Kickstarter For Open Source GPU 108
First time accepted submitter eekee writes "The targets are high, but so is the goal: releasing Verilog source code for a GPU implementation. The source will be open source, LGPL-licensed, and suitable for loading onto an FPGA. The first target is for a 2D GPU with PCI interface; perhaps not terribly interesting in itself, but the first stretch goal is much more exciting: full OpenGL and Direct3D graphics."
Unlike the Open Graphics Project, this is starting from a working 2D accelerator and mostly working 3D accelerator cloning the features of the Number Nine Ticket to Ride hardware. If they get a meelion bucks they'll overhaul the chip to support something other than PCI (although you can bridge between PCI and PCIe) and implement a modern programmable rather than fixed-function chip. Also unlike OGP, they do not appear interested in producing hardware, instead focusing entirely on the core itself for use in FPGAs (anyone want to dust off the OGD1 design?)
Dubious Market? (Score:3)
How is this not an Ouya in the making? Is there actually a market large enough for this thing to do anything but vanish without a trace?
Apologies for the negativity. Please do correct my world view. I'm amenable.
Re: (Score:3)
The Kickstarter is for a code drop, no hardware.
According to Francis Bruno [phoronix.com], the 2D version fits on a Cyclone II 25 or a Spartan 3.
Re: (Score:3)
Why would embedded manufacturers prefer buying a GPU from someone who isn't established over the usual players? I don't think they care about having a 100% open source GPU driver.
I applaud this in principle, I just don't think there is a massive audience for this project, especially with Nvidia's recent mea culpa to the Linux community and their new promise to help deliver solid open source drivers for their cards. Given that Nvidia wants to be the heart of every Steam box (as well as every Android gaming d
Re: (Score:1)
Did you even read the Bio. on the kickstarter link,
did you look at the company link provided?
More like:
1. Design GPU, then sell to S3
2. License design back from S3
3. Produce in the embedded space (aka Profit)
4. Start kickstarter to improve and openly release the Design from #1
Re: (Score:2)
Wrong. None of the projects on opencores have something that's anywhere near this far along and feature-complete.
There are LCD controllers, text mode VGA designs, and one or two framebuffer-level VGA adapters. And framebuffer is essentially garbage.
Re: (Score:2)
This company already has somewhat of a market (air traffic being one, IIRC). But yes, it is niche.
Now, the full product has some nice features like a builtin VNC server...
Re: (Score:3, Informative)
The point of this is NOT to produce a graphics card you'll stick in your PC. The card they're emulating was released in 1998. It's for embedded system designers to add graphics capabilities to a project that's already using an FPGA without completely reinventing the wheel. Speaking as someone who implemented a (very basic) 3D GPU for a class project once, it is quite a lot of work.
Re: (Score:1)
Well if they can find 20 commercial FPGA projects that need a 2D display controller for their FPGA that are willing to fund $10,000 each, then this Kickstarter might get funded.
However I'm sure you can buy 2D cores (proprietary) for less money already.
It is important in some ways to have a 2D LGPL GPU though, especially an optimised, proven design, one with hardware acceleration for drawing, text and blitting - it could benefit a lot of FPGA projects - e.g., FPGA reimplementations of classic hardware, espec
Re: (Score:1)
I guess thats what their betting on.
That enough people see the value in the concept, regardless of an immediate need for the product.
Re: (Score:2)
Re: (Score:2)
Clearly they aren't looking for markets otherwise it would be closed source and they would be proposing to sell IP licenses or actual hardware.
What they are doing is making a freely available Verilog implementation of a graphics chip. This means electronics designers (whether they be hobbyists or people designing embedded hardware for their business) have an accelerated graphics core that they can use with their design. Given it's an LGPL and open source design, they can adapt it to use something other than
Re: (Score:1)
To me it looks like a last-gasp effort to get $200,000 from their old graphics IP, and possibly a nice year or two gig to enhance it.
This is hardware (Score:4, Insightful)
Re: (Score:2)
This fits on some pretty small Xilinx and Altera boards:
Spartan 3, Cyclone II 25, and Cyclone III.
If you look in the right place, $30 is likely to be enough.
Re: (Score:1)
Keep in mind that these guys are former #9 Visual Tech. engineers, so it's not just some shmuck of the street saying lets clone a #9 card and fully open the specs.
They have done hardware before.
Re: (Score:2)
This isn't aimed for plugging in to a PC. It's for embedded devices.
As am embedded developer.. (Score:4, Insightful)
Open reference designs are invaluable.. they allow integration of this into an existing design or SoC, but more interestingly, provide an easy platform to customize at the hardware level.
Re:This is hardware (Score:4)
Not really - PCI (not PCIe) is fairly easy these days (it's 33MHz, so it's not "hard"). And since it's an FPGA, you just need one with a PCI compatible interface. The other parts would be a video DAC to output to VGA or an FPGA with TMDS lines (yes, they make those) to hook to DVI directly.
Connect to a computer is easy - it's PCI, most modern PCs have one. If not, they make PCIe-to-PCI bridges that do the same thing (albeit with more work). Or I'm sure if you look around, there's a reference design card you can have that has PCIe, an FPGA and a variety of connectors and ports for plugging straight into a PC.
And hobby manufacturing is actually fairly easy these days - given how kickstarter seems to have spawned a small industry of contract manufacturers and such with pick and place machines and all that that are reasonably affordable to use to build a small (under 1000, above 20 or so) run with.
No, the biggest problem these guys will encounter is it's impossible to do an open-source GPU. Because everything they need to do is patented, some of it quite heavily (like S3TC - a core part of OpenGL and DirectX these days, of which there is a software and a hardware part, all owned by Via).
And that's just 3D graphics. 2D graphics is also a minefield (stuff like overlays are patented). Or if you want to do hardware assisted video decode (patent minefield! Even if you don't want to do h.264). Since the drivers are probably going to be open, that means pushing a lot of the patented material into hardware.
Hell, the hardware is to be honest the easiest part. Even doing the GPU is fairly straightforward.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Most patents affecting what they have, especially the 2D portion, should be expired by now.
If they aren't, this probably counts as prior art in its own right.
Re: (Score:2)
A lot tougher to get right than software. In software you can implement anything you want, as badly as you want. It doesn't cost anything and it's easy to start over.
readmefirst.txt [amazon.com]
Re: (Score:2)
Its a softcore for an FPGA so you can get as sloppy as you want. Doesnt work? Debug, compile and re-load the bitfile into the FPGA. Then optimize in places that need it. I doubt this is going to make its way into an ASIC unless the design can compete with other embedded IP like the Mali or PowerVR. If you read the article they say specifically they will change the interface from PCI to another bus: "generic interface as well as AXI, Avalon and Wishbone". AXI, Avalon and Wishbone are embedded busses for on c
Re: (Score:1)
You can get away with starter kits.
For one example, Altera has a starter kit with HDMI support. This is for previous generation of their FPGAs (Stratix IV), but still.
What you have to do right, is to write the driver. And here we have a huge task, because no current 3D accelerator card will do any good without optimization of shaders and, more importantly, the combination of shaders and other calls, like texture loading, etc.
It's fine to be fabless (Score:3)
But at least make something that can be fabbed. MIPS and ARM are doing great with implementing reference designs.
Heck, I'd *love* a Chinese factory to take an open graphics card spec and crank out $9 low-power cards for me to use in a server. PCI is fine, but it kills me to put a $50 GeForce card with a heatsink and fan in a normally headless machine!
FPGA is a great target for distributed development; just support those who want to implement it in an an ASIC.
Re: (Score:2)
Why not just buy a motherboard with onboard GPU? Plenty of these around, though the offers are diminishing with the onset of APU generation of CPUs (which you probably don't really want on a server, modulo special cases). I think you are paying around $9 for the graphics card then.
Re: (Score:2)
All server mobos have integrated GPUs.
That would be great if it were so, but it's not.
Re: (Score:2)
No it's your own stupidity that kills you in this case if you buy those for your servers.
Can you can point to something better [newegg.com] or are you just spouting off again, AC?
Re: (Score:2)
> it kills me to put a $50 GeForce card with a heatsink
> and fan in a normally headless machine!
Then take off the heatsink and fan, duh. :-)
Re: (Score:2)
utterly utterly worthless (Score:1)
With the rise of ARM, SoC parts with fully open GPU APIs of amazing power are essentially almost ZERO cost. Tiny circuit boards are available for experimenters and developers with first class 2D, 3D, Video and JPG acceleration, and even video ENCODING is becoming a common hardware feature in low-end parts.
What exactly does the cretinous idea of 'open-source' hardware bring to the table. Sure, it's an environment for budding chip designers to cut their teeth on, but SIMULATIONS on powerful PCs actually provi
Re:utterly utterly worthless (Score:5, Interesting)
With the rise of ARM, SoC parts with fully open GPU APIs of amazing power are essentially almost ZERO cost. Tiny circuit boards are available for experimenters and developers with first class 2D, 3D, Video and JPG acceleration, and even video ENCODING is becoming a common hardware feature in low-end parts.
Care to name any? Most of the ones I've heard of with any form of acceleration are using a proprietary GPU core, where you get a binary blob for Android and bugger-all else. Maybe things have changed since, but last I hard the driver situation was worse for ARM cores than it was in the PC space. Indeed, that was the rationale behind Mir - that it would be able to use the Android blobs under Ubuntu.
Re: (Score:2)
Unlike the Open Graphics Project, this is starting from a working 2D accelerator and mostly working 3D accelerator cloning the features of the Number Nine Ticket to Ride hardware.
Not surprising since one of the reasons Number Nine closed its doors was its failure to embrace 3D acceleration quickly. Still, this is a good thing and I would love to see some one add some really cool graphics to an ARM SoC or something. Wouldn't it be cool to see some hot gfx action out of a system the size of a USB stick?
Re: (Score:2)
No mod points so I'll comment. The binary blobs are generally not updated after a short period - sometimes as little as year. At that point you've effectively been EOL'd by your GPU because it is incompatible with later kernels. Sometimes, as is the case with the IMX6 SOC, the drivers are old when the device is released. Mir helps because android is more likely to get updates but serously... being unable to update your two year old ARM box is super lame.
Fortunately there are efforts to reverse engineer
Re: (Score:1)
Seeing as how I don't have a holographic display, why in hell do I need 3d capabilities in the fscking GPU? The first Pure Linux system I built was based on the Nvidia TNT2-M64 AGP card specifically because of 2d Graphics. Now days, you can't get a fucking card that supports 2d graphics as every fucking thing is 3d this 3d that and 3dnow. Fuck this crap. Give me a solid card that's low power, doesn't need a fucking Reactor just to power my desktop and does over 10k in GLGears like that TNT2 card did and I'
Re: (Score:2, Informative)
These guys have licenses covering the original #9 tech that was sold to S3.
The are former #9 engineers, not as you say non-too-bright fiddlers.
"By mid 2000, S3 had completed the acquisition of Number Nine's assets and Number Nine had ceased operations. In 2002, two former Number Nine engineers, James Macleod and Francis Bruno, formed Silicon Spectrum, Inc., and licensed Number Nine's graphics technology from S3 to implement in FPGA devices."
http://en.wikipedia.org/wiki/Number_Nine_Visual_Technology#Number_N
Re: (Score:2)
None of the GPUs in current ARM SoCs are open in any sense of the word.
Some of them are getting painfully reversed engineered open source drivers - Lima (for Mali), Grate (for Tegra), etc. But the hardware and firmware is still closed source.
This isn't about getting cheap hardware en-masse. It's about getting fully open source hardware at any cost. And the market for that (from a user perspective) is quite small, although the benefits for FPGA computer projects could be massive - essentially a free GPU if t
Who wants this? (Score:2)
Re: (Score:3)
People who fetishize the term "open source" almost to the point of religion.
Re: (Score:2)
Embedded hardware designs. This doesn't get it's own FPGA - it shares an FPGA with whatever else the application requires.
Re: (Score:2)
UltraSPARC cores already fully opensource (Score:1)
Sun released the (IIRC) VHDL code for their UltraSPARC cores used in the T1 and the T2 multicore CPUs.
Re: (Score:1)
I was wondering about open source CPU cores to go along with this GPU - thanks. A better bet than the T80 or T68 core!
Sounds Expensive (Score:2)
I've been out of the FPGA world for a while, but I would expect anything interesting in terms of a 3D chip to require a largeish FPGA chip, and if you have to use a $200-$2000 FPGA (with exotic surface-mount and power requirements) then this seems like maybe not that much of a win.
G.
Re: (Score:2)
Page says the 3D part is optional - you can leave it off to save gates, if you don't need the acceleration.
Re: (Score:2)
Even the highest end Virtex can't touch one of AMD/nVidia's ASICs for 3D rendering, but that isn't the target market for this design.
Here's an evaluation board for a low end FPGA chip, emulating a full 80186 PC including CPU, BIOS, keyboard controller, and VGA video:
http://zet.aluzina.org/index.php/Altera_DE1_Installation_guide [aluzina.org]
The Cyclone II on that card is several years old, a modern equivalent of the FPGA is about $12. Now if you haul the x86 and everything else out, you've got plenty of room freed up to
WTF is a "meelion" (Score:2, Insightful)
That kind of shit should be stripped out by "editors." No slang, please.
Re: (Score:2)
Re: (Score:2)
Ummm (Score:1)
If you want to learn, buy a Xilinx demo board with a VGA or HDMI connector. You can learn something from it, and begin to see what a huge amount of work something like this requires.
Leave your high-minded ideas for open source (or free money from Kickstarter) and get a vague of idea of what you are proposing. The costs to fab anything are enormous. A $35 PCI-Express video card with no fa
How about focus on OpenCL (Score:2, Interesting)
I would prefer OpenCL-only chip, for massively parallel computations. I would like some really simple CPU, like MC68000 + vector FPU, with some 128KB of private memory for each core, with thousands of cores, working on 3GHz or more. It would be even better, if cores are setup in 2D grid, so that each core can access private memory of neighbouring 4 cores.
Re: (Score:2)
Not exactly what you asked for. [parallella.org] Only 1 GHz, and less cores than you asked, but there are already plenty of people complaining that it's too distributed to be useful, so I guess something like that needs to prosper before people try that OpenCL chip.
Patents? (Score:2)
Designing a GPU is much easier than desiginging a GPU which doesn't violate any of the massive number of GPU patents. For a while, it seemed like 'doing something a CPU does but on a GPU' was the patent equivalent of 'doing something computers have been doing for years, but on a phone'.
Drivers? Resource utilization? (Score:2)
Software drivers are a challenge, and we will work on providing some level of drivers, with the hopes that the community takes them up and pushes them to new levels and provides problem reports to us.
I assume the #9 is not nearly as complex as a modern-day GPU, but this sentence really concerns me. As a hardware guy myself, I'd want some more experienced hands on board for the software side. And a Linux driver needs to be part of the deliverables at least.
Also, what's the resource utilization? If this thing only runs in large Virtex-class chips then it's not terribly useful to the open-source community.
It would be really cool to have a graphics core of this level open-sourced, but I think the audience f
Oh wow, Number(N)ine! (Score:2)
I remember having a Number(N)ine card back in the day (an Intergraph card?). It was a little obscure but came with the territory (I think that was the card I paired with a Gravis Ultrasound, as well). Good to see someone using that tech again.
But Why? (Score:2)
They aren't exactly asking for pocket-change here. They want fairly extensive funding in order to produce a product that will be utterly worthless in the marketplace... (running a GPU on an FPGA isn't exactly going to be at the upper-right of the price/perf graph.) It'd be nice if the top contributors at least got a copy of a reference design, not just a USB stick with source on it.
It's nice that it will be "open", but I kind of think they'll have trouble getting funding for a "hobby" project that few, if
Re: (Score:2)
This would be like contributing money to a new car company so they could design (but not actually build) a hand-built, but otherwise generic, mid-size sedan, but with a riding lawnmower engine in it instead a car engine. Yeah, the whole world will get a copy of the blueprints, but why would anyone ever want to build it?
So they could still travel after the NSA backdoored their 'driverless car' and shut it down?
Not really an Open Source development (Score:2)
It's funny, but it seems their development model isn't really open source, they just plan to release the source when they're done.
Mind you, I'm not saying they're lying. They never stated their development model was open source.
What's the big difference?
An open source development model means the source is open from day 0, both for contributions, testing, etc, etc, etc.
This is actually a closed development (ie: behind closed doors), which later gets published.
The quality of the end result of both development
Number Nine, Silicon Spectrum, old IP (Score:2)
This is from Silicon Spectrum, who bought the graphics accelerator from Number Nine (i128, Ticket-to-Ride, etc.). It appears they've deviced to open source that IP. The i128 was mostly 2D with some very limited 3D support. I don't know if they managed to develop anything with programmable shaders, but they might have. Also, I think they've been selling graphics accelerator IP for FPGAs for a while; not sure.