From: Richard Humphrey rlhumphrey@gmail.com

Thanks that did the trick

Although nVidia does put people on the Freedomware "nv" driver, and it can handle multiple framebuffers for Xinerama, VIVO (video in, video out), etc..., the Standardware "nvidia" driver is far easier to deal with for single framebuffer (no Xinerama) multiple displays, VIVO, etc..., in addition to adding GLX (OpenGL over X11) support.

-- Bryan

Political Notes:

nVidia actually released the GLX code for the GeForce series back in the XFree86 3.3.x days, and obscuficated some of the code provided by 3rd parties (Intel's AGP trade secrets, SGI and Microsoft patented OpenGL, etc...). Unfortunately the community quickly started reverting the obscuficated code to the formal identifiers and other blocks that the 3rd parties didn't like to see, and lawsuits were threatened. So nVidia decided that it would no longer provide such code on the GLX portion once XFree86 4.0 came out with the ability to use binary drivers. The kernel memory-AGP interface driver portions are largely Intel's trade secrets, although it seems that is actually going GPL as Intel no longer considers AGP a trade secret, and PCIe (PCI-Express) is a formal PCISG standard.

ATI should be also commended for attempting to build a "clean room" DRI/GLX implementation. But as lack of resources to continue to "fork" mounted along with lagging performance, ATI had to follow in nVidia's footsteps by creating a closed-source, but standards-driven unified driver by withholding specifications as of the R300 series (Radeon 9500+). Just like nVidia, the "clean room" GLX work is still available, and might even work on newer GeForce/Radeon cards, but the newer features of those cards are not documented.

Lastly, it should also be noted that Matrox also has closed source drivers as well. If there is an industry that Microsoft is poised to strike the hardest on patents, it's OpenGL. They've con'd most of the graphics companies into giving them non- exclusionary rights to them -- especially SGI, the inventor.

-- Bryan J. Smith mailto:b.j.smith@ieee.org