Kickstarter-Based Open-Source GPU Launches

[scionicspectre]

So they need a million dollars before we can even get to something comparable in performance and features to an integrated GPU, let alone a discrete card? Oy vey.

I mean, what's better- an open GPU, or an open GPU people want to use in daily life? Not to harsh on them- I'm sure it's absurdly difficult to come up with even a simple GPU- but this seems a little pointless for anything aside from research. It could lead to a better open design from someone else building on it, I suppose. Am I missing something obvious?

[tga.d]

I think what some people fail to realize is that this is an open *hardware* solution. Intel's FOSS drivers are very nice (I use them myself), but that is distinct from the hardware itself being open. Things like this are actually extremely useful for security purposes, as it allows hardware-level DDC (Diverse Double Compiling, used to counter the Trusting Trust attack most famously explained by Ken Tompson), and in general allows for much more thorough security auditing at a hardware level. The other useful aspect is that it would be fully free and therefore could be customized for different purposes. It's not going to be replacing gaming hardware, but it's at least a baseline for security and research purposes.

[tarceri]

So they need a million dollars before we can even get to something comparable in performance and features to an integrated GPU, let alone a discrete card? Oy vey.

I mean, what's better- an open GPU, or an open GPU people want to use in daily life? Not to harsh on them- I'm sure it's absurdly difficult to come up with even a simple GPU- but this seems a little pointless for anything aside from research. It could lead to a better open design from someone else building on it, I suppose. Am I missing something obvious?

Well the obvious point that most people seem to overlook is that you need to start somewhere. This was never going to compete with a modern GPU from the begin, but it could be a start towards a future with open hardware, no longer the need to wait for companies to release technical documents in order to write drivers, and worldwide collaboration to create a superior product. Linux itself started off as a small project well behind the Unix's of the day, now its taking their market share a a phenomenal rate.

The funny thing is most people will spend more money on a coffee than they would put towards a project like this even thought they find it interesting (not that I'm saying people don't have a right to decide what they do with their money but I find it interesting). In my opinion its a real shame, projects like this have a chance to really stir things up but its projects like Ouya the worst games console ever that seem to get the big $. I guess I'm just a dreamer.

[maiden]

PCI etc.. is complex. maybe I'd want to implement my own simple bus like wishbone.

Dunno about you but I have quite a bit of fun and rather often. Sure some aspects of life suck but we can't focus on those and become depressed.

I like nature myself... perhaps I can use this GPU to render some nice nature scenes to distract me from the facist utopia that surrounds us ;-)

You mean like we did? orgfx is based on the wishbone bus.

http://opencores.org/project,orsoc_graphics_accelerator

http://www.youtube.com/watch?v=qiielWxDH24

it got vector, 2d and 3d graphics and is available for free now. I might need some more work

[scionicspectre]

Well the obvious point that most people seem to overlook is that you need to start somewhere. This was never going to compete with a modern GPU from the begin, but it could be a start towards a future with open hardware, no longer the need to wait for companies to release technical documents in order to write drivers, and worldwide collaboration to create a superior product.

True enough. Of course, by today I would think building a bulky, less elegant GPU with similar capabilities of the tiny chip in a four year old phone wouldn't be absurdly laborious, unless the core concepts of building a 3d shader model GPU are downright fringe science, regardless of the size.

Of course, not everyone has the combined talent of a chip manufacturer and 3d software engineer, so I definitely see how that's not a good match for volunteer work. I'd say I agree with you that having a crappy 2D GPU is better than nothing, and gives at least hobbyists an alternative option in creating devices.

Not to mention, with a sufficient CPU, software-rendered GL could still run on a 2D chipset, I take it. Heck, maybe I should contribute a little.

[cb88]

True enough. Of course, by today I would think building a bulky, less elegant GPU with similar capabilities of the tiny chip in a four year old phone wouldn't be absurdly laborious, unless the core concepts of building a 3d shader model GPU are downright fringe science, regardless of the size.

Of course, not everyone has the combined talent of a chip manufacturer and 3d software engineer, so I definitely see how that's not a good match for volunteer work. I'd say I agree with you that having a crappy 2D GPU is better than nothing, and gives at least hobbyists an alternative option in creating devices.

Not to mention, with a sufficient CPU, software-rendered GL could still run on a 2D chipset, I take it. Heck, maybe I should contribute a little.

Yeah... I've acutally ran Minecraft using Mesa LLVMpipe.. it runs decently on an i7 laptop better than a Geforce 2MX actually you can even adjust how many cores to give it with environment variables.

[uid313]

They're selling not funding

Kickstarter is for funding someone to do something.

What they're doing is not collecting funds to do something, what they're doing is selling something.
They already have the 2D accelerator done, it's already made.
So there is nothing to fund since it's already done.
Basically they're selling it for $200,000.

They're not getting funding to make something they don't have.
They're selling something they already have.

[intellivision]

Using the LGPL for hardware seems to be outside the scope of that license.
Has the FSF approved this sort of use or given any legal advice to that effect?
How do you define static and dynamic linking with hardware? How do you make sure that a library can be swapped out with one of another version?

[Kivada]

Waaaaaa

Their best case scenario is 2001 era OpenGL1.4 class hardware, you can get that on eBay for $5 TODAY with OSS drivers in PCI, AGP and probably even in ISA if you dig.

Unless theres some HUGE untapped market for ancient GPUs and i586 or UltraSPARC CPUs this project is a waste of time.

The million would be far better spent on hiring more debs for the GPUs people actually want to buy.

[Kivada]

GPU shaders are programmable in GLSL?
OpenCL is supported ?

From the $400,000 goal

3D Option

This Stretch Goal adds optional* full OpenGL and Direct 3D (7.0/8.0) 3D graphics.

Thats OpenGL1.4, from 2001, so "shader based" apparently means vertex and pixel shaders.