Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!cs.utexas.edu!uunet!mcsun!sunic!tut!santra!hemuli.atk.vtt.fi!tml
From: tml@hemuli.atk.vtt.fi (Tor Lillqvist)
Newsgroups: comp.windows.x
Subject: Re: Xtank
Message-ID: <4292@hemuli.atk.vtt.fi>
Date: 1 Nov 89 06:15:21 GMT
References: <127042@sun.Eng.Sun.COM>
Reply-To: tml@hemuli.atk.vtt.fi (Tor Lillqvist)
Organization: Technical Research Centre of Finland
Lines: 27

In article <127042@sun.Eng.Sun.COM> naughton@sun.com (Patrick Naughton) writes:
>
>Has anyone worked out all of the problems running Xtank on Suns? Has the
>multi-thread package been ported, or does the RT version work? I spent

This must be the first time I come across a program that is easier to
port to the HP9000 Series 800 (the HP Precision Architecture RISC
machines) than to Suns....  Usually all interesting free software is
developed on Suns or VAXen, and there is some effort needed to port it
to to the hp9000s800.  After staring at the thread library source for
maybe an hour I got enlightened and understood how it was supposed to
work (I think).  After that it was easy to get Xtank running.  The
dynamic loading code had to be rewritten of course, but luckily I had
saved some articles from comp.sys.hp discussing dynamic loading.  The
xtank source compiled without any warnings, really amazing.  A good
job, congratulations to the author!

Now if I only could find out where it gets all these math errors
(DOMAIN error, TLOSS error).  And usually it dies with a floating
exception after running OK for perhaps ten minutes.  The code seems to
use trigonometric functions rather heavily, I think I will profile it
and maybe use tables instead.

I will send the hp9000s800 diffs to the author in a few days, and put
them up for ftp here.
-- 
Tor Lillqvist, VTT/ATK