i'm trying to consolidate what i know about hardware acceleration and
graphics, and i'm going around in circles...
opengl is an API that allows a programmer to do 2D/3D graphics. the most
important implementation of opengl as far as we're concerned is mesa. any
given implementation may or may not be hardware accelerated. the important
point is that opengl is device independent.
glide is another graphics API (which also happens to be an implementation
since it's written solely by 3dfx graphics inc). glide is NOT opengl,
although some of its functions look like opengl. glide functions make use
of the hardware acceleration capabilities of voodoo boards.
mesa is a free implementation of opengl (leaving legalities out of the
discussion. it aims to be an opengl implementation).
yet i've seen people say that mesa is also a drop-in replacement for glide.
there are versions of mesa which have 3dfx drivers, which presumably use
3dfx acceleration.
so does mesa-glide use hardware acceleration for opengl functions? or does
mesa-glide implement both opengl and glide?
if mesa-glide is a "drop in replacement" for glide, then the latter has to be
true, i think.
all this discussion is independent of
1- the actual device driver for the card itself.
2- the X server, which has to have support for the card.
now, DRI == direct rendering infrastructure allows X clients to write 3D
rendering information to the video card. this can be hardware accelerated
or not.
glut is a windowing toolkit system, like gtk, but implements opengl.
finally, utah glx is the device dependent opengl implementation for xfree86.
i'm sure a lot of this isn't right. does anybody have this stuff down?
i'd like to have a clearer big picture of all this stuff.
thanks!
pete
