Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.mira.net.au!news.vbc.net!news.cais.net!news.jsums.edu!gatech!newsfeed.internetmci.com!news.msfc.nasa.gov!sol.ctr.columbia.edu!startide.ctr.columbia.edu!wpaul
From: wpaul@ctr.columbia.edu (Bill Paul)
Newsgroups: comp.os.linux.development.apps,comp.os.linux.development.system,comp.os.linux.x,comp.os.linux.hardware,comp.os.linux.setup,comp.unix.bsd.386bsd.misc,comp.unix.bsd.bsdi.misc,comp.unix.bsd.netbsd.misc,comp.unix.bsd.freebsd.misc,comp.sys.ibm.pc.hardware.video
Subject: Re: Matrox Millennium and hardware acceleration
Date: 10 Apr 1996 16:47:13 GMT
Organization: Columbia University Center for Telecommunications Research
Lines: 79
Message-ID: <4kgomh$6j5@sol.ctr.columbia.edu>
References: <4j21ph$crr@slappy.cs.utexas.edu> <4ja1g6$ddj@hiekkalaatikko.cs.hut.fi> <4jolnu$a42@idefix.eunet.fi> <3160A88F.2153@engin.umich.edu> <4kdq9i$mgm@ninurta.fer.uni-lj.si>
NNTP-Posting-Host: startide.ctr.columbia.edu
X-Newsreader: TIN [version 1.2 PL2]
Xref: euryale.cc.adfa.oz.au comp.os.linux.development.apps:14318 comp.os.linux.development.system:21012 comp.os.linux.x:29004 comp.os.linux.hardware:35925 comp.os.linux.setup:49795 comp.unix.bsd.386bsd.misc:515 comp.unix.bsd.bsdi.misc:3084 comp.unix.bsd.netbsd.misc:2864 comp.unix.bsd.freebsd.misc:17013 comp.sys.ibm.pc.hardware.video:56353
Daring to challenge the will of the almighty Leviam00se, Arijan Siska
(arijan@news.fer.uni-lj.si) had the courage to say:
: Jimmy Wan (vecna@engin.umich.edu) wrote:
: : Well, for one thing, NT isn't exactly a mainstream platform(yet).
: : Don't judge them by just their NT drivers. Their Win 3.x drivers are
: : stable as hell, and blazing fast.
: I have problems with Matrox Millenium drivers 1.2 from their ftp site:
: - Photoshop crashes when using "hand" tool
: - cinepack avi files crash if rescaled
: Is this stable?
: Arijan
Well, that depends on your perspective. From the vendor standpoint,
I'm sure the code quality is acceptable, but that's only because they
never really test drivers as thoroughly as they should. To be fair,
they can't easily test their drivers with every possible PC hardware
configuration and every software package on the market -- they'd be
bogged down for years. And don't forget that Windows itself is broken
in a couple of fundamental ways which even the best driver developer
can't work around. Even so, they should test more than they do.
From a user standpoint, particularly a user who can't work because they
just happen to have the magic hardware/software combination that causes
the drivers to eat themselves, it just plain sucks. And it shouldn't
be tolerated either.
Unfortunately, here you have a situation where a couple of applications
that blow up rather than the whole OS (and I use the term OS loosely since
we're talking about Windoze here). Now you have to figure out who's
responsible: the application, the OS, the drivers or the hardware. If
you go to the board vendor and say "Hey: this program crashes with your
graphics card," you'll have to prove to them that it really is their
card and/or driver that's responsible before they'll even begin to do
anything about it. So how do you prove it? Well, things would be easier
with source code, but of course they don't want to let you near that,
othwerwise you might discover the awful kludges they had to perpetrate
to get everything to work. At best, you have to spend time trying different
hardware and software configurations until you can reproduce the problem
reliably and prove that it goes away if you use a different card.
Being able to say "Well gee, this TurboGronkulator 1000 card is a lot
more reliable than yours -- I think I'll just keep it and return your
card" might just make them act faster.
I had the pleasure of installing a Parallax Sbus video board in a
SPARC 20 with Solaris 2.5 the other day. This board has its own frame
buffer which you must use to display incoming video. This requires both
a kernel driver and a special ddx shared object for the Sun X server.
The software that we got came with a stack of yellow sheets marked
'Release notes' that outlined several open bugs which the developers
haven't fixed yet. One of them, which I find both humerous and
heartbreaking, states that if you run Netscape (*spit*) and try to
resize it by dragging anything except the lower righthand corner, the
X server will crash.
What I can't figure out is how the software came to be released that way.
They could claim that the problem is Sun's fault, but that doesn't explain
why it doesn't happen with Sun's drivers and ddx shared objects. If it was
a bug that just slipped by quality control, then I have to ask just
what sort of quality control is being used here.
There are times when I see this and I wish for just five minutes alone
in a room with the driver source code or, barring that, the guy who wrote
it.
-Bill
--
=============================================================================
-Bill Paul (212) 854-6020 | System Manager, Master of Unix-Fu
Work: wpaul@ctr.columbia.edu | Center for Telecommunications Research
Home: wpaul@skynet.ctr.columbia.edu | Columbia University, New York City
=============================================================================
"If you are in trouble, go the CTR. Ask for Bill. He will help you."
=============================================================================