LinuxFund and the Open Graphics Project are teaming up to raise funds and supply 10 Open Graphics Development boards to open source developers. After several years in development the Open Graphics project is offering pre-orders of development boards. The Open Graphics Project aims to design an open source hardware/open architecture and standard for graphics cards, primarily targeting free software/open source operating systems. LinuxFund is accepting donations on their website to help fund the project. Additionally you can pre-order an OGD1 board for yourself through Traversal Technology.
if these can be made available as a boxed product that has a reasonable price point and OpenSolaris support I would definitely fork out some cash, even at a premium to support an open solution.
The final product is something I would seriously consider buying if it really does work out to be something that is not only easy to develop open source drivers for but also gets decent graphics performance. It doesn’t need to be top of the line, only usable for a 3D desktop and popular 3D applications (other than games). It’s not an easy market to get into though and I fear it’s going to be a very difficult prospect to sell enough of these to make them price/performance practical compared Nvidia, ATI, and even Intel. Now that laptops are more popular than desktops it’s going to be even more difficult.
While I’m excited to see this project progress, I wouldn’t get my hopes too high about performance… According to this page (http://wiki.opengraphics.org/tiki-index.php?page=OGPN17) they are expecting this card to perform a bit above a Radeon 7000 or a GeForce 2, which was state of the art nearly ten years ago. I’ll probably get one of these in the future if they can produce them cheap enough (<= $40), but I don’t expect something that will perform on the same level as even the cheapest video cards you can purchase… Right now from taking a quick look at Newegg, it looks like you can pick up a low end GeForce 8 series card for $30 with an additional $10 MIR, but I’m willing to pay a bit of a premium for something that is fully open and I _know_ will be supported for years to come, however I don’t think most consumers look at it quite like that.
Just to clarify, the FPGA is expected to perform at that speed. An ASIC could be clocked very fast.
Hi,
A “top of the line” Nvidia card with no device driver (e.g. using VBE and doing everything in software) would probably be a lot slower than one of these with full support for hardware acceleration. I think you’re under-estimating the value of open documentation, especially for every OS that isn’t Windows.
-Brendan
While this is true if you can get a system that comes with an onboard Intel video card that has better performance, for a lower price, and has open source drivers then what niche will this card fill? AMD/ATI has also been opening up documentation and providing code.
I applaud the effort but the graphics stack and documentation has come a long way since this project was conceived and its relevance is in question if it cannot offer something above and beyond what is available now. It has to beat existing manufacturers in price, performance, or openness. From the beginning they were gunning for openness but this has changed a great deal in the past couple of years and I doubt they have the resources to offer better performance or a better price.
I think it is something of a chicken/egg problem. the efforts around OGD probably made Nvidia and ATI aware that there indeed is a third dimension next to price and performance. And having one open hardware vendor around would ensure ATI and NVIDIA stay open enough.
Whoever is more open will win the Linux market, with performance being of minor importance and the price even less so.
Loads of people are perfectly willing to spend $150 to $200 for a graphics card with a quarter of the performance of a top notch NVIDIA or ATI graphics card which would just cost the same. They would do that because it would mean ultra-long support times in the kernel, no installation/upgrade hassles, just plug an play.
I know it would be the first thing to buy once it is available, and I guess millions on this planet would think similar. I am even thinking of buying the development board once it comes “pre-flashed” as a graphics card, a good documentation how to reprogram it, and some tools to load up the new hardware layout to it.
The problem with this line of thought is that Intel drivers are completely open and probably perform on par or better than this solution. ATI has been releasing docs and code for a little while now. Nvidia is the only one holding back but considering 2 out of the top 3 graphics providers are already offering open code and documentation I don’t see where a product based on OGD fits in.
Just another blurb.
<a href=”http://www.fsf.org/blogs/community/help-the-open-graphics-project“….
Also, we may have VGA working this week. I’m hoping to take a video of a PC booting up with an OGD1 board as the console.
Bah. This thing won’t let me edit. Here’s some correction:
Just another blurb.
http://www.fsf.org/blogs/community/help-the-open-graphics-project
Also, we may have VGA working this week. I’m hoping to take a video of a PC booting up with an OGD1 board as the console.
Note that OGD1 isn’t a graphics card, so don’t expect it to be cheap. It’s a platform for developing graphics cards (among other things).
So one thing I might suggest, regarding donations, is to think about how much you might want to spend on a graphics card, and maybe use that as a rule of thumb for how much you might want to donate. Enough small donations will add up. We’re tired of hardware vendors not giving us documentation, and we want to do something about it.
Just to clarify. As I stated in my initial post, I am not referring directly to the development board but what final product may come out it. I think it’s great that someone is providing a open graphics hardware to develop a completely open graphics solution. I just wonder how practical open development is on sophisticated hardware. The big boys have access to a lot of capital and expensive equipment that most people don’t have. It is only going to become increasingly difficult to provide decent performance for a decent price without significant funding. Perhaps some Linux or other open source OS business will find it particulary useful and help with development.
Assuming the developer’s version of the card is more expensivethan… does anyone know what regular users can explect to pay? Even an initial ball park figure would be interesting.
Edited 2009-04-15 05:27 UTC
So far, there is only a developer card, and an end-user version wouldn’t be any cheaper if it too was based on an FPGA. Flexibility is expensive. Some people looking to buy an OGD1 want to drive two Apple 30″ cinema displays. Others want to do things like cryptography and audio (i.e. not graphics).
On the other hand, with enough resources, they could develop an ASIC version of their GPU and sell it cheaply.
“On the other hand, with enough resources, they could develop an ASIC version of their GPU and sell it cheaply.”
That’s what I am talking about/hoping for.
The challenge is getting those resources. Millions of dollars just don’t drop into your lap. We’ve tried getting help from Google, but it doesn’t seem to be their thing.