Linked by Thom Holwerda on Sat 15th Sep 2007 20:16 UTC, submitted by Michael
AMD AMD started delivering on their word of providing GPU specifications to the open-source community without a Non-Disclosure Agreement, and now with the 2007 X Developer Summit having come to a close, we asked several key members of the X.Org community on how they judge AMD's recent move. They were also asked if they believe NVIDIA will follow suit in helping the open-source community. Those that responded were David Airlie, Daniel Stone, Jerome Glisse, Stephane Marchesin, and Oliver McFadden. Mark Shuttleworth had also previously commented on AMD's efforts.
Thread beginning with comment 271704
To read all comments associated with this story, please click here.
SIS?
by klimg on Sun 16th Sep 2007 11:33 UTC
klimg
Member since:
2007-08-03

The sis driver is a one man project w/o any support from the manufacturer and pretty much a hit or miss thing depending on the chipset.
Not a good idea to bring that mess up as an example for a great driver (don't mean to put down Winischhofer - he does an excellent job with the resources he got).

ATI is/was a mess for years and had so many shortcomings I don't even care to mention it.

As they get ready to integrate the CPU/GPU for low cost systems and Linux gets some traction in that market they decide to release specs.
Somebody could get the idea that they couldn't get it right for years and now decided to let the xorg folks do their job for free to be ready in time.

Nvidia supported Linux and other OS's for a long time and was out there with an 64 bit driver when Ati was out there with a bunch of promises.
Not being a close relative of RMS I don't see myself switching to ATI anytime soon.

Reply Score: 2

RE: SIS?
by cyclops on Sun 16th Sep 2007 13:25 in reply to "SIS?"
cyclops Member since:
2006-03-12

"Not being a close relative of RMS I don't see myself switching to ATI anytime soon."

Be very careful trying to label RMS.

I can say why I chose my X800 card, because it gave me a secure card, with reasonable frame rates in those games available on the GNU platform, while allowing me to take advantage of cutting edge versions of the kernel; X; and the compositing desktop.

No other group of cards could do that. I would be happy for you to try and dismiss that comment.

So why would you personally have a hard time to dismiss Richard, apart from his personal contributions to free-software and the open-source landscape, and his tireless work in the free-software movement...

No, Because Nvidia's binary drivers are garbage.
http://www.news.com/2100-1002_3-6126846.html

"It is our opinion that Nvidia's binary driver remains an unacceptable security risk based on the larger numbers of reproducible, unfixed crashes that have been reported in public forums and bug databases," Rapid7's advisory said.


You have to remember when talking about binaries and open-source that you have to be very careful. I expect a higher standard. Personally I thank my lucky stars for the FSF that have worked tirelessly to provide alternatives to this software. If I was to pick the top complaint in the forums it would be "Flash on 64bit", and I see that a credible alternative is coming to my desktop. In only Linus was as fastidious. Maybe this 3D problem that has existed for so long would not have been as significant factor working against GNU mainsteam acceptance.

Now please don't waste my time with off the cuff remarks.

The sis driver which I have fond memories off for many reasons. The top one being developer accessibility, regular easy to understand updates, the choice of a drop in binary or source code..and a forum. In contrast I trail 3 message boards 3 development blogs a combined blog and several git commit branches and two bugzilla boards and an irc log, and have a hard-lock problem on my ati for the last four revisions with no easy access to anyone who is able to address the problem. It was/is a disgrace that sis did not support him. As a side note I find power-user support in Linux excellent, Its a shame its developers are not as approachable.

With Ati you have to remember there is a difference between the r100/r200 and the r300/r400 drivers. I have given an accurate description of the r300 driver as *I* see it, and I think few would disagree with that description. I have little to say on the r100/r200 drivers apart from they are considered more feature complete than that of the r300/r400 ones and have support officially dropped by ati, it seems that RMS was at least right about those drivers.

If you are trying to pick a *solid* reason for ati have releasing some of there specifications etc. Then you are better than me I think its a combination of factors from good publicity; the imminent release of the avivo driver; acquisition by a company selling a hardware platform not a graphics card; The development of a reverse engineering tool; following Intel's working example who are now have a successful real 3D alternative. Want to regain some control of the kernel fir-to-busting with those on Intel's payroll...just look at all the kernel develops in their Intel t-shirts. Have lost a major mind share in the technical community to Intel's core duo etc etc. I suspect you will have some well paid analysts try to figure out all the variables...but I wont even try.

Now you can support a *graphics card* company if you want, and be derogatory about a organization that puts *users* first, but I think your wrong. personally I wish companies put users first too.

Edited 2007-09-16 13:32

Reply Parent Score: 4