I’d dontate to that too. Good thread
3D Hardware acceleration is a daunting task. There is a reason multiple people have tried over the years and failed.
- We aren’t Linux or FreeBSD and don’t use Xorg or wayland
- Most of Mesa is written around this ecosystem + DRM/DRI. Haiku’s OpenGL stack is unique.
- DRM/DRI is not easily cross platform (even though it is marketed as such)
- Haiku isn’t a priority platform for Mesa. The Mesa project has been awesome accepting our patches, however as soon as developers start to trip over our code, it will likely face calls to remove it. (especially since we don’t have an R1)
- Haiku Mesa + DRI/DRM, if it was completed, would need one or more people 100% dedicated to its grooming in Mesa 24/7/365
We currently have a software llvmpipe renderer which does get half-way decent performance and can play games like Quake 3 pretty easily. (the latest master Mesa has some build breakages due to “mesa progress” though i’ve been working through)
I started on a complete replacement for DRI/DRM in Mesa, but I would have to get the Mesa developers on board, and I honestly don’t have the skills to keep up.
So… that is why things are where they are today. I did make some big improvements to the situation by getting our OpenGL kit into Mesa, however, a lot of work needs to be done. (more work than a single contract will fix)
I would love to see hardware 3D on haiku. I loved BeOS and developed a synthesizer(SHS-101) app back in 2001. Then BE died. So I stopped. Meanwhile I have created a 3D engine which I would love to port to Haiku. Maybe I would even touch the synthesizer app again and finish it. Unfortunately SHS-101 can not be found in WWW any more. But I rescued the source code from a CD a while ago.
But due to I am into games development now, Haiku is no choice for me until 3D hardware acceleration is available.
But then again I guess Haiku is not yet ready for daily stuff. On the other side I would love to have haiku running on my arm devices like PineBook or ODROID XU 3 Lite.
I would spend some money for 3d hardware acceleration or ARM port.
Check out my 3D engine. It would be easy to port I guess and has a lot of stuff already.
Haiku! Go! Go! Go!
The guy who ran those sites got a flea up his rear and decided to take his ball and go home. In short, he got tired of helping the BeOS/Haiku community and left everyone in the lurch by taking away access to all the software he was hosting. Not a nice thing to do.
Indeed. Thats not nice! He could have transferred ownership to someone else somehow.
There are a list of software sites (besides Haikuports repo, of course):
Most of the Bebits and Haikuware can be found the sites detailed there.
I have one burning question to “kallisti5”. How much time/money would you need (If you feel able to get hardware accellerated 3D working on Haiku).
I’m back in the UK for my Christmas break. Once I’m back Stateside in January I’ll kick this off. I think there’s sufficient interest to give it a shot!
Seriously, this is a major league waste. Make a crowdfunding to support actual development in Haiku that’s propelling things forward but this entire 3D thing is a total waste of people’s money, time and what not else. If you ever worked with 3D graphics on a professional grade level you would know what kind of a beast 3D drivers are. This is not simply fiddling with some interrupts to switch graphics mode around. You’ve got to write a complex architecture to drive various kinds of hardware (heavily differing from one architecture to the next) on an OpenGL 4.x+ level… and this at a performance level which does not compare to a snail.
then no… just do nothing.
There’s already funding for Haiku Inc; you can donate one time or monthly.
Thanks for the patronising response, but yes, as a senior software developer with 20 years experience I do have knowledge of the gargantuan effort that is required to bring 3D acceleration to the OS. However, rather than just sitting doing nothing, I’m prepared to try and bring some kind of funding in and try and make some progress.
Hopefully, some of our other fans will have a more constructive attitude.
I’d much rather see proper 2D accelerated Intel/ATi-AMD/nVidia drivers that support correct widescreen resolutions, multiple displays, video overlay etc. before anything related to hardware 3D. That is worth donating money to. Personally I’m not that interested in 3d games and heavy apps that require GPU-acceleration at the moment, it’s not like we’ll get modern GPU-accelerated Photoshop or similar software anytime soon.
Nobody told it would be easy. BTW: As far as I know are Linux/FreeBSD open source drivers only GL 2.x or maximum 3.0. No core at all.
We and haiku are part of the free world. We can be into what ever we want and support specific aims of Haiku when we want. If someone kicks in for 3D drivers with maybe a funding specific for that, then I ask whats the problem?
This isn’t a bad point, actually. I think a better description for this project would be “modern graphics support” or something. The reason (in my own mind) I thought of acceleration is for things like compositing.
Do you have a demo video or a demo game for 3D Mini?
Thank you for the interest.
Yes. I have 3 game prototypes done with it. And the engine itself has some tests for GUI, Engine, Physics and tools like model editor, level editor and particle systems editor. I can show a lot of videos.
Problem is this week I have no WWW as I am currently switching my provider. Currently I am using my handy to connect to the world.
I will post some videos here when I have internet again.
Sorry, a bit off topic but…
I see you have Godot running also. Thats impressive!!!
I am currently working on a OpenFL project, OpenFL would be a nice match for Haiku too I guess.
I guess they have lots of games in the world too.