Into a third dimension
We add some depth to the story and learn about the first, bona fide, GPU.
T he first consumer 3D graphics cards appeared under 3DFX’s “Voodoo” branding in late 1996. These were add-in cards that still required a regular video card (or chip) to display 2D scenes. But they enabled games like
Tomb Raider, Descent, Grand Theft Auto
and
Unreal
to use 3DFX’s proprietary Glide API and render things (well, some things) much faster than software rendering. A number of other proprietary 3D graphics APIs appeared around this time, including the S3 3D Engine (used in Virge cards), the Nvidia Multimedia Library, the Creative 3D Library (used in Blaster 3D devices) and the even more imaginatively titled ATI 3D C Interface (for 3D Rage hardware).
Silicon Graphics, makers of expensive workstations, also had a proprietary API, in the form of IRIS GL. This could do 3D things, but was also much more general purpose (it was also a windowing system). There was industry demand for parts of this API to be licenced (and competition from the likes of Sun and IBM), and in 1992 SGI tailored a version of IRIS GL into a new open standard called OpenGL. They also organised the OpenGL Architecture Review Board (OpenGL ARB), a group of companies that would maintain the specification in future. Today, OpenGL is steered in much the same way by the Khronos consortium.