“One project that I’ve been following quite closely lately is a project started by chip-designer Timothy Miller, called the Open Graphics Project. His goal, along with the rest of the project, known as the Open Graphics Foundation, is to make a 3D accelerated video card which is fully documented, free-licensed, and open source.” We have already covered the OGP a few times, but this article gives a nice overview of the project.
Perhaps a title other than “introducing …” would be appropriate for something “already covered … a few times”.
It appears the site is down.
Same for me. I used google’s cache to read it.
http://72.14.221.104/search?q=cache:y4G4wfIjXxsJ:www.freesoftwarema…
I think this deserves more attention from free/OSS advocates.
It really is a great idea.
If NVIDIA and ATI won’t make it easy for you to make drivers for video cards, just make your own video card.
Supposedly, the reason for them not making it so easy is because they each violate the others patents.
Supposedly, the reason for them not making it so easy is because they each violate the others patents.
‘Know what? Here’s my take: it has nothing to do with patents and everything to do with class action suits. People “in the know” (driver modders for example) have been patching drivers to re-enable functional units on crippled cards which are sold for a lower price than their non-crippled counterparts. Sometimes these cards are crippled because of yield issues (some functional units are actually defective); sometimes, just to broaden up the market and sell these crippled cards as budget models in order to cover for the ASIC manufacturing costs. Some Radeon 9500 models could be soft-modded to a Radeon 9700, for example. Same with the GeForce 6200, to a 6600. There are plenty of other examples, these two I know by heart because of the many times I installed the Omega drivers on my computers, and because I own a 6200TC, which to my dismay can’t be soft-modded. =P
If they opened their source code either they’d have to disable the parts that enable the extra juice on selected “premium” models and people will complain that the open source drivers (mind you!) are crippled/a bad joke, or “inferior” models are suddenly going to perform much quicker than before, and people who spent more cash on more expensive models will get so pissed that they’ll definitely sue the cardmakers. The thing they definitely won’t do is let this differentiation explicit on the code.
Of course, there’s always the possibility of patent wars, but IMHO that’s a lesser evil than the bad PR and probable class action suits that would follow any of the alternatives just described.
Note that such lawsuit would be laughed out of court quite quickly: ATI or NVidia promise you a set a features for a certain price.
It doesn’t matter whether the board has exactly the hardware for those feature or whether they have more but some of the additionnal features are disabled.
Yep, just make your own card… that costs more and does less than Intel integrated graphics which already has open drivers. Can someone remind me again what is being accomplished here?
The open drivers for the Intel chips still lack some performance features that are covered by certain Intel IP.
ATI radeon boards have open sourced drivers, if you have a r100 or 200 class chip. They might not be the speediest performers by today’s standards, but surely one of the top-end r200’s can rock the heck out of any intel integrated chip.
My firegl 8800 128MB was only 60 bucks off of Ebay.
Here, take a look:
http://dri.freedesktop.org/wiki/ATIRadeon
http://www.phoronix.com/scan.php?page=article&item=463&num=1
The OSS/DRI drivers are not as fast as ATI’s closed drivers, but there’s been alot of talk about the FGLRX drivers not working properly with r200s anyways.
Even the specs on the R100 and R200 are not completely open. Important features like HyperZ are not documented, and had to be reversed engineered.
In any case, with the HyperZ patches, the DRI drivers are actually faster than the FGLRX drivers, at least on Quake III.
Yep, just make your own card… that costs more and does less than Intel integrated graphics which already has open drivers. Can someone remind me again what is being accomplished here?
Intel’s graphics chipsets can not be used with non-Intel CPUs or with arbitrary motherboards. To my best knowlegde, they provide neither TV out nor dual head capabilities.
fs
If NVIDIA and ATI won’t make it easy for you to make drivers for video cards, just make your own video card.
Oh yeah, should be an afternoon project.
The real question is, when this thing is done who is going to manufacture it?
Obviously, like any feature based open source project, “when it will be ready”.
But here, they will need heavy test against FCPGA based boards, since ASIC production is expensive and cannot be fixed afterwards. I guess, depending on success of FPGA cards – read : number of testers – that we will have to wait 1 year, at least 6 months : Our determination is that the 3S4000 is the only chip available within the next 6 months that will meet our needs ( http://wiki.duskglow.com/tiki-index.php?page=OGPN10 )
Edited 2006-08-02 22:10
You didn’t read my question correctly.
Who is going to manufacture these cards?
Who is going to manufacture these cards?
Timothy Miller founded Traversal Tech with two other Tech Source employees. They hope to raise enough money (from venture capitalists and other sources) to produce and sell these cards although this is one of the biggest obstacles on the way towards a completely free graphics card.
fs
Open hardware technologies such as this one may be able to find synergies in NPO-driven initiatives such as the line the One Laptop Per Child program.
Creating and more even distributing a piece of hardware is a very different beast then creating and distributing a piece of software.
On my computer I have a Via Unichrome graphics card, built into the motherboard. I sat for about a year trying to get it to work with various combinations of closed source drivers, then different versions of the open source drivers, DRI modules, X server versions, Linux distributions, and even Windows 98 (2 closed source driver versions). No matter what I did, something – logging out of X, playing a movie with mplayer, certain scenes in a game – would just lock up my PC (no magic SysRequest keys worked, no watchdog timer helped, nothing – totally dead – had to hold power button for 4 seconds).
This is with an open source driver – the maintainers of which, by the way, refused to help me.
After hearing NVidia makes good Linux drivers, and finding a cheap FX5200, I went and bought it, and haven’t looked back. My PC hasn’t crashed once – Linux or Windows.
“Open source” does not imply “good” in software, and I assume in hardware too.
“Open source” does not imply “good” in software, and I assume in hardware too.
This statement contains the assumption that closed source does imply good. Which isn’t true either.
It’s not the label open source or closed source that removes or gives quality to a product or project. It is the dedication and skill level of the crew making the goods.
Some groups will produce high quality stuff, some groups are stuck producing stuff ready for the landfill.
Because you still did not have open source! You had open software but closed hardware. There was no way to confirm the software was writting/reading the correct values from the correct hardware.
Infact, if the hardware was open you could have written you own driver from scratch – not easy, but at-least you probably could have gotten the card to act like a simple frame buffer without locking up your system, and then improve it over time from there.