[comp.sys.sgi] X-windows

mccalpin@masig3.ocean.fsu.edu (John D. McCalpin) (12/15/89)

I am considering buying a 4D/25-turbo, and wondered what experiences
people have had with X-windows on this box (or similar boxes).  I have
heard that the new machines have a faster raster chip which should
help the X performance noticeably....

I want to run X so that I can get more software (NCAR graphics, NCSA
XDS tools) and so that I can run the same system on remote X-windows
terminals (NCD).  Does the 3-D graphics system run well under X?

--
John D. McCalpin - mccalpin@masig1.ocean.fsu.edu
		   mccalpin@scri1.scri.fsu.edu
		   mccalpin@delocn.udel.edu

ssr@gauss.Berkeley.EDU (Steve S. Roy) (12/17/89)

> Does the 3-D graphics system run well under X?

The 3-D graphics system does not run at all under X.  X, at this time, 
is purely 2-D, and an X program running on an Iris sees ordinary X.

Here is something of what is going on, as I've been able to piece it together.

The Iris runs its own windowing system, 4Sight, which knows nothing of X, but
it does know a lot about 3-D.  To run X you start a program in the background
called Xsgi and then run your program that expects an X environment.  Somewhere
in the initialization of your program a link is made between it and the Xsgi,
which then links to 4Sight.  When your program wants to draw something it executes
an X command, which knows nothing of 3-D, and the request goes thru Xsgi to 
4Sight and gets drawn.

I do not know of any way to do both X and native drawing in the same window,
but one program can open two windows, one under X and one under 4Sight and do
whatever is appropriate in each.

Some people are working on something called PEX (Phigs Extensions to X) which
is intended to allow 3-D drawing in X, and when that's available presumably
you will be able to draw in 3-D, but still not with the native calls and there
will probably be a perfomance hit.  I have no idea of the timescale involved
with this.

----------------------------------------------------------------------
Real Name: Steve Roy
e-mail: ssr@acm.princeton.edu
telephone: (609)258-5324 (office)
Address: office:
         Program of Applied and Computational Mathematics
         Princeton University
         Princetion NJ 08544-1000
----------------------------------------------------------------------