Utah GLX
More info
- Released
- Type
- GLX module for XFree86
- Source
- utah-glx.sourceforge.net
The first attempt to give XFree86 hardware-accelerated OpenGL, around 1999, with every drawing call routed through the X server — and made redundant by DRI within a year.
Before 2000 there was no 3D acceleration in free X. Mesa rendered in software, and anyone who wanted more bought a commercial X server. Utah GLX, named after the University of Utah where the first pieces were written, filled the gap by loading a GLX module into XFree86 3.3 that passed OpenGL calls on to the card. Rendering was therefore always indirect: every call went from the application over the X protocol to the server, which drove the hardware using Mesa as its base. Drivers existed for the Matrox G200 and G400, ATI Rage Pro, Intel i810, S3 ViRGE and NVIDIA Riva, the last built on code NVIDIA itself contributed before the company switched to closed drivers.
The model had a ceiling. Copying everything through the server cost CPU and bus bandwidth, and fast games noticed. Precision Insight chose direct rendering instead: the client talks to the card itself through a kernel module, and the server merely keeps track of whose turn it is. That became DRI, shipped with XFree86 4.0 in March 2000, and Utah GLX went quiet shortly after. Several of its drivers, Matrox among them, made the move across.