I admit, I know nothing about AmigaOS, so I don't know the limitations. I would assume that OpenGL would probably be an issue? Currently, xoreos runs on my own naive fixed-function OpenGL 1.2 (+ some extensions) code, but it's being overhauled to utilize a shader pipeline. The target would be OpenGL 3.2, with an OpenGL 2.1 fallback path. Would that work on your hardware and OS?
There is no shader support for older cards like the Radeon R200 based cards which are still the most commonly used cards on OS4. They are supported by the native Warp3D API and there's an MiniGL implementation. Hans De Ruiter is probably the best person to ask questions about what functions are implemented.
Hans has also created a new shader based API for A-EON called Warp3D Nova, currently certain Radeon HD7000, R5 and R9 cards are supported. An OpenGL ES2 wrapper for Nova is being developed by Daniel Müßener(Daytona).
There is no shader support for older cards like the Radeon R200 based cards which are still the most commonly used cards on OS4.
R200 based systems are probably going to be too slow for those games anyway, Neverwinter nights (my favourite) was abysmal on a less than 1ghz mac, so sam440 & flex, micro, xe & se are going to be pushed to their limits on CPU, add the memory constraints on some of them and it's looking bad.
Personally I would love to see Neverwinter Nights on my X1000 & R7770HD using Nova
Amiga user since 1985 AOS4, A-EON, IBrowse & Alinea Betatester
I said no, because development is still in its early stages
Okay. :)
@Severin Even my laptop with an RV350 back in the day was only okay-ish with NWN, yeah. KotOR and KotOR2 were basically unplayable.
@Rob Ah, so targetting OpenGL ES instead of desktop GL would be the way to go? Right now, xoreos doesn't support that (and GLEW, which we're using to mangle GL extensions, doesn't do ES either, unfortunately), but it's on the long TODO list.
@tlosm Please do tell me (or better yet, file a bug report on the GitHub issue tracker) if it doesn't work. :)
Especially if you're running in big endian mode. It should still work (since the file reading code is based on ScummVM's), but I have only tested xoreos on little endian machines so far.
Even my laptop with an RV350 back in the day was only okay-ish with NWN, yeah. KotOR and KotOR2 were basically unplayable.
FYI: The machines with R200 cards (SE, XE, Micro) are G3 or G4 800Mhz and usually 512MB of ram. Sam flex's are slower (mine is 733Mhz) but have 1GB ram. The basic sam440 is even slower.
The ones that would reach a minimum spec are X1000, X5000, & Sam460, The Tabor might just manage it but it's classed as entry level so not sure. The Tabor and X5000 are not available to the public yet anyway.
Amiga user since 1985 AOS4, A-EON, IBrowse & Alinea Betatester
i strongly suggest to make a warp3d first and nova on 6xxx. i know users that have this cards. one is me. the si are not supported on linuxppc or better there is a really huge endianess there ... glamouregl crash when invoched. radeon 6xxxx are a good cards that make x1000,x5000,tabor and 460 have 3d working on os4 and 3d working on linuxppc. and like me i hope in future many of us will use only their ng as the only one computer . note 3y ago was sold many 6xxx because was annonced the compatibility. i love my amiga
ps: i have a si too but not used because working only on os4 and good
I did the same bought an 6850 card, wasnt cheap that time and later i had to get a 7xxx card. I can guess many people did the same as announced. Today still nothing. Pretty bad
i strongly suggest to make a warp3d first and nova on 6xxx...
Yes, I know that the Radeon HD 6xxx situation sucks. However, realistically I only have the time to work on drivers for one GPU series. Normally there are multiple people working on drivers for one series, and not the other way around.
I'd welcome drivers for other GPUs, but other people need to step up and make it happen.
As for the Linux endianness issue, you could ask the Linux porting team about running it in little-endian mode. That's what IBM's doing.
@Marko It definitely gives me some motivation. I'll get back to it next month to figure out what breaks the non-cycle based SID emulation and small fragment sizes.
This is just like television, only you can see much further.