Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!brutus.cs.uiuc.edu!lll-winken!decwrl!ucbvax!A.ISI.EDU!ENGLE
From: ENGLE@A.ISI.EDU
Newsgroups: comp.sys.transputer
Subject: (none)
Message-ID: <[A.ISI.EDU].9-Mar-90.15:09:16.ENGLE>
Date: 9 Mar 90 20:09:00 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 30

> From Mark Edmonds:
> Does anyone know if X-windows is supported on any of the graphics TRAMS (or
> boards) available, preferably written in occam.  I heard INMOS were porting
> X-windows to their graphics TRAM - any info available? 

> From Steven Ericsson Zenith
> Yes, X-Windows does run on an INMOS G300 board. I've seen it running,
> it isn't written in Occam (silly :-) ). Helios has  X. It crashes
> of course if you try to open too many windows (no memory management
> remember) but was very fast. Maybe range checking has been added
> since I saw it last, in which case in won't now be very fast.
>

As far as I know, Helios has the only version of X-Windows running on 
transputers.  This is the version that Inmos is showing.   S. Zenith is 
correct in thinking that range checking has been added, it seems that the 
original MIT guys languished in virtual memory and rarely checked their 
mallocs to be sure that they weren't returning NULLS (parents take note,
never let your children get away with unchecked mallocs).

X-Windows under Helios currently supports the Inmos G300 TRAM, the Transtech
graphics board (sorry, its name escapes me), and the Parsytec product, with
additional boards being added all the time.  Given certain conditions, a 
port is fairly straight forward.  X is particularly nice on Transputers 
because of the 2D block moves available, and also the fact that you can
often dedicate a transputer to X (though under Helios you don't have to) and
thus you aren't competing against the X-server for cycles.

Steven W. Engle
MIMD Systems, Inc.