Submitted by Open Graphics Project founder Timothy Miller recently noted on the project’s mailing list that they are set to announce that their first hardware, the OGD1, is ready for pre-order. “The OGD1 design has actually been finished for a couple of months now,” he began, explaining that they’ve been setting up a way to process pre-orders for the first 100 boards. The board will retail at USD 1500, with a USD 100 discount offered for the first 100 pre-orders. “These are pre-orders, not orders,” Timothy continued, “that means the lead time is unpredictable. We don’t have a stock. We will purchase a stock based on the number of pre-orders we get. Also, this means that if we never get a large enough number of pre-orders, we will be unable to fulfil them; all pre-orders would be cancelled, and no one would be charged anything.”
$1500? Ouch!
Just looking at their goal of running Quake III at a frame rate of between 20 and 30 FPS at 1280×1024.. not worth it.
I can see they’re trying to position this card as more than just a graphics card, but at that price I wont be bothering.
From the link:
This is NOT a graphics card. This is a hardware development prototype card. This may be used to develop a graphics card, and even used as one (it can also be used as a paper weight and a hammer), but it isn’t a graphics card. If you don’t want to develop hardware then don’t get this card.
This would have been useful for a friend of mine who was doing research on building hardware to speed-up bio-informatics calculations. Or if you wanted to build an open-source wireless card. This is not that expensive for what is meant to be, however, taken as a graphics card -or a paper weight or a hammer- it is outrageously expensive.
Edited 2008-04-16 21:56 UTC
Yeah, me neither. I can buy an Optimus Maximus for that price
…that IS definitely too much. I certainly don’t have that kind of cash lying around, and I’m sure most users won’t be able to afford it either.
I seriously hope this won’t be the end of the project as I admire what they’ve accomplished so far. But something definitely needs to be done to make it more affordable. đ
Edited 2008-04-16 21:33 UTC
This is a developers boeard, not a graphic card.
It contains a FPGA (freely programmable gate array) which is some sort of rewireable chip.
It is needed to design the chip layout (which transistors have to sit where, and how should they be connected).
When the chip layout is finished, it will be given to a chip manufacturer who will then make the actual special purpose fixedly wired graphics chip. This step will cost close to 2 Million US$ up front.
The actual graphics chip will cost much less than 1500 $. It will be close to 150$, something I at least would gladly shell out to get a graphics card which will be the best supported one in every Linux and BSD distro.
Please check the link:
We often get inquiries about the use of OGD1 as a graphics card. It can easily-enough function as a graphics card, but for most such uses, it is badly over-priced. On the other hand, OGD1 is very competitively priced as an FPGA development kit.
With ATi kind of trying to make their documentation open and iNTEL pretty much opening theirs, will there be a place for an “open source” graphics card?
Ok, I know this is both cheeky and a bad comparison:
With Microsoft trying to make their protocols open, is there really a place for open source operating systems?
How exactly do you learn how to build a graphics card and improve on existing designs without having reference implementations to learn from and play with?
Open source is more than having information on how something is built… Drivers and specifications are just the software that allows the OS to communicate with the hardware – it doesn’t define how the hardware works internally.
This is not just an opensource graphics card, its a kind of roadmap on how to build your own graphics hardware, something you are not going to get from Intel or ATi. This is more for those who want to build their own graphics card and want to see the ins and outs of real on with full documentation, specs, possibly howtos, and source code showing you how and why it works. At least that is what i get form this project, otherwise the thign is rather useless as a video card. the only real benefit is educational.
Sorry but as far as I know ATi and Intel are only open-sourcing the interface to their cards not the actual gate layout itself. Actually, I don’t believing they are revealing any of the actual implementation of the chips. If you want to build your own graphics card (for a cell phone or for research) there is nothing you could reuse except for maybe the interface… but that’s probably patented.
In summary: Intel and ATi may give us OSS drivers, but still no OSS graphic cards.
But do we know that the machines that will be manufacturing these cards will be running OSS? Because if they’re not, it’s not truly an OSS card.
Sarcasm duly noted. I’m just saying that their goal is a pure OSS card, and that is why they are doing what they are doing. After all, there is the potential that the OSS hardware market will be just as important as the OSS software market. With the declining prices of FPGAs there are those possibilities.
I’m no purist myself (written from XP), but I see there point and where they are going with this. That said, I think it would be nice if we had a complete open-source patent-free stack with lithography and everything!
Edited 2008-04-17 00:55 UTC
You see, the new “openness” of ATI is not really reflected by it’s driver development efforts.
You still see ATI not pouring lots of resources into the open source drivers. Its more a sort of “we finally no longer block you open sourcers” attitude.
OGA will be completely different, ALL of the specs will be open, down to the hardware level, not just some ATIBIOS ABI thrown out to the public.
In essence there will be open drivers AND open firmware, something no other manufacturer offers. It’s the ultimate open source approach.
This is not the “retail” graphics card they intend to eventually produce.
This board is primarily for hardware hackers who want to play with FPGAs or for those who are actually participating in the development of the “retail” or user-oriented ASIC version of the OpenGraphics card.
1500 is a lot for a graphics card, but its very competitive for what it is. In fact, its probably the only “video-oriented” FPGA board of its kind.
I recall there being an FPGA ray-tracer implimentation implimented in a 50Mhz FPGA beating out a 2.8Ghz P4 — imagine what someone could do with the speed and additional gates this card provides.
Yeah, think about some special-purpose number crunching.
Want to have a huge matrix-inverter? Get 5 of these cards, stick them intou your 5 PCI slots and get hacking.
I am quite convinced, that you can blow away most $100.000 Number crunchers.
I did some looking on their website. Sadly, If the card only can do 20-30FPS in Quake III, then it’s less than half the speed of an obsolete Nvidia Geforce 6600GT. I think it would compare to an Nvidia 5000 series, or perhaps a 3TI.
It’s too way too slow for modern gaming, and it might also be too slow for some CAD programs.
At $1500, it would compare to a $30-$50 commercial card.
I believe they set their sites way too low. It might have been better to use a regular CPU, and boot it into Linux to run the rendering code in software.
Edited 2008-04-16 23:10 UTC
I did some looking on their website. Sadly, If the card only can do 20-30FPS in Quake III, then it’s less than half the speed of an obsolete Nvidia Geforce 6600GT. I think it would compare to an Nvidia 5000 series, or perhaps a 3TI.
It’s too way too slow for modern gaming, and it might also be too slow for some CAD programs.
You have completely misunderstood the whole point of OGD1, it is NOT a graphics card. If they had just planned to create a 3D graphics card with open specs then it would have been a whole lot cheaper than this. But this is a board meant for hardware developers! You should read their site a little bit more carefully
Then why set themselves a goal of 20-30fps for Quake 3? It’s positioning it as a graphics card with lines such as that.
Well you are not going to put out something that doesn’t work just because its open. You at least want to give people who are going to buy it something that works. If they were really shooting to make a graphics card 20-30fps in Q3 is a joke, they could do a WHOLE lot better than that without even trying.
God almighty – doesn’t anyone read the damn website; its a friggin development kit; of course it is going to be expensive. Once the ground work for 1.0 is released, the cost of pumping out GPU’s will be next to nothing – the development kit right now costs a fair bit.
For me, I’d love to see this succeed; I’m not a big gaming fan, and for most people, that is the case. If it means that they can purchase a machine with decent 2D performance and ok 3D, and fully supported by opensource operating systems, it would give a great push forward.
With that being said, I don’t think it should stop at video cards; lord knows the PC is filled with proprietary junk that should be replaced with open components using open standards.
It is an important milestone, I think we should buy these cards in a donation manner
1500 * 100 = 150.000
so we need 3.000 donations of 50$. If large companies step in the donation amount can be reduced.
I am interested to donate to allow developers to have one handy. And for the rest who compare it to a graphics card, indeed it is competitively priced as a development board but please do not troll. If you are not interested in it or you wait for a final product there is nothing for you here.
$1500 is the retail price, but discounts are available. Also, the board itself cost a lot to develop and costs a lot in parts. As such, the amount of “profit” that would be earned from sales of those cards isn’t nearly what you’d expect. If we were to make (OPTIMISTICALLY) $500 per card, that’s $50’000 in profit, but that’s money that would be immediately reinvested into more open hardware development, not put into anyone’s pocket. And $50’000 may sound like a lot, but it’s not enough to pay anyone’s salary. We’re paying up-front costs out of our own pockets and working for free.
At this point in time, the most valuable thing that can be done by anyone who is enthusiastic about this is to publicize this and get people to place pre-orders. The web site order form still needs some work, but we can take pre-orders NOW. If you believe in Free Software, get the word out.
Yes, this product is not for playing Quake..
– like it clearly says in the story, if you read it.
Now that the Open Graphics Project has already got this far, I cannot see why their products couldn’t find their own niche market later too? Of course, most computer users won’t probably be interested in OGP, but not everyone is interested in, say, professional CAD/3D cards, and not everyone needs support for three or more monitors from their video card either. These projects have their own niche market too.
I could expect that users and developers of many alternative operating systems from Syllable to OpenBSD could be interested in these kind of open graphic cards. The users of those operating systems could have working 3D support then too. Oh, and by the way, I don’t think that many of them would be too interested in wasting their time playing the latest, most fashionable video games anyway…
How well a video card may do Quake is actually a very minor detail for most computer users… Believe me but most people on the planet don’t know nor care to know what the heck “quake” might be. Many couldn’t care less for video games in general either. For most (smart) people there are plenty of other uses for computers and video cards than just playing some stupid games…
Edited 2008-04-17 16:06 UTC
I think the point of this card isn’t to rival nVidia or ATI cards in performance, just yet. It is, however, a step in the right direction to do that. The open source software community is full of at least thousands of software developers, all working to improve the software. Open source hardware may get the same treatment. If a dozen electrical engineers donate their time to the project, that’s a dozen annual salaries that won’t be included in the cost of a finished product. In the end, we’ll only be paying for the silicon and gold, basically. The cost to mass produce a GPU is probably a lot less than the wholesale price, one would think. Right now though, at US$1500 each, they won’t be on anyone’s hardware shortlist for building a gaming machine. They will be cheaper in years to come, I think.