The Open-Source Graphics Card Is Dead
Prior to LinuxTag Berlin later in the week, I have been visiting with Egbert Eich, the SUSE engineer, long-time X.Org developer, and former RadeonHD driver developer. Among the many Linux graphics topics being discussed in Frankfurt-Darmstadt, Egbert and I realized "that project to come up with an open-source graphics card" hadn't been heard of in years. Hell neither of us could recall the name of the main project even though it was presented just four years ago at FOSDEM.
The project in question was Project VGA. Project VGA hoped to ship in 2008 (just one month after FOSDEM) as what was slated to be a low-budget, open-source, VGA-compatible graphics card. The initial Project VGA hardware was said to be PCI-based (i.e. the standard PCI bus, not PCI Express), run on a 100MHz FPGA, and to have 16MB of SDRAM. Even for back then the hardware specifications were quite bad, but they were basically limited to using a slow Xilinx FPGA as their graphics processor, there was very limited funds, and also very limited contributors / experienced hardware engineers.
Beers are needed, of course, when talking about the former RadeonHD days...
All of the Project VGA software and hardware was to be made available under the GPLv3 license. The first-run Project VGA graphics cards were expected to retail for some $200 USD but without any warranty -- or that the PCI graphics card with a 100MHz FPGA would even work in the first place. Well, since FOSDEM of 2008, it was never really hard from again. After Egbert and I recalled this project, we attempted to track it down to no avail.
The project's web-site (ProjectVGA.org) no longer resolves and there's no signs of it transforming into anything else, it was just another open-source project that died. This isn't a terrible surprise considering the slow hardware making it really not viable for anything but enthusiasts/developers wishing to tinker with an open-source GPU, die-hard open-source zealots wishing for an open-source GPU, etc. Unfortunately most community-based open-source hardware projects for complex components such as GPUs are poised for failure due to the cost of design and manufacturing / large orders being mandated. As well, the increasing complexity, experience, patent access, and other abilities of the leading IHVs with proprietary designs makes it out of reach to have a product that would be competitive with consumers and attract real attention/interest/sales. (Yes though there are "open hardware" possibilities in other less-complex forms such as with the VGA switch or ColorHug.)
Another dead topic to reminisce over was the initatives for bringing audio to the X Server.
This effort is basically kaput.