nV News Forums


nV News Forums (http://www.nvnews.net/vbulletin/index.php)
-   NVIDIA Linux (http://www.nvnews.net/vbulletin/forumdisplay.php?f=14)
-   -   Satisfaction / dissatisfaction with nvidia (http://www.nvnews.net/vbulletin/showthread.php?t=92386)

kriko 05-30-07 10:34 AM

Satisfaction / dissatisfaction with nvidia
netllama gave me this idea, so I'm opening this thread to discuss opinions on nvidia.

What is bothering me is the almost "fully" closed development of new drivers. You don't know which bugs are going to be fixed in future release, when are they going to be fixed or if they are assigned to be fixed at all.
So I'm proposing an open bug-tracking system, where customers would report bugs, see their priority and progress.

Next thing is moderation on this forums. Why netllama closed this popular thread:
of course it was a bit offtopic, and it bring a bit of shame on nvidia, but why not instead clean it a bit (delete or move to a new thread offtopic posts)?
As you can see, we were trying to find a way around a problem, so you can't just close it.

Thats it for now. Please leave your comments, or ideas on improvements. Maybe someday the'll become reality.

EDIT: added poll

evilghost 05-30-07 10:56 AM

Re: Satisfaction / dissatisfaction with nvidia
Compared with the alternatives, such as ATI, yes I'm quite satisfied. It's also nice to be able to directly interface with NVidia employees though this forum.

Ulayo 05-30-07 11:19 AM

Re: Satisfaction / dissatisfaction with nvidia
Closed source doesn't bother me much. I just wish they'd put down more time to fix issues and performance.

kriko 05-30-07 12:50 PM

Re: Satisfaction / dissatisfaction with nvidia
evilghost: yes, compared to alternatives is much better, but it could be even better than it is. I pointed out some ideas, if someone else got them, bring them. Maybe nvidia will listen.

@mods: Why was this thread closed for some moments?

Sylvain59 05-30-07 12:50 PM

Re: Satisfaction / dissatisfaction with nvidia
I say Yes, but there are still some things to improve.

I can understand that open sourcing the drivers is not possible due to patents. But perhaps it is possible to give no-patented informations which are sufficient to develop open source drivers from scratch.

People could have the choice between open source drivers which are perhaps less buggy due to less complexity, and proprietary drivers giving better performance.

jeisom 05-30-07 12:56 PM

Re: Satisfaction / dissatisfaction with nvidia
Not really one way or the other. I'd like to see improvement to video accel though. but compared to the alternative currently available i lean to satisfied

gfxdrone 05-30-07 12:58 PM

Re: Satisfaction / dissatisfaction with nvidia
satisfied/dissatisfied is far too black and white. Is nVidia driver support better than something that doesn't even exist, ATI support? Of course, but that's not saying anything meaningful.

Do nVidia drivers work better than the current crop of ATI drivers? Since ATI currently doesn't even have a single driver that works with xorg 1.3 type identifying xorg, again, no question, nvidia is doing the job very well, and continues to do a very good job.

But after that it gets far less black and white. Personally, after reading these forums for a while, I see the entire project of trying to provide working non-free, binary blob drivers for the linux kernel as doomed to fail long term.

And the issues like nvidia drivers requiring reinstalls after every xorg update also point to problem areas, in this case conflicts with the packaging systems of the distro, debian and all its derivatives, for instance.

Then the inability to function with things like paravirt_ops, without doing gpl violating hacks, again, not a recipe for long term success. In this case, for example, we have to release new kernels with less than full features enabled to allow them to work with nvidia drivers.

So I really don't envy nVidia employees at all, they are faced with what is really not a realistic long term task, but they can't do anything about it, since the decision to open the drivers is not their's to make, so they have to deal with what they can handle on their end. And in that limited field, they do a very good job.

But the moment that AMD/ATI comes out with an open driver, if they actually do follow through on their hint that they might do that, nvidia will fast lose all relevance in the Linux market place. Already, for non-gamers, choosing a Intel mobo based video chipset, for it's free/open driver xorg support is becoming a no brainer if you can find an adequate system re price/performance/features. The binary blob driver is not going to survive as a Linux kernel business model long term, there's almost no question about that in my opinion.

But for now, nobody comes close to nVidia in the current Linux kernel non-free binary blob driver marketplace, their drivers continue to 'just work', release after release, they are easy to handle, they do not fail for our users, and they cause me almost zero headaches, month after month, and for that I really have to say a big thanks, with the above qualifications.

(I didn't vote, because I can't say I either am or am not satisfied, some things are good, some aren't, so to vote satisfied would ignore the problems, and to vote dissatisfied would fail to give credit for the good work nvidia does do on an ongoing basis.)

TPG 05-30-07 03:48 PM

Re: Satisfaction / dissatisfaction with nvidia
Honestly i don't care whether drivers are opensourced or not, till they work perfectly.

I'm fully satisfied with _current_ nvidia support for *NIX :D

mehukatti 05-30-07 03:50 PM

Re: Satisfaction / dissatisfaction with nvidia
Despite of my recent problems with integrated GeForce 6150, I have been satisfied with NVidia's drivers. Yes some people have problems, but I think one has to remember that most people only post onto this board when they have some kind of a problem. Atleast this is why I registered here in the first place. Of course a more open development model would be nicer, but atleast they are fixing bugs even now.

m-singh 05-30-07 05:20 PM

Re: Satisfaction / dissatisfaction with nvidia
Nvidia deserves the linux community's thanks for being the first major graphics chip maker to provide linux drivers. That has won nvidia an immense amount of goodwill.

However, you cannot ride on goodwill alone forever. There has to be performance to match the growing numbers and needs of the linux community.

The persistent bug with nvidia's drivers on certain chipsets in laptops which borks hibernate (or for some people, suspend to RAM) is a case in the point. Many nvidia bug reports have been filed, every scrap of information asked for has been provided, but stable/beta versions come and go and this bug remains unfixed (upwards of 6 months at this point in time).

Nvidia owns the IP on its drivers and has a perfectly legitimate right to do whatever it wishes with them. However, if it wants linux users to invest in nvidia hardware on new machines, it has to demonstrate either of the following three :

1. Fix the bugs in time (yes - it takes a week or two to track it down, but 6 months ?? - this might be passe in the windows world, but it is extremely extremely rare for bugs to survive this long in well-supported Linux OS/applications) and keep the drivers closed source.
2. Open the specs of the chipsets so that people (such as the authors of nv) can add 3D support etc. I would like to mention that the particular bug I mentioned does not occur with nv drivers.
3. Open source the nvidia drivers themselves, so that bugs are hunted down and fixed ASAP.

Its not our place to dictate what nvidia should do. We neither have the right nor perhaps the legal knowhow to do that. However, the picture of closed source drivers that go months with unfixed bugs is not a pretty one. Earlier today, I had to make a purchase decision for two laptops for some of the new people in our group. I specifically chose intel 9xx cards for them (and this is when one of them is a windows user, where the windows version of the driver probably does not have the bug - just what happens when you piss off your fans) because I cannot in good conscience inflict the kind of experience I have had over the past few months on some other people.

I would like to propose that a clean set of drivers (open source or closed or even competing (option 2 above)) with a process that quickly fixes those bugs would do a lot to attract and retain linux users. How Nvidia arrives at that sound business decision is Nvidia's call. My job is only to provide customer feedback as I see it.

lloeki 05-30-07 06:20 PM

Re: Satisfaction / dissatisfaction with nvidia
yes, satisfied, as critical issues were fixed for me. but there's room for improvement. out of the blue, maybe there's:
- a lack of manpower
- an inefficient/inappropriate workflow
- an improvable involvement of the community

closed source doesn't bother me, as long as I have long term, reactive, efficient support, but as of now, to me we're only half-way there.

zbiggy 05-30-07 06:23 PM

Re: Satisfaction / dissatisfaction with nvidia
Nvidia has the best GPU driver for Linux. The most feature rich and fast developed. The support for new geforces is added really faster than before.
The tech support here is great. In comparison ATI support is none and new radeons add to driver can take months to years.

Nvidia should have public bugzilla. It will save developers time: users will add comments to existing bugs instead of making new threads again and again.

Another bugzilla should be made for website related things - the place to report wrong files, old links, outdated information etc. The nvidia ftp server needs some cleanup and "nvidia driver download beta" application needs some fine tuning for Linux users. There is no webmaster e-mail address on nvidia.com so there is no possibility to leave Nvidia a hint on these issues.

All times are GMT -5. The time now is 04:09 PM.

Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.