Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!ames!sgi!msc@ramoth.SGI.COM
From: msc@ramoth.SGI.COM (Mark Callow)
Newsgroups: comp.sys.sgi
Subject: Re: SGI licencing of X windows
Summary: From the inside
Message-ID: <32066@sgi.SGI.COM>
Date: 4 May 89 20:41:46 GMT
References: <237@noether.UUCP> <30571@bu-cs.BU.EDU>
Sender: daemon@sgi.SGI.COM
Organization: Silicon Graphics, Inc., Mountain View, CA
Lines: 53

In article <30571@bu-cs.BU.EDU>, eap@bu-cs.BU.EDU (Eric Pearce) writes:
> 
>   I just installed the X source tape we got from SGI.  I don't know if
>   we are paying for it or not.  It seems to have the same problems as
>   the one I got off expo.lcs.mit.edu, but I can't verify that they are
Source tape???  Hmmm..  I'm not sure what tape you are talking about.  We
have been sending out an X Window *binary* on special request at no charge.
This tape is a compiled version of the stuff we put in the R3 contrib
directory.  We were aware of the build problems and we posted a set of patches
to this newsgroup as well as to the consortium.

>   I was rather disappointed in it once I got it running.   This is on
>   an IRIS "power series" GT and it was slower than a Sun 3/50 for most
>   of the clients and demos.  "ico" runs at a good clip, but that's
>   about it.   The server seems to crap out often.  I also had problems
>   with resizing the windows.  I found it useful to be able to xterm
>   into it from my Sun though.  
> 
>   If anybody has a "fast" X11R3 running on their Sgi, I would like to
>   hear about it.
> 
We only became members of the X Consortium a very few weeks before R3
was released. The server you have is based on the infamous cfb code with
just enough changes to make it run on an Iris 4D.  That's all we had time
or since we wanted to get it on the R3 tape.

As you may be aware X was not designed with real graphics machines in
mind.  It takes substantial work to make it fully utilize, where the
imaging model lets it, the 4D hardware.  We have been doing the work.
The about to be released (within 3-4 weeks) Irix 4D1-3.1G software
release includes a new X server that is anywhere from 5 to 100 times
faster than the one you have now.  Our performance comes at or near the
top of the recently posted X benchmark table.  We have applied over 130
MIT patches and fixed a number of our own bugs.  I think you will be
more than satisfied with its performance.  Further performance
improvements will come with Irix 4D1-3.2.

The currently released  microcode for the "Power Series" GTX (*not GT*)
graphics has very poor performance on moving small spans of pixels to
the screen.  (It has great performance for large spans.)  X mostly
deals with small spans so it suffers.  New microcode that fixes the
problem is also in Irix 4D1-3.1G.

On present SGI hardware XOR performance is at best adequate.  We therefore
recommend you run our window manager which does rubberbanding in the overlay
planes (it's written in PostScript, but that's another story) rather
than something like uwm which uses xor all over the place.

>   (In all fairness, I have not asked SGI about these problems yet).
> 
I'm glad you said that.
--
	-Mark