🚀 go-pugleaf

In <3jdq68$j...@golem.wcc.govt.nz> harri...@kosmos.wcc.govt.nz writes:

>Have just played Descent on a network of Pentium Overdrives at m
>old school, WAY COOL - 8 player descent and me the only one with a
>sound card (my own from home) JOY.! Well paralax <sp?> I am VERY  
>IMPRESSED! All I have to do now is get a computer of my own good 
>enough to play it, buy the registered version, get a faster modem,
>and then find someone else with all of the above to play against! 
>Anyone else have these problems?

Yeah,  Pentuium would be nice but it will be a P6 by the time I upgrade.  
Let me just state that a good 486 works just fine for Descent.  Memory 
and graphics cards are the bottle necks.  :D

-- 
                      
Dave Chernetz                       "-and the truth shall
cher...@mbnet.mb.ca                  make you flee."        --S. Robinson
                    

Excerpts from netnews.alt.games.descent: 6-Mar-95 Re: Descent 8-way on
pents! by Dave Chernetz@access.mbn 
> Yeah,  Pentuium would be nice but it will be a P6 by the time I upgrade.  
> Let me just state that a good 486 works just fine for Descent.  Memory 
> and graphics cards are the bottle necks.  :D

    Not true...  When you have 8 people playing over an IPX network, the
486 DX/2 66's are a BASELINE machine...  Even then you have to turn all
the detail down, and make your viewpors as small as possible...  The
network communications take too much of the processor...  I am saying
this based on my experience with my 486/66 that has 32MB ram and a
stelth 64 with 4MB vram...  Descent over the network really does require
a Pentium...

 ________________,^-.                _______________________________   
|                   ,>              |   Ethan Connor                |
|     Arlington->* {                |   Local:  x2-2387             |
|        MA         ?     )         |   Email:  ec...@andrew.cmu.edu |
|___ _____________   `--~ ;         |   IPX:    128.2.46.133        |
    "              \_,-__/           -------------------------------

In article <IjKq3XG00...@andrew.cmu.edu> Ethan M Connor <ec...@andrew.cmu.edu> writes:
>From: Ethan M Connor <ec...@andrew.cmu.edu>
>Subject: Re: Descent 8-way on pents!
>Date: Mon,  6 Mar 1995 14:22:11 -0500

>Excerpts from netnews.alt.games.descent: 6-Mar-95 Re: Descent 8-way on
>pents! by Dave Chernetz@access.mbn 
>> Yeah,  Pentuium would be nice but it will be a P6 by the time I upgrade.  
>> Let me just state that a good 486 works just fine for Descent.  Memory 
>> and graphics cards are the bottle necks.  :D

>    Not true...  When you have 8 people playing over an IPX network, the
>486 DX/2 66's are a BASELINE machine...  Even then you have to turn all
>the detail down, and make your viewpors as small as possible...  The
>network communications take too much of the processor...  I am saying
>this based on my experience with my 486/66 that has 32MB ram and a
>stelth 64 with 4MB vram...  Descent over the network really does require
>a Pentium...

I know my setup isn't the fastest but come on... I've played on a network on a 
386.  True, the guys with the 486's kicked my butt...

But, the only reason you would *have* to use a Pentium on the network is to 
keep up with the competition.

Randy Lisle
rdl...@cris.com
(speaking for myself)

Ethan M Connor (ec...@andrew.cmu.edu) wrote:
> Excerpts from netnews.alt.games.descent: 6-Mar-95 Re: Descent 8-way on
> pents! by Dave Chernetz@access.mbn 
> > Yeah,  Pentuium would be nice but it will be a P6 by the time I upgrade.  
> > Let me just state that a good 486 works just fine for Descent.  Memory 
> > and graphics cards are the bottle necks.  :D
>     Not true...  When you have 8 people playing over an IPX network, the
> 486 DX/2 66's are a BASELINE machine...  Even then you have to turn all
> the detail down, and make your viewpors as small as possible...  The
> network communications take too much of the processor...  I am saying
> this based on my experience with my 486/66 that has 32MB ram and a
> stelth 64 with 4MB vram...  Descent over the network really does require
> a Pentium...

Oh, come on - just because you're used to your machine, doesn't mean that
lower level critters aren't playable. We've got 486/33's, with 4meg and 512K
built-in VGA running 8 player games over cheap NE2000 clones, and they play
exactly no slower than playing solo on the same machines.

(The machines are University undergrad ones, in Britain, which is why they're
such cheap ones)

There's something wrong with your network, I'd hazard a guess...

BTW, since when was a 4meg video card necessary for 320x200x256? Hmm? I like
the Stealth, but you're just showing off now!

In article <3je3qb$b...@canopus.cc.umanitoba.ca>, 
cher...@access.mbnet.mb.ca says...
>Let me just state that a good 486 works just fine for Descent.  Memory 
>and graphics cards are the bottle necks.  :D

If you like being a sitting target, sure a 486 works fine. (brag session 
to follow) There's not a single person that could beat me using a 486 when 
I'm on my P5-90.

Mr Jaimie Vandenbergh (ch11jv@PROBLEM_WITH_INEWS_GATEWAY_FILE) wrote:

: Oh, come on - just because you're used to your machine, doesn't mean that
: lower level critters aren't playable. We've got 486/33's, with 4meg and 512K
: built-in VGA running 8 player games over cheap NE2000 clones, and they play
: exactly no slower than playing solo on the same machines.

	Didn't I read somewhere that net-Descent runs at YOUR computer's 
speed, that it doesn't slow down to the speed of the slowest computer 
that's playing the net-game, unlike DOOM which does, in fact, slow down 
to accomodate the slower machine...?

In article <IjKq3XG00...@andrew.cmu.edu>, Ethan M Connor <ec...@andrew.cmu.edu> says:
>
>Excerpts from netnews.alt.games.descent: 6-Mar-95 Re: Descent 8-way on
>pents! by Dave Chernetz@access.mbn 
>> Yeah,  Pentuium would be nice but it will be a P6 by the time I upgrade.  
>> Let me just state that a good 486 works just fine for Descent.  Memory 
>> and graphics cards are the bottle necks.  :D
>
>    Not true...  When you have 8 people playing over an IPX network, the
>486 DX/2 66's are a BASELINE machine...  Even then you have to turn all
>the detail down, and make your viewpors as small as possible...  The
>network communications take too much of the processor...  I am saying
>this based on my experience with my 486/66 that has 32MB ram and a
>stelth 64 with 4MB vram...  Descent over the network really does require
>a Pentium...


Actually, what kind of network card do you have? I KNOW FOR CERTAIN that
a 486/66 works very well even with just a Steath 24 w 1meg DRAM, and 8 megs
memory... It could be better, but that is mostly my graphics  card and RAM.
If you have a cheap network card, then that is probably your problem.  The group
with whom I play with all have 3COM Etherlink III cards, and as I said the game 
works perfectly well.  Truthfully, the game even runs descently (pun not intended)
on the 486-33 on our network, and it's and IBM (<- read Slow for a 33).
If you have cheap NE2000 compatible cards or something like that, there is your
problem.  I suggest getting 3COM or Intel cards for replacement.


Todd Wagner
--todd...@vt.edu

fr...@genesis.demon.co.uk (Lawrence Kirby) writes:
>That seems excessively high, what are the conditions that generate that?
>I've done some detailed serial port load testing (though clearly not with
>Descent) Under SCO Unix. If you run a 16550A serial port at a
>*continuous* 38400bps the system load on a 486/33 is about 5%. (2 player)
>Descent uses a vastly lower bandwidth than that so I'd venture it should
>generate about a 1% overhead or there is something wrong with the driver.
>With an 8250/16450 you could see about 10% overhead.

The difference is that the current serial driver in Descent uses a polled
output routine rather than an interrupt driven one.  If we used interrupt
driven serial output I think your figures would be just about right.

As to why Descent uses polled output, it has to do with the Greenleaf
serial I/O lib used in the game.  Probably future Descent games will
have a more efficent serial send routine utilizing interrupt driven output.
----
Rob Huebner