Hello Bas,
vrijdag 01 oktober 1999 04.06, Bas Heijermans wrote to Kris Steenhaut:

 BH> Kris Steenhaut made noise to Bas Heijermans:

 BH> Hi Kris,


 BH> I have used one of the early Dani drivers and that one has screwed my
 BH> desktop totaly, after that happend I never touched or will touch that
 BH> driver again.

You can't be blamed for having an open mind. Well, I'm going to play the "it's 
only you" trick myself: yoy are the only one now having problems with Dani's
drivers..

 BH> It still is, maybe if you stop messing arround so much in OS/2 your
 BH> problems would go away:-)

I'm no messing around: that ONE problem on my system has beel pinpointed to
the Matrox Drivers. And not by me, btw, but by someone having the same amount
of knowledge about these things as Sam Detweiler.
And I take the liberty to believe her instead of you, because what she says
match _exactly_ what is happening on my system.


 BH> Have you ever tried to work with the newer range of ATI, like the 128
 BH> series?

No, I'm talking about stuff you wery well know who has installed it.


 BH>  Well have fun! Ati is very un-cool in the OS/2 area. The S3
 BH> cards are only good where it's the S3Trio series about, they are
 BH> stable and well written, all the rest of them are unpredictable of the
 BH> stability.

I wasn't talking about the at128, but about Mach64, wich you do know very
well.
I do know, unlike the newer Matrox, the S3Savega4 works excellent, even with
S3 propriatary drivers.


 BH>> the butt, drivers and stability. I suspect that de Dani drivers
 BH>> are the major problem with the Matrox problems we see arrise now.

I don't use the Dani driver, as I don't need it, you are very well aware about 
that.


 BH> O please!
 BH> Do you have any idea what you are talking about?
 BH> Shared IRQ's has little to do with the drivers but all with the
 BH> hardware design of the cards used and/or the motherboard. In case you
 BH> don't know most motherboards can turn off or on the IRQ that can be
 BH> assigned to the videocard, it doesn't make much of a diverance either
 BH> way.

YOU don't knwo what you are talking about! The Aopen AX59PRO has no
possibility to reassign the video-IRQ. Look at my previous message, look at
the fucking manual (it's in Spanish, you sure are able to read): the video-irq 
is ALWAYS bouns to the same one as the mass-controler.



 BH>  I grant you that the Matrox 2.22 and the 2.23 drivers are not
 BH> good in some cases, but they certanly won't hog the CPU, they only
 BH> hang the WPS.

Wich is _exactly_ the case here. Not only with 2.22 - 2.23 but 2.22 @ 2.31 GA.



 BH>  The only thing software can do with IRQ's is clearing
 BH> them (CLI instruction) but to let that have much impact the CLI has
 BH> to
 BH> be looped, and I still have to see a programmer who is that stupid to
 BH> use that in their driver. And in case if they have done so it will
 BH> lock your entire system (if even possible with OS/2) that you would
 BH> have to use the reset button to get out of it, even
 BH> ATL-CTRL-NUMLOCK-NUMLOCK won't work anymore, and that's something I
 BH> have never seen in any driver.

_EXACTLY_ what happens here on my system and other fellows I had a mail with.

 BH> If we take a vote on that here in the area you will lose big time:-)

This matters aren't solved by voting, but by reading the readme's 1st and
scrutenizing the system.


 BH> Well here in all of my systems NS give lots of problems, with dropped
 BH> FTP's, crashing randomly, leaving threads open after closing, refusing
 BH> to start, showing pages only after reloads and so on. But only 1 out
 BH> of the 3 systems has a Matrox card, the others have WesternDigital
 BH> 90C31 and a S3Trio64, the results are the same. Even the OS/2 versions
 BH> are diverend, one is Warp 4 US-fp10, the other Warp 4 US-fp9 and the
 BH> last Warp Server Advanced 4 SMP-US-fp42. Also 3 diverend machines, one
 BH> is an AMD-K6-2-350 with 192MB ram, the other an AMD 5x86-133
 BH> (overclocked to 160Mhz) with 48MB ram and the Server has 2x Intel
 BH> PI-133 with 96MB. I have tried NS202, 404 and the latest 4.61, all
 BH> with the same problems. The only program that's giving me big problems
 BH> on all of them is NS, so why should I be convinced that it has
 BH> anything to do with Matrox??????

Do take a test: switch the Matrox drivers by the SDD, and you'll see the
difference. There is a new beta7 out since yesterday, btw, will install them
within a few houres.

 BH> and it has the same problems. The only good versions of NS that I have
 BH> seen so far are the ones for BeOS and Linux. This brings me to the
 BH> conclusion that NS is crap, allmost any version.

Even with Matrox drivers  loaded, the NS461 behaves well. Nearly all of the
problems are gone, except for the Alzheimer-bug: NS went down a few times
since I had it installed. But the FTP-http- and other files transports are
fine and well: last night I downloaded the 4 Mg SDD drivers withhout a hitch
at 4.3 kb, and days before I had the same experience with even larger files.




    Groeten uit Gent,
    Regards/2

      Kris

--- GoldED/2 3.0.1  FMail/2 1.48/g