Documentation?
Tuma opened this issue · comments
Hi Bumblebee community!
I wanna ask if there is some documentation about how it works.
Some theory and explanation what is/should be done to install and make working the whole software.
I use Slackware and want to setup Bumblebee on my system.
I've done some RE but it's not enough for me to install everything needed.
So at first I'm asking someone to explain me how it really works.
Thanks.
I can give all what you want on that part, but I need an information : are you wanting to know how work Optimus exactly, what would be an ideal Bumblebee, or what we're doing now ?
I want to know how Bumblebee works (in few words), what it consist of,
what are the requirements and what are the steps to install.
Then I'll install it manually on my system and test it.
I'm experienced in Slackware good enough. So after I'll get understanding what is Bumblebee
and how it should be used we can add support for Slackware in Bumblebee install scripts.
Ok.
For MrMEEE version, is works like described just below :
We use VirtualGL server on a second X server, and use an associated VirutalGL client on the original X server.
The fisrt X server is running with the intel card, and we launch the second one on the nVidia card.
VirtualGL is a quite new software that is rendering graphics on a (in practice) powerfull server, and then send back the rendered frames to the client.
We're creating the second X server when launching optirun. The server is removed after end of execution.
In our current implementation, it is quite the same, but it uses a daemon.
So it requires VirtualGL (and its own dependencies), nVidia drivers if you want to use it (the work for nouveau isn't done for now), and then you need to configure all that.
See also https://github.com/Bumblebee-Project/Bumblebee/wiki/Bumblebee-Overview
Should you have any development questions, feel free to join #bumblebee-dev on Freenode
ArchangeGabriel, please confirm if I understood you right (or correct me).
We use nVidia GPU for rendering only. Then we get fresh rendered image and display using Intel GMA.
As I understand it works exactly as on Windows.
And what do you mean "For MrMEEE version"? There are some other versions?
And another question - can I configure my system to start whole X server on nVidia, not for single application?
I mean full time nVidia OpenGL from starting KDM to KDE + all applications started by it and till shutdown of X.
So the goal is to run X like if I had the only GPU - nVidia.
I think it's possible. We should start primary X server, then start secondary with VirtualGL on nVidia, and then start KDM/GDM/XDM on secondary X. Login in selected window manager and work with OpenGL and full-time HW acceleration.
Please confirm if my idea is correct.
Your view is correct. Note that some (OpenGL) features might not be fully supported by VirtualGL which is used by Bumblebee for transferring frames from the nvidia card to the intel one.
We've forked this project and continued working on https://github.com/Bumblebee-Project/Bumblebee
MrMEEE has decided not to join us for now and continues on https://github.com/MrMEEE/ironhide.
I've never tested it myself, but you should be able to start the whole DE with optirun by creating a new session type in /usr/share/xsessions/ and adjust Exec accordingly.
Closed here, opening an issue on the new tracker: Bumblebee-Project/Bumblebee#45