If this is your first visit, be sure to
check out the FAQ by clicking the
link above. You may have to register
before you can post: click the register link above to proceed. To start viewing messages,
select the forum that you want to visit from the selection below.
- From market perspective, there are no market forces which would be forcing Nvidia to open-source their OpenGL drivers.
You are mixing up the driver, and OpenGL implementation. Driver should be responsible for hardware support, OpenGL implementation should be a library for supporting that API. I'm talking about the driver (lowest level hardware interaction layer). OpenGL API implementation for Nvidia is already pretty good in Mesa, so I don't care much about Nvidia's closed version. If they'll provide open driver that works well with hardware and fully exposes its features, Mesa can be combined with it, I'm sure.
Originally posted by atomsymbol
- Including parts of Nvidia drivers into the Linux kernel would introduce a significant lag: Support for new Nvidia hardware would arrive to Linux users months after the hardware is released (and the same amount of months after it is available in Windows). Closed-source drivers do not have such a lag problem. ... AMD's open-source efforts are a proof that the lag exists: it is very uncertain which Linux kernel release will gain support for Polaris GPUs (because of the centralized nature of Linux driver development (centralized = almost all drivers are in the Linux kernel git repository; out-of-kernel drivers are an exception rather than the rule)).
What prevents them from providing flexible kernel modules that ship their driver at their own schedule (for those who want to be ahead of what standard kernel packs)? It still can be perfectly open. Linux supports modularity for years already, so it shouldn't be an issue, unless I'm missing something.
Now let's stop polluting all threads touching Nvidia with this bullshit. There are some nice mid-range GPUs coming tomorrow, and one of them is going to be the next big seller, even among Linux gamers.
Surely there will be some masochists. But most will probably just buy 3 ultra high-end AMD GPUs, a house and a new car for the same price.
I have some expectation of this. With more stuff being pushed into firmware blob, making an open driver is probably something Nvidia will eventually do.
Sure... The again pigs will also fly.
Seriously thou, Nvidia obviously doesn't really have any interest in open sourcing any more things than they already are. If their recent behavior with things like moving to signed firmware images for 900-series cards is anything to go by it's only going to get worse.
As for the actual topic at hand, Nvidia is obviously trying to get the drop on AMD and officially unveiling the first Pascal cards before AMD can unveil the first Polaris cards at Computex, which is at the end of the month.
Seriously thou, Nvidia obviously doesn't really have any interest in open sourcing any more things than they already are. If their recent behavior with things like moving to signed firmware images for 900-series cards is anything to go by it's only going to get worse.
But that's the catch. It gets worse in firmware sense, since you can't use those cards without firmware blobs, and on the other hand it can make it much more straightforward for Nvidia to open the driver. Kind of a weird situation.
Also why would nvidia make their driver more modular? now theres nvidia_drm nvidia_modeset and nvidia_uvm. Perhaps they plan to release some of it to be open source?
I think that's supposed to help with Display offloading support so that Bumblebee notebooks are no longer totally unsupported and you can drive additional displays from onboard Intel graphics when using a discrete Nvidia GPU at the same time.
Comment