivers until new ones arrive.

[bloodline]

I rather like disscusions like these, but I do have to point out that I can really see any future for this.

68k is not going to get any new hardware*... So the C64 and P96 systems support everything that is available. Both are basically compatible... Solid and in use... This idea is just reinventing the wheel without any tangible benefits.

*I really can't see any PCI-E boards for the Amiga being either comercially or technically viable...

There could be some merit in porting AROS's gallium gfx thing to MOS and AOS4... maybe?

[Tumbleweed]

I rather like disscusions like these, but I do have to point out that I can really see any future for this.

68k is not going to get any new hardware*... So the C64 and P96 systems support everything that is available. Both are basically compatible... Solid and in use... This idea is just reinventing the wheel without any tangible benefits.

*I really can't see any PCI-E boards for the Amiga being either comercially or technically viable...

There could be some merit in porting AROS's gallium gfx thing to MOS and AOS4... maybe?

There may be some benefit for those with PCI busboards if it allows them to use a broader range of GFX boards. I agree though that there is limited upside for those with classic Amiga's but there are some benefits - faster raster and greater stability (bug fixes). Also the poll seems to be pointing in favour of developing a unified open RTG standard.

Weed

[arnljot]

I agree that new hardware for classic amigas is unlikely. But not impossible. One new hardware rumor is the Indivision A1200 mkII. But I expect Jens has all the SDKs he needs.

However, a more tangible discussion recently was the Virge emmulation discussion in FPGA Arcade (Minimig AGA thread). It would open up much better possabilities if such challenges could be sorted out in the RTG system.

Thank you Piru for your response. I think I now understand some of the concern that you have.

The idea we're tossing around here is however not to create a third way, but a common way. For software, clients of the RTG system if you will, it would be vital that API compatability is retained. Or else the users wouldn't want it as it would render too much software allready compiled and no longer maintained useless.

As it stands in the moment, it seems that a new RTG initiative would have to specify a new API which would be P96 and CGX compatible, not so hard in it self as P96 currently to some extent does this.

The hard thing, like you and others have pointed out is the driver situation. And everything in interfacing with the hardware as that is what the RTG layer does.

As someone asked here, is it possible to "mimic" the other RTG systems, and retain compatability with existing drivers? I think that would be hard, and cost way more than you'd like to put into such a project, even though the project we're talking about here would be huge.

Please remember that one huge user benefit would be that this new RTG system should also exist in a minimal footprint version, which you could have in Kickstart so that RTG is available from power on. Much like linux and other OS' today have their VESA modes.

[bloodline]

But someone is going to have to write drivers for the new system that support the cards that are already supported by either P96 or CV64... No one is going to do that...

[kolla]

I think it's unrealistic to expect someone to write drivers for all the existing amiga graphics cards.

Why would it be someone? Typically there might be someone who will write a driver for his/her card. However, without SDK there's almost a guatantee that there will be no new drivers at all - ever..

Writing graphics card drivers is quite tricky and requires special kind of skillset. There aren't many around who would be capable, not to mention willing.

And again, by not providing the SDK needed, it is almost guaranteed that noone will bother. God forbid any new people who might stumble into this freakshow of a plattform and who might be interested in writing drivers.

This would basically lead to a situation that instead of two systems that support different set of hardware we'd have three. Instead of two different systems with slightly different quirks, we'd have three systems with different quirks.

Ofcourse, but one of them at least having the advantage that it can be upgraded and fixed on - that's what this is all about.

New driver system would also mean that 3d drivers would need to be written from scratch.

No it doesnt. 3D on m68k is not that much of an issue in the first place.

I don't have an easy alternate solution to offer, unfortunately.

I just wonder what people like you will do when MorphOS eventually is in the same situation

[kolla]

Just top put things in perspective - imagine if AHI was closed up like P96 and CGX, and what that would mean.

[kolla]

But someone is going to have to write drivers for the new system that support the cards that are already supported by either P96 or CV64... No one is going to do that...

What do you know about that? As I wrote before, the typical scenario is not someone writing all the drivers to replace CGX and P96 all over, it's someone writing a driver for his/her card, and to be able to tweak and test out things with his/her own hardware. Ofcourse the best would be if one could do this using existing SDKs for CGX and P96, since they no doubt are superior to anything a new RTG would be able offer in years. However, today this is not really possible without reverse engineering and and a lot of guesswork.

[Golem!dk]

Just top put things in perspective - imagine if AHI was closed up like P96 and CGX, and what that would mean.

Not much if you consider who wrote the existing drivers.

[arnljot]

Not much if you consider who wrote the existing drivers.

But it's AHIs open nature who enabled other developers besides Martin to do so. That is part of the argument here.

[wawrzon]

perhaps the existing ahi drivers just work perfectly, dont recall who actually wrote the pci ones, but all in all i think soundcards are not that a problem atm and if they were a driver could be written.

but what concerns other things like rtg and especially warp3d there is an unnecessary amount of work people (i mean actually matthey) are forcred to invest in fixing it after disassembling. while if it was open maybe amiga might have working, as much as possible up to date, 3d accel solution and not a pile of rubbish what it is atm.

[Golem!dk]

But it's AHIs open nature who enabled other developers besides Martin to do so.

That isn't the impression I got.

[arnljot]

If you look here
http://arp2.berlios.de/ahi/#downloads

Then you can see that the driver developer SDK is freely downloadable.

That means that openpci guys have made ahi drivers, elbox has made mediator pci drivers, and grex dudes have made their drivers.

That's the way we would like it to be for the rtg systems too.

[Golem!dk]

That's the way we would like it to be for the rtg systems too.

I'm well aware of what is available, but the thing is... look again at who actually wrote the existing drives, somehow I think they would have gotten access even if it was closed source.

[wawrzon]

@golem: once again, what ahi drivers you immediately require that are missing? i think the ahi job has been done quite decently, what isnt true for rtg anymore. there is a lot of things that might be improved or fixed if the standards were open source. i dont say they would be fixed but they might. such an example is inexistant support for w3d on mediator radeon that pretty much rules this card out against voodoo, which drivers still have to be fixed. the best supported to date are i think b- and c- visions. how many years ago the drivers were written exactly. i dont recall.

and i think we seriously have to stop thinking of amiga software on profit basis. it is a hobby, so it is either open source or nothing. the consumer base is too small to generate income. not so hardware wise since it cant be done in the kitchen. i dont know exactly if this is sad , but it is the truth.

[Golem!dk]

@golem: once again, what ahi drivers you immediately require that are missing?

The ones I mentioned obviously.

and i think we seriously have to stop thinking of amiga software on profit basis. it is a hobby, so it is either open source or nothing. the consumer base is too small to generate income. not so hardware wise since it cant be done in the kitchen. i dont know exactly if this is sad , but it is the truth.

Good thing we have AROS.