Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Why I'm going with 3DFX and Voodoo...

42 views
Skip to first unread message

Eric Calcagni

unread,
Jun 29, 1996, 3:00:00 AM6/29/96
to
Ok. I've been researching heavily into 3D cards in the past several
months. I have seen all the specs, all the gee-wiz items and
useless polygon counts. After all of this, I have decided I am going
to get the Voodoo chipset. They seem to have a winning chipset
on their hands (I hope :) But I just wanted to add...

3DFX has won my over by their wonderful technical support online.
Every e-mail I have sent, they have answered with a personal answer.
EVERY one. And not once have I gotten the hard sell from these people.
If anything, they have been very modest and told me what they think
the advantages are of some other 3D cards as well. Finding a company
that will do that is almost impossible.

In fact, I have gotten better support from 3DFX than from any company
I have ever purchased anything from - and I don't own any 3DFX stuff
yet.

I know I'll eventually have to deal with technical support from
the companies that actually make the video cards that support
have the voodoo on board, and I'm not expecting much from the current
offerings. But I am so impressed by 3DFX's complete and sound answers
to my questions that they have my business.

Plus, the voodoo will probably be the hottest thing out there (next to
PowerVR and Read 3D). ;)

So if you guys are looking for a company that seems to care about
this 3D "revolution" we are in, I'd suggest you ask them a few
questions. Damn nice people.

Later - Eric C.

Brad

unread,
Jun 30, 1996, 3:00:00 AM6/30/96
to calc...@3-cities.com

Eric Calcagni wrote:
>
> Ok. I've been researching heavily into 3D cards in the past several
> months. I have seen all the specs, all the gee-wiz items and
> useless polygon counts. After all of this, I have decided I am going
> to get the Voodoo chipset. They seem to have a winning chipset
> on their hands (I hope :) But I just wanted to add...
>
> 3DFX has won my over by their wonderful technical support online.
> Every e-mail I have sent, they have answered with a personal answer.
> EVERY one. And not once have I gotten the hard sell from these people.
> If anything, they have been very modest and told me what they think
> the advantages are of some other 3D cards as well. Finding a company
> that will do that is almost impossible.
>
> In fact, I have gotten better support from 3DFX than from any company
> I have ever purchased anything from - and I don't own any 3DFX stuff
> yet.
>
> I know I'll eventually have to deal with technical support from
> the companies that actually make the video cards that support
> have the voodoo on board, and I'm not expecting much from the current
> offerings. But I am so impressed by 3DFX's complete and sound answers
> to my questions that they have my business.
>
<snip.........>


The only problem is that the Verite chipset by Rendition is supported
directly by "Microsoft" and all the major CAD and PC game developers. I
was lookin at the Virge chipset till I talked to many people and read up
on the Verite chipset. If your at all interested in games and interested
in getting the "standard" chipset that will be widely supported for a
long time to come then I personally would go with the Verite chipset.

BTW .. I really dont see what the huge deal is with tech support or
customer service with a video board. I mean you basically get it up ans
running and thats it besides getting the latest drivers from the net
--
("`-''-/").___..--''"`-._
`6_ 6 ) `-. ( ).`-.__.`)
(_G_.)' ._ ) `._ `. ``-..-'
_..`--'_..-_/ /--'_.' ,'
(il).-'' (li).' ((!.-'
br...@isci.net
http://206.26.158.13/free/blj19.html

Daron Myrick

unread,
Jun 30, 1996, 3:00:00 AM6/30/96
to

|> The only problem is that the Verite chipset by Rendition is supported
|>directly by "Microsoft" and all the major CAD and PC game developers. I
|>was lookin at the Virge chipset till I talked to many people and read up
|>on the Verite chipset. If your at all interested in games and interested
|>in getting the "standard" chipset that will be widely supported for a
|>long time to come then I personally would go with the Verite chipset.

I like the Verite chipset by Rendition also.
Why?
3DFX is made by a company I've never heard of till a couple of months ago that
seems to not have any history, while Rendition has been around from the good
old 286 days and has been making top of the line, extremely bleeding edge
cards since the 286 came out.
Rendition has been doing 3D hardware acceleration since the early 1980's so
they know their stuff to where they can do it in their sleep.
Which is faster, I don't know but I would feel more comfortable with Rendition
since I know they are going to be there for us, with often
software/driver/firmware updates.
I speak from experience since many times have I bought computer products made
by "no namers" that was the fastest at the time, but had no solid programming
team, no solid hardware team and the like, so their drivers were either buggy,
slow to release or nonexistant.
The above is mainly aimed at advanced gravis.

Chris Ciccarello

unread,
Jun 30, 1996, 3:00:00 AM6/30/96
to

Brad wrote:
>
> Eric Calcagni wrote:
> >
> > Ok. I've been researching heavily into 3D cards in the past several
> > months. I have seen all the specs, all the gee-wiz items and
> > useless polygon counts. After all of this, I have decided I am going
> > to get the Voodoo chipset. They seem to have a winning chipset
> > on their hands (I hope :) But I just wanted to add...
> >
> > 3DFX has won my over by their wonderful technical support online.
> > Every e-mail I have sent, they have answered with a personal answer.
> > EVERY one. And not once have I gotten the hard sell from these people.
> > If anything, they have been very modest and told me what they think
> > the advantages are of some other 3D cards as well. Finding a company
> > that will do that is almost impossible.
> >
> > In fact, I have gotten better support from 3DFX than from any company
> > I have ever purchased anything from - and I don't own any 3DFX stuff
> > yet.
> >
> > I know I'll eventually have to deal with technical support from
> > the companies that actually make the video cards that support
> > have the voodoo on board, and I'm not expecting much from the current
> > offerings. But I am so impressed by 3DFX's complete and sound answers
> > to my questions that they have my business.
> >
> <snip.........>
>
> The only problem is that the Verite chipset by Rendition is supported
> directly by "Microsoft" and all the major CAD and PC game developers. I
> was lookin at the Virge chipset till I talked to many people and read up
> on the Verite chipset. If your at all interested in games and interested
> in getting the "standard" chipset that will be widely supported for a
> long time to come then I personally would go with the Verite chipset.

Yes, the Rendition card is the "reference" platform for Direct 3D, although I'm not sure that means much. It
doesn't mean that other cards can't support Direct 3D as well or better. In fact from what I've read lately
(CGW, PC Gamer), 3DFX is also getting a heck of a lot of support (as much as Rendition) and developers are
excited about it too. Hopefully Direct 3D will catch on and the best performing card will be the best card to
own, and right now that card looks to be 3DFX.

-Chris

Daron Myrick

unread,
Jun 30, 1996, 3:00:00 AM6/30/96
to

|>What? These companies (the independent startup 3D chip makers) have been
|>around since only 1993 or 1994. Verite is the first Rendition product they
|>started in 1993. I do believe there was some other company called Rendition
|>or somesuch, but this isn't it.

Your right Fred.
Boy I feel stupid.
Anyhow, there was a company called rendition or did #9 used to make a board
called the rendition?
Anyhow, I'll take my comments back since both companies are new.
I like the technology behind the Verite but the Voodoo is generaly faster.
What about the GLINT series?
I remember visiting a WEB site and there was a 3D accelerator chip for PC use
that does over 1 billion pixels a second copy, move and fill(nontextured)
rate.
Wish I could remember that site.

Daron Myrick

unread,
Jun 30, 1996, 3:00:00 AM6/30/96
to

|>Theoretically: If both products are released at approximately the same
|>time, yet the 3DFX delivers twice the performance of the Verite, are
|>you going to go with Rendition because they've been around longer? I
|>really doubt it.

I correct myself by stating that the Rendition I was thinking of was not the
Rendition company that makes the Verite.
As far as which company is newer, Rendition is but only by less than a year.
I'm sold on the Voodoo after visiting their site.
Too bad we have to buy an Orchid card for Voodoo power since I always hated
Orchid.
I'm tempted to buy the Voodoo card make by 3DFX themselves since at least I
know I won't be getting support and drivers from a company that's one step up
from Diamond.
Hopefully the card will be in the $500 to $600 range so that I can afford it
though it's still quite steep.

Eric Calcagni

unread,
Jun 30, 1996, 3:00:00 AM6/30/96
to

Brad wrote:

> The only problem is that the Verite chipset by Rendition is supported
> directly by "Microsoft" and all the major CAD and PC game developers. I
> was lookin at the Virge chipset till I talked to many people and read up
> on the Verite chipset. If your at all interested in games and interested
> in getting the "standard" chipset that will be widely supported for a
> long time to come then I personally would go with the Verite chipset.

Just because the Verite chipset was used in the design of Direct 3D
doesn't make it the best chipset to use with Direct 3D. Many 3D boards
will have the same features as the Verite but do them better and faster
using Direct 3D.

I can say the the Verite chip is the only chip I have ever seen in
use. I saw it running Terminal Velocity on a television program
(the piece was on the Rendition company and the Verite). I liked
the way it looked, but it wasn't very fast in high res. Of course,
that was a few months ago, so maybe they have improved it since, or
perhaps it was a simple port.

The Voodoo chipset is the only chipset I have heard of that people
(who have seen it) said it knocked their socks off. No other
product has done this to my knowledge. In fact, many people said the
3DFX chipset blew all the other 3D accelerators out of the water
at E3.

>
> BTW .. I really dont see what the huge deal is with tech support or
> customer service with a video board. I mean you basically get it up ans
> running and thats it besides getting the latest drivers from the net
> --

Chances are I won't use technical support for my 3D video card. But
I am really impressed with 3DFX's desire to answer questions and
not give me a canned answer like many other companies. This shows
me that they are interested in the market they are penetrating and
will probably translate this attitude into better products manufactuered
by them.

- Eric C.

Frederick Y Mah

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

dar...@onramp.net (Daron Myrick) writes:


>I like the Verite chipset by Rendition also.
>Why?
>3DFX is made by a company I've never heard of till a couple of months ago that
>seems to not have any history, while Rendition has been around from the good
>old 286 days and has been making top of the line, extremely bleeding edge
>cards since the 286 came out.
>Rendition has been doing 3D hardware acceleration since the early 1980's so
>they know their stuff to where they can do it in their sleep.

What? These companies (the independent startup 3D chip makers) have been


around since only 1993 or 1994. Verite is the first Rendition product they
started in 1993. I do believe there was some other company called Rendition
or somesuch, but this isn't it.

--
Fred Mah --- fm...@ecn.purdue.edu
http://widget.ecn.purdue.edu/~fmah

Dave Glue

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

On Sun, 30 Jun 1996 10:24:57 -0700, Brad <br...@icsi.net> wrote:


> The only problem is that the Verite chipset by Rendition is supported
>directly by "Microsoft" and all the major CAD and PC game developers.

The Voodoo actually has the longest list of direct developer support.

EVERY chip is supported "directly" by Microsoft. The Verite was
simply the "base" platform for Direct3D because it was one of the
first decent-performing accelerators out at the time Direct3D was
being designed. The Voodoo will do everything the Verite will do, but
more- too bad the current incarnation of Direct3D can't take full
advantage of it, but that points to the 3DFX's expanded feature set,
not to any liability on its part compared to the Verite.


>was lookin at the Virge chipset till I talked to many people and read up
>on the Verite chipset. If your at all interested in games and interested
>in getting the "standard" chipset that will be widely supported for a
>long time to come then I personally would go with the Verite chipset.

Compared to the Virge, yes. Now, I suggest you read up on Direct3D.
There is no hardware standard, Direct3D eliminates this. And if you
want to talk about games directly coded for the cards, then the Voodoo
currently has the best support.

Intelligent Gamer, July Issue:

"In the last year, one of the biggest questions on the minds of
developers and high-end PC gamers was which 3-D accelerator card would
be the one to emerge as the gaming standard. After viewing several
impressive games are chatting with game developers at E3, it appears
that 3DFX Interactive's Voodoo chipset is the front runner."

See also:

http://www.3dfx.com/press/960521.html

>BTW .. I really dont see what the huge deal is with tech support or
>customer service with a video board. I mean you basically get it up ans
>running and thats it besides getting the latest drivers from the net

Glad to see you've had no problems, other's haven't been so lucky.
Tech support and customer service _are_ the drivers. People hate
Diamond because they provided useless drivers for so long that didn't
take advantage of their cards under 95 or were extremely flaky under
3.1, were completely unresponsive to the users needs and lord help you
if you have a problem and want to get through on the phone.

All that being said, picking _any_ 3D accelerator at this point as the
"one you're going to buy" is foolish, IMO. Wait at least until you
see the software and hardware prices for godsakes!


Dave Glue

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

On Sun, 30 Jun 96 05:31:19 GMT, dar...@onramp.net (Daron Myrick)
wrote:


>I like the Verite chipset by Rendition also.
>Why?
>3DFX is made by a company I've never heard of till a couple of months ago that
>seems to not have any history, while Rendition has been around from the good
>old 286 days and has been making top of the line, extremely bleeding edge
>cards since the 286 came out.
>Rendition has been doing 3D hardware acceleration since the early 1980's so
>they know their stuff to where they can do it in their sleep.

>Which is faster, I don't know but I would feel more comfortable with Rendition
>since I know they are going to be there for us, with often
>software/driver/firmware updates.

Probably the same mentality which keeps people buying Microsoft and
Diamond. Not always a very wise one, but I understand your concern.
However, I suggest listening to the people who are actually working
with the product and more importantly, seeing the product yourself
before making any decisions.

As well, the 3DFX staff has been _more_ than helpful with any
questions I have asked. Another user also commented on just how
personal they have been, I have been very impressed with their
professionalism and eagerness to talk to the average Joe, which is not
as common as it should be these days.

Frank Lachance

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

> Anyhow, I'll take my comments back since both companies are new.
> I like the technology behind the Verite but the Voodoo is generaly
faster.

Tell me.. If you have a P166 and +, why do you absoluteely need more
speed? It's like if you buy a slower gfx card, the P166 will still be
fluid... it's just that everybody is dazzled by the Voodoo.. Ok, it's
powerful but not flexible..

ciao.!

Dave Glue

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

On Sun, 30 Jun 96 18:45:33 GMT, dar...@onramp.net (Daron Myrick)
wrote:


>Hopefully the card will be in the $500 to $600 range so that I can afford it
>though it's still quite steep.

I doubt the card will be that expensive if they hope to sell it in the
consumer market, which it is targetted for. Developers wouldn't be
supporting it if they felt it had no market. I would expect a price
of around $300 for a 2 meg frame buffer/2 meg texture model.


Jon

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

On Sun, 30 Jun 1996 18:51:11 -0700, Chris Ciccarello
<chri...@gnn.com> wrote:

>Brad wrote:
>>
>>
>> The only problem is that the Verite chipset by Rendition is supported

>> directly by "Microsoft" and all the major CAD and PC game developers. I


>> was lookin at the Virge chipset till I talked to many people and read up
>> on the Verite chipset. If your at all interested in games and interested
>> in getting the "standard" chipset that will be widely supported for a
>> long time to come then I personally would go with the Verite chipset.
>

>Yes, the Rendition card is the "reference" platform for Direct 3D, although I'm not sure that means much. It
>doesn't mean that other cards can't support Direct 3D as well or better. In fact from what I've read lately
>(CGW, PC Gamer), 3DFX is also getting a heck of a lot of support (as much as Rendition) and developers are
>excited about it too. Hopefully Direct 3D will catch on and the best performing card will be the best card to
>own, and right now that card looks to be 3DFX.

One problem with the Verite chip is it's a low-end low-cost solution.
Creative Labs even speaks of it as their first low-cost solution PCI
based 3D card. They also say they will have high-end more expensive
cards out later.

That's why I'm not going to buy the Verite based board, because it's
low-end, I want speed!

Life is grand, don't mess it up..
Existence is meaningless, buck up!
----------
Email : ampl...@tamu.edu
PGP-Key : email TO:pgp-pub...@pgp.mit.edu SUBJECT: get ampl...@tamu.edu
Web Page: http://blackhole.dorms.tamu.edu
Ftp Site: ftp://blackhole.dorms.tamu.edu

Peter Calderwood

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

On Mon, 01 Jul 1996 08:09:21 GMT, dav...@interlog.com (Dave Glue)
wrote:

>On Sun, 30 Jun 96 05:31:19 GMT, dar...@onramp.net (Daron Myrick)
>wrote:
>
>

Another reason to go with Verite is because 3D shutter glasses like
SimulEyes currently are working on support for it, not VOODOO, mainly
because of the power of Creative Labs.

--
Peter Calderwood
pet...@bc.cybernex.net
http://www2.cybernex.net/~peterc

Eric Calcagni

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

Ever try running Magic Carpet in Hi-Res on a P166?

Ever try to run US Navy Fighters in hi-res on a P-166?

Fact is, games are passing us by in terms of technology. Software has
surpassed what hardware can do efficiently. The above two games
are playable at hi-res, they are not really good performers.

You are correct when you say the Voodoo is not flexible. But, to be
honest, I really don't care that much. Chances are that when I actually
outgrow a Voodoo, the 3D market will have matured to a point that I will
be able to get a kick-butt 3D card that has yet to be seen.

Of course, this is speculation, since I haven't really seen the Voodoo.
But if it increases rendering speeds 10 times (just hoping) faster than my P100
can do by itself, I consider the $300 well spent. There is NO other
upgrade that will give you such a cost/performace increase.

BTW, on another note, I received mail from a 3DFX person and I would
like to share my question and his answer to you guys. This is
an example of how helpful the 3DFX people have been to me...

>2. I have a P100... I was wondering if your board could pull 30fps at a 800x600.

Your cpu performance mainly affects triangle rate, not fill rate. Our fill rate
is the same whether you have a P66 or a P200. 800x600 is about .5 Mpixels,
so you need 15 Mpixel/sec fill rate at a minimum to hit 800x600@30 fps. We
can easily do that. Now, if your game draws the same pixel multiple times
(aka overdraw) then that eats into your fill rate. If you use a lot of small
triangles, then that uses up a lot of cpu time and maybe you can't push
triangle data out to the graphics board fast enough.

Not a bad answer to a simple question eh? If he is correct with his answer,
then I know of no other board at the Voodoo's price range can pull that off.
Of course, these are still words until I actually see it working, but
it does sound promising.

Oh, I believe the 800x600 at 30fps was with 16 bit color, not 8 bit color.
The Voodoo doesn't even support 8 bit color if I remember right.

Of course, achieving such a good from rate at such a good resolution
will depend on how well the game programmer can efficiently uses his polygons.

- Eric C.

Daron Myrick

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

|>Another reason to go with Verite is because 3D shutter glasses like
|>SimulEyes currently are working on support for it, not VOODOO, mainly
|>because of the power of Creative Labs.

I don't get it....
Why would anyone intentionaly purchase a 3D video card that's slower?
Voodoo seems to be the fastest one can get under $500 at this time, so it
would make sense to get the Voodoo over anything else as long as they have
fast/bugless drivers and fast/bugless hardware.
I don't understand, why anyone would intentionaly go with a card that is half
the speed?

Daron Myrick

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

In article <4r941u$q...@news.interlog.com>, dav...@interlog.com (Dave Glue) wrote:
|>I doubt the card will be that expensive if they hope to sell it in the
|>consumer market, which it is targetted for. Developers wouldn't be
|>supporting it if they felt it had no market. I would expect a price
|>of around $300 for a 2 meg frame buffer/2 meg texture model.

I'm talking about buying the Obsidian from 3DFX and avoiding the Orchid card.
The Obsidian has to be at least $500 to start, right?
I don't want some dinky card made by a company that is one step up from
Diamond, I want a solid/fast card with full solid/fast drivers.
I don't want to feel at the Mercy of anyone, and that's what one will probably
get through normal channels.
On the other hand, the Obsidian has to have full Win95 support before I would
buy it or it's unfortunately Orchid or another chipset alltogether.

Daron Myrick

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

|> If you use a lot of small triangles, then that uses up a lot of cpu time and maybe you can't push
|> triangle data out to the graphics board fast enough.

The above is from 3DFX right?
If so then I'm concerned since that means that the Voodoo doesn't do bus
mastering since a bus mastering chip could get drawing instructions from main
memory at the same speed, no matter what speed one is using as long as the bus
speed stays the same.
A bus mastering card will let the CPU process instructions while the bus
mastering card itself grabs drawing/vector data from main memory.
According to the above 3DFX statement, the Voodoo does not do bus mastering,
which is a thorn in it's side.

Charles Kendrick

unread,
Jul 1, 1996, 3:00:00 AM7/1/96
to

Daron Myrick wrote:
>
> |> If you use a lot of small triangles, then that uses up
> |> a lot of cpu time and maybe you can't push
> |> triangle data out to the graphics board fast enough.
>
> The above is from 3DFX right?
> If so then I'm concerned since that means that the Voodoo doesn't do bus
> mastering since a bus mastering chip could get drawing instructions from main
> memory at the same speed, no matter what speed one is using as long as the bus
> speed stays the same.

The note about the importance of CPU speed is referring to computation speed
in rendering setup and the software geometry engine, not the rate at which the
CPU can push geometry data (once it has been determined) across the bus.

cken...@shgc.stanford.edu

Peter Calderwood

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

Another reason to go with Verite is because 3D shutter glasses like
SimulEyes currently are working on support for it, not VOODOO, mainly
because of the power of Creative Labs.

--

Dave Glue

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

On Mon, 1 Jul 1996 12:31:01 -0400, "Frank Lachance"
<fra...@accent.net> wrote:

>> Anyhow, I'll take my comments back since both companies are new.
>> I like the technology behind the Verite but the Voodoo is generaly
>faster.
>
>Tell me.. If you have a P166 and +, why do you absoluteely need more
>speed? It's like if you buy a slower gfx card, the P166 will still be
>fluid... it's just that everybody is dazzled by the Voodoo.. Ok, it's
>powerful but not flexible..

Really. Try running games with 16-bit colour and filtering on your
P166. Cripes, it can't even handle all games in 640*480 in 8-bit
colour with no effects at a smooth frame rate- I doubt EF2000 or Quake
is hitting 30fps in that mode.

And a graphics card still has a tremendous impact. Why cripple your
P166 with a slow graphics card. "Flexible"? Explain that- what
graphics card is "flexible"? Something I can tweak on the Verite to
make it push twice the amount of polygons/second?


Eric Calcagni

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

Daron Myrick wrote:
>
> |> If you use a lot of small triangles, then that uses up a lot of cpu time
> |> and maybe you can't push triangle data out to the graphics board fast enough.
>
> The above is from 3DFX right?
> If so then I'm concerned since that means that the Voodoo doesn't do bus
> mastering since a bus mastering chip could get drawing instructions from main
> memory at the same speed, no matter what speed one is using as long as the bus
> speed stays the same.
> A bus mastering card will let the CPU process instructions while the bus
> mastering card itself grabs drawing/vector data from main memory.
> According to the above 3DFX statement, the Voodoo does not do bus mastering,
> which is a thorn in it's side.

I can't say if the Voodoo uses bus mastering or not. But if I read him
correctly (and 3DFX's WWW page), this can be an advantage to some. The
speed at which your computer is able to draw graphics on your CRT is
dependent on both your main CPU and the Voodoo chipset. So, if you
have one P100 and another P166 (both with Voodoo), the P166 will be able to
get better frames per second over the P100 simply because the P166 is
able to throw out more polygons to the Voodoo more often.

So, in essence, if you ever upgrade your CPU to a faster speed, you
should see a performance gain with your Voodoo as well. You're not
"stuck" at the same speed no matter what system you have.

Another piece of information. Supposably there is going to be a big
announcement of a major mass-market graphics board manufacturer
for the Voodoo at any moment. This is good news if you don't like
Orchid or Fujitsu.

One last thing. For those that have been wondering whether to get a
Voodoo based Righteous 3D or a Voodoo based Obsidian system, another 3DFX person
told me...

"Obsidian isn't much more than a Voodoo and memory.
It has lots of expansion capabilities, but the performance of a
Righteous 3D and a base Obsidian should be the same."

So there you go.

- Eric C.

Dave Glue

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

On Mon, 01 Jul 96 15:59:38 GMT, dar...@onramp.net (Daron Myrick)
wrote:

>I don't get it....


>Why would anyone intentionaly purchase a 3D video card that's slower?
>Voodoo seems to be the fastest one can get under $500 at this time, so it
>would make sense to get the Voodoo over anything else as long as they have
>fast/bugless drivers and fast/bugless hardware.
>I don't understand, why anyone would intentionaly go with a card that is half
>the speed?

Price, feature list (the Verite supports DOS/Windows/Video
acceleration as well). I bought a Tseng4000 W32 Windows accelator a
while back- sure, there were faster Windows cards, but for the price,
it was "good enough" and I was quite happy with it. And we don't know
how much faster the Voodoo is over the Verite at this point, let's see
the products being put through their paces first.

Dave Glue

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

On Mon, 01 Jul 96 16:12:37 GMT, dar...@onramp.net (Daron Myrick)
wrote:

>The above is from 3DFX right?


>If so then I'm concerned since that means that the Voodoo doesn't do bus
>mastering since a bus mastering chip could get drawing instructions from main
>memory at the same speed, no matter what speed one is using as long as the bus
>speed stays the same.
>A bus mastering card will let the CPU process instructions while the bus
>mastering card itself grabs drawing/vector data from main memory.
>According to the above 3DFX statement, the Voodoo does not do bus mastering,
>which is a thorn in it's side.

That's not what he's saying. The vast majority of these cards use the
CPU to handle the geometry, regardless of how fast your bus
interaction is, you're still going to need a fast Pentium in order to
deliver enough triangle computations to satisfy these accelerators.

From what developers have said and comments from lookers-on at 3DFX
demo's, it certainly doesn't seem to be a "thorn" in its side if it
does indeed not perform bus mastering. I don't care if a hampster is
running around in a wheel on the board, what matters is the end
result.

EB

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

dar...@onramp.net (Daron Myrick) wrote:
>Hopefully the card will be in the $500 to $600 range so that I can afford it
>though it's still quite steep.

Sam here (below) says Orchid's Voodoo-based card should be about $300.
(ie. Righteous 3D)

-EB

P.S. I'd like to know which stores will be authorized to demo the
Righteous 3D when it comes out. Anyone know (Canadian stores)?

------- Forwarded Message Follows -------
From: pa...@3dfx.com
Date: Wed, 26 Jun 96 17:36:49 PDT
To: Kry...@lightspeed.bc.ca
Subject: Re: 3D cards rated by speed, & MW2
Organization: 3Dfx Interactive

In article <31CC4F...@lightspeed.bc.ca> you write:
>It looks like I'll be most interested in the 3DFX's Obsidian cards.
>(http://www.3Dfx.com/products/obsidian.html)
>OR VideoLogic's PowerVR
>(http://www.videologic.com/prodinfo/pvr.htm)

According to the flyer I got from VideoLogic at E3, the home
version of PowerVR has a maximum triangle throughput of about
1/4 what ours is (they support multiple chips to get higher
performance, which we do also). They also don't have bilinear
filtering. They do support lod blending (linear mipmapping) at
little or no performance cost, while we require either a second
texture chip or reduced performance. I think their fill rate
is lower too, but it is harder to compare given the radically
different architectures.

>I'm not sure of their prices. Can anyone list them yet?

Orchid's Voodoo based board should be around $300. It is
possible it could be lower, but it also may be higher
(the final price will be largely determined by what software
gets bundled).

I expect PowerVR to be signficantly cheaper.

Sam Paik
--
Samuel S. Paik / pa...@3dfx.com / 415-919-2441 / I speak only for myself

Brian Hook

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

> If so then I'm concerned since that means that the Voodoo doesn't do bus
> mastering since a bus mastering chip could get drawing instructions from main
> memory at the same speed, no matter what speed one is using as long as the bus
> speed stays the same.

The Voodoo is not bus mastering. The Voodoo has the highest performance of any
current or announced 3D graphics accelerator, so I ask you: how is lack of
bus mastering a performance bottleneck?

And I believe your technical assertions aren't entirely accurate. Bus mastering, as
far as I know, operates at bus speed, but CPU speed determines how fast you can
perform triangle setup (gradient calculations). Whether you are using PIO or bus
mastering, assuming that bus speed is constant you're still going to push the same
amount of data across the PCI bus.

Just because something happens "in the background" doesn't mean it's free.

Brian

Steve

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

On Tue, 02 Jul 1996 18:32:01 GMT, dav...@interlog.com (Dave Glue)
wrote:

>Price, feature list (the Verite supports DOS/Windows/Video


>acceleration as well). I bought a Tseng4000 W32 Windows accelator a
>while back- sure, there were faster Windows cards, but for the price,
>it was "good enough" and I was quite happy with it. And we don't know
>how much faster the Voodoo is over the Verite at this point, let's see
>the products being put through their paces first.

And not forgetting that iD Software are very excited about Rendition's
Verite, and they've been running a pre-production version.

I like many others are more persuaded by this kind of enthusiasm than
by obscure benchmarks that differ from chip to chip.

Cheers

Steve

Steve

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

On Tue, 02 Jul 1996 12:25:00 -0700, Brian Hook <b...@netcom.com> wrote:

-snipped-


>And I believe your technical assertions aren't entirely accurate. Bus mastering, as
>far as I know, operates at bus speed, but CPU speed determines how fast you can
>perform triangle setup (gradient calculations). Whether you are using PIO or bus
>mastering, assuming that bus speed is constant you're still going to push the same
>amount of data across the PCI bus.

-snipped-

AHEM, PIO is an enhanced rate of flow for Enhanced IDE disk drives ;-)

Steve

Jean Gagnon

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to Dave Glue

> Glad to see you've had no problems, other's haven't been so lucky.
> Tech support and customer service _are_ the drivers. People hate
> Diamond because they provided useless drivers for so long that didn't
> take advantage of their cards under 95 or were extremely flaky under
> 3.1, were completely unresponsive to the users needs and lord help you
> if you have a problem and want to get through on the phone.>>>


Another company having a great support is ATI. You allways reach tech support at first
ring (at least for me). You didn't mention this company so do you feel their 3D card is not match
against the competition?

Brian Hook

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

*AHEM*, read a computer architecture text book when you get a chance and come back and remind me
what PIO means (in the context of my post).

Brian Stretch

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

With the turtle icon enabled on Quake 0.91 (displayed when the framerate dips
below 10fps), I see it an awful lot running 640x400 on my Pentium 133 burst-cache
board and Matrox Millenium video card. Still my favorite res for the game, but
it's reinstated a PPro motherboard to my toy list. I'll buy a Voodoo card too so
long as enough games I like are supported.

Wimpy video cards start choking real quick at 640x480 res (single-ported memory
cards first and foremost, such as DRAM-based cards). If it doesn't use dual-
ported memory (VRAM, WRAM, etc), I don't want it.
--
Brian Stretch (str...@arborweb.com) Software Engineer, Certified Technojunkie
Visit Automotive Reviews Online by Steven D. Gatt, http://autopage.arborweb.com
ArborWeb Inc., Ann Arbor Michigan's Full-Service Web Provider and Consultancy.
Visit our home page at http://www.arborweb.com, or email in...@arborweb.com

Daron Myrick

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

In article <31D9B6...@netcom.com>, b...@netcom.com wrote:
|>*AHEM*, read a computer architecture text book when you get a chance and come
|> back and remind me
|>what PIO means (in the context of my post).

PIO as in Peripheral Input/Output device, or a simple chip that takes the
place of several logic chips, giving I/O capability including handshaking and
port to bit level priority processing.
I remember PIO chips very well back from the early 1980's.
Actualy there were 8085 and 8080 PIO chips along with Z-80 PIO chips too, long
before the 8086 and 8088 and the PC in general.
Ahhh, brings back mammories of the good old days :-) .
Have no Idea what PIO stands for today, though I know what it does and how it
works.

Daron Myrick

unread,
Jul 2, 1996, 3:00:00 AM7/2/96
to

|>>Tell me.. If you have a P166 and +, why do you absoluteely need more
|>>speed? It's like if you buy a slower gfx card, the P166 will still be
|>>fluid... it's just that everybody is dazzled by the Voodoo.. Ok, it's
|>>powerful but not flexible..

Depends on how complex the software/game is.
Sure, it's easy to push around a few thousand light shaded polygons, one can
do that with just a Pentium 133.
But the more complex the game is, and the more polygons used, the slower
everything gets.
So, you might be able to calculate even millions of those 3D transforms with a
Pentium, more than fast enough, but the pentium shure wouldn't be able to
calculate AND draw fast enough, so that is why we need 2D and 3D graphics
accelerators.
Now, taking it one step further, what if a software/game becomes so complex
that even a slower graphics accelerator can't pump out enough polygons a
second to keep up with a liquid frame rate?
Common low end arcade 3D games these days need to pump out around 150,000 to
350,000 texture mapped light shaded, transparent, overlayed and all kinds of
other gibberish, polygons per second.
If one has a card that can only pump around 90,000 polygons per second (Like
the Millenium, which can't even do 3D polygonal texture mapping) then one is
confronted with what one has been avoiding all along which we see when using
just a pentium alone, which is jerky, slow, graphics.
Also, I know exactly what is going to happen when the first slew of 3D
cards come out and they catch on.
What will happen is that software/games will become much more complex and then
we will be right back where we started, with slow, jerky, graphics within a
years time and then we will be doing this 3D card evaluation thing all over
again, and repeating this thread again, just like it was year after year
before, since the IBM monochrome days.

Dave Glue

unread,
Jul 3, 1996, 3:00:00 AM7/3/96
to

On Tue, 02 Jul 1996 20:52:20 GMT, 0@0.0 (Steve) wrote:

>On Tue, 02 Jul 1996 18:32:01 GMT, dav...@interlog.com (Dave Glue)
>wrote:
>
>>Price, feature list (the Verite supports DOS/Windows/Video
>>acceleration as well). I bought a Tseng4000 W32 Windows accelator a
>>while back- sure, there were faster Windows cards, but for the price,
>>it was "good enough" and I was quite happy with it. And we don't know
>>how much faster the Voodoo is over the Verite at this point, let's see
>>the products being put through their paces first.
>
>And not forgetting that iD Software are very excited about Rendition's
>Verite, and they've been running a pre-production version.

Yes, but keep in mind those quotes from iD are when the only other
accelerators they could work with were the VL-BUS 3D Blaster and the
Nvidia. They chose it as the accelerator for Quake because it was the
only decent-performing solution available at that point.


EB

unread,
Jul 3, 1996, 3:00:00 AM7/3/96
to

Well... now you will of course have to elaborate & back up your statement that
the voodoo chip is "not flexible" with accurate facts, ci vous plais. :)
I'm very interested in hearing this, as I too am very impressed by this 3Dfx
chip & have never heard of its 'inflexibility' as you put it, so PLEASE back
up your statements (& list all valid references if were possible).
THANKS!

-EB

"Frank Lachance" <fra...@accent.net> wrote:

>Tell me.. If you have a P166 and +, why do you absoluteely need more
>speed? It's like if you buy a slower gfx card, the P166 will still be
>fluid... it's just that everybody is dazzled by the Voodoo.. Ok, it's
>powerful but not flexible..
>

>ciao.!

Frank Lachance

unread,
Jul 3, 1996, 3:00:00 AM7/3/96
to


> Kry...@lightspeed.bc.ca (EB) wrote in article
<31d9d495...@news.teleserve.ca>...


> Well... now you will of course have to elaborate & back up your
statement that
> the voodoo chip is "not flexible" with accurate facts, ci vous plais.
:)
> I'm very interested in hearing this, as I too am very impressed by this
3Dfx
> chip & have never heard of its 'inflexibility' as you put it, so PLEASE
back
> up your statements (& list all valid references if were possible).
> THANKS!

The Voodoo has a SOLID setup of polygons. Verite, due to his RISC core, is
programmable so that the setup can be expanded if a new method/setup is
coming on the market... And RISC core MEANS SPEED AND POWER for those
kids who want speed... here, you have power AND speed...


Salut M. Si vous plait...

Brian Hook

unread,
Jul 3, 1996, 3:00:00 AM7/3/96
to

> The Voodoo has a SOLID setup of polygons. Verite, due to his RISC core, is
> programmable so that the setup can be expanded if a new method/setup is
> coming on the market... And RISC core MEANS SPEED AND POWER for those
> kids who want speed... here, you have power AND speed...

Out of curiousity, why does a RISC core mean speed and power, given that ASICs in general are
significantly faster by design (you don't need to interpret instructions).

Also, please check the performance of the Verite and the Voodoo Graphics and tell me, once
again, how the RISC score is faster and more powerful.

Brian

Daron Myrick

unread,
Jul 3, 1996, 3:00:00 AM7/3/96
to

|>The Voodoo has a SOLID setup of polygons. Verite, due to his RISC core, is
|>programmable so that the setup can be expanded if a new method/setup is
|>coming on the market... And RISC core MEANS SPEED AND POWER for those
|>kids who want speed... here, you have power AND speed...

Yes, you are right, the Verite is code expandable, but at the cost of being 4
times slower than the Voodoo already, and Verite based cards will only get
slower as extra graphics features are added and image complexity increases.
I would much more prefer a core which is 4 times faster already, with
expansions done by the system CPU, than have everything done by a single RISC
which has already been proven 4 times faster without enhancements since once
enhancements are added, the Voodoo/CPU combination would at least be 4 times
faster still using a P166 or P200.

Frank Lachance

unread,
Jul 3, 1996, 3:00:00 AM7/3/96
to

>
> One problem with the Verite chip is it's a low-end low-cost solution.
> Creative Labs even speaks of it as their first low-cost solution PCI
> based 3D card. They also say they will have high-end more expensive
> cards out later.
>


Verite chipset IS NOT A LOW-END LOW-COST SOLUTION... Verite is a powerful
chipset that can match the Voodoo but with a slower processing of
poly/s...

> That's why I'm not going to buy the Verite based board, because it's
> low-end, I want speed!


BTW, The Verite can do over 180,000 poly/s fully rendered polygones.. the
Voodoo, 655,000 poly/s .. HMMMMM... SATISTIC?? CANNOT COMPARED THOSE...


Frank Lachance

unread,
Jul 3, 1996, 3:00:00 AM7/3/96
to

You are Mentioning it.: > Common low end arcade 3D games these days need


to pump out around 150,000 to
> 350,000 texture mapped light shaded, transparent, overlayed and all
kinds of
> other gibberish, polygons per second.

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^

> If one has a card that can only pump around 90,000 polygons per second
(Like
> the Millenium, which can't even do 3D polygonal texture mapping) then
one is
> confronted with what one has been avoiding all along which we see when
using
> just a pentium alone, which is jerky, slow, graphics.

Here you say 90,000 poly/s, but are they FULLY RENDERED or PLAIN poly ?
AAAAAH... Here is the cross.. Some company will say that they can do
700,000 poly/s, but I bet a brown (it's 100$ in canada) that it's PLAIN
poly.. Do you really think that a card can handle 700,000 poly/sec FULLY
rendered? Why then SGfx computers exist?????? hmmm... it would be a good
stole from 3D chipset makers VS SGfx..


Daron Myrick

unread,
Jul 3, 1996, 3:00:00 AM7/3/96
to

|>Verite chipset IS NOT A LOW-END LOW-COST SOLUTION... Verite is a powerful
|>chipset that can match the Voodoo but with a slower processing of
|>poly/s...

Fast poly's are %95 of what a 3D card is needed for in the first place.
So, what you are unknowingly stating above, is that you know that the Verite
is slower than the Voodoo but your going with the Verite since you've made up
your mind.
Ok, go for it, but I'm going with the Voodoo which is already at least 4 times
faster than the Verite and will be even more faster once extra RISC code is
used for "add on graphics enhancements" which the Voodoo and the system CPU
together, can do, still 4 times faster than the Verite doing everything
itself.
I was sold on the Verite also until I actualy studied www.3dfx.com and learned
the facts, then I realized that this RISC stuff "just aint gonna cut it"
compared to 3D litteraly DONE IN hardware, not just BY hardware like RISC
does.

Steven Begley

unread,
Jul 4, 1996, 3:00:00 AM7/4/96
to

>I remember PIO chips very well back from the early 1980's.

I dont want to remember them :-)

>Ahhh, brings back mammories of the good old days :-) .

Err... MAMMORIES?!?!?! Hahahah!!!!

>Have no Idea what PIO stands for today, though I know what it does and how it
>works.

Polled Input/Output.

Cya,

Steve


Gary McTaggart

unread,
Jul 4, 1996, 3:00:00 AM7/4/96
to

In article <4r7s6d$c...@news.onramp.net> dar...@onramp.net (Daron Myrick) writes:

> |>Theoretically: If both products are released at approximately the same
> |>time, yet the 3DFX delivers twice the performance of the Verite, are
> |>you going to go with Rendition because they've been around longer? I
> |>really doubt it.
>

> I correct myself by stating that the Rendition I was thinking of was not the
> Rendition company that makes the Verite.
> As far as which company is newer, Rendition is but only by less than a year.
> I'm sold on the Voodoo after visiting their site.
> Too bad we have to buy an Orchid card for Voodoo power since I always hated
> Orchid.
> I'm tempted to buy the Voodoo card make by 3DFX themselves since at least I
> know I won't be getting support and drivers from a company that's one step up
> from Diamond.


> Hopefully the card will be in the $500 to $600 range so that I can afford it
> though it's still quite steep.

The $300 card from Orchid is every bit as powerful as the 3Dfx
Obsidian board in my machine. You will get the same 3Dfx developed
and tuned drivers with the Orchid board that we use with the
Obsidian. The Obsidian is meant to be more expandable for
arcade/location based entertainment (LBE) use and also a reference
board for OEMs.

Gary McTaggart
3Dfx Interactive

--
~/.signature

Jon

unread,
Jul 4, 1996, 3:00:00 AM7/4/96
to

On Wed, 3 Jul 1996 16:49:58 -0400, "Frank Lachance"
<fra...@accent.net> wrote:

>>
>> One problem with the Verite chip is it's a low-end low-cost solution.
>> Creative Labs even speaks of it as their first low-cost solution PCI
>> based 3D card. They also say they will have high-end more expensive
>> cards out later.
>>
>
>

>Verite chipset IS NOT A LOW-END LOW-COST SOLUTION... Verite is a powerful
>chipset that can match the Voodoo but with a slower processing of
>poly/s...

The part you cut out was important. The fact that Creative Labs said
this themselves. If you want to argue, argue with CL.

>
>> That's why I'm not going to buy the Verite based board, because it's
>> low-end, I want speed!
>
>
>BTW, The Verite can do over 180,000 poly/s fully rendered polygones.. the
>Voodoo, 655,000 poly/s .. HMMMMM... SATISTIC?? CANNOT COMPARED THOSE...
>

I don't understand what you are trying to say.

Life is grand, don't mess it up..
Existence is meaningless, buck up!
----------
Email : ampl...@tamu.edu
PGP-Key : email TO:pgp-pub...@pgp.mit.edu SUBJECT: get ampl...@tamu.edu
Web Page: http://blackhole.dorms.tamu.edu
Ftp Site: ftp://blackhole.dorms.tamu.edu

Jon

unread,
Jul 4, 1996, 3:00:00 AM7/4/96
to

See:

http://www.hyperstand.com/SITE/NMToday/1107/3Dfx_Voodoo_Graphics.html

The voodoo can do 350,000 z-buffered, bilinear, LOD MIP-mapped,
alpha-blended 50-pixel triangles per second.

Although as the page mentions, this is less than the Permedia chip
slated for Creative Labs later high-end 3d boards.

Greg Pierce

unread,
Jul 4, 1996, 3:00:00 AM7/4/96
to

Its not, they even admitted that themselves at E3, claiming that price
was the primary concern of their first board.
--
------------------------------------------------------------
Gregory Pierce ---- Template Software, Inc.
pie...@template.com \ / 13100 Worldgate Drive
(703) 318 - 1000 \/ Suite 340
(703) 318 - 7378 fax Herndon, VA 22070-4382
http://www.template.com/
'The opinions above are mine alone...'

Gary McTaggart

unread,
Jul 4, 1996, 3:00:00 AM7/4/96
to Brad

In article <31D6B8...@icsi.net> Brad <br...@icsi.net> writes:

> Eric Calcagni wrote:
> >
> > Ok. I've been researching heavily into 3D cards in the past several
> > months. I have seen all the specs, all the gee-wiz items and
> > useless polygon counts. After all of this, I have decided I am going
> > to get the Voodoo chipset. They seem to have a winning chipset
> > on their hands (I hope :) But I just wanted to add...
> >
> > 3DFX has won my over by their wonderful technical support online.
> > Every e-mail I have sent, they have answered with a personal answer.
> > EVERY one. And not once have I gotten the hard sell from these people.
> > If anything, they have been very modest and told me what they think
> > the advantages are of some other 3D cards as well. Finding a company
> > that will do that is almost impossible.
> >
> > In fact, I have gotten better support from 3DFX than from any company
> > I have ever purchased anything from - and I don't own any 3DFX stuff
> > yet.
> >
> > I know I'll eventually have to deal with technical support from
> > the companies that actually make the video cards that support
> > have the voodoo on board, and I'm not expecting much from the current
> > offerings. But I am so impressed by 3DFX's complete and sound answers
> > to my questions that they have my business.
> >
> <snip.........>
>
>
> The only problem is that the Verite chipset by Rendition is supported
> directly by "Microsoft" and all the major CAD and PC game developers. I
> was lookin at the Virge chipset till I talked to many people and read up
> on the Verite chipset. If your at all interested in games and interested
> in getting the "standard" chipset that will be widely supported for a
> long time to come then I personally would go with the Verite chipset.
>
> BTW .. I really dont see what the huge deal is with tech support or
> customer service with a video board. I mean you basically get it up ans
> running and thats it besides getting the latest drivers from the net
> --
> ("`-''-/").___..--''"`-._
> `6_ 6 ) `-. ( ).`-.__.`)
> (_G_.)' ._ ) `._ `. ``-..-'
> _..`--'_..-_/ /--'_.' ,'
> (il).-'' (li).' ((!.-'
> br...@isci.net
> http://206.26.158.13/free/blj19.html

I recommend that you contact Microsoft to see which graphics chips
they support directly versus which ones they support indirectly. I
think you'll find that the semantic of "directly" versus "indirectly"
is superfluous.

3Dfx has worked closely with both D3D teams in the UK and Redmond.
The combination of 3Dfx's internally developed base drivers and tuning
by the D3D teams has given very impressive results. For example, at
the WinHEC conference in March, Bill Gates demonstrated Direct3D on a
consumer level Voodoo Graphics equipped board as part of his Simply
Interactive PC initiative. The application Bill G. ran was Valley of
Ra under D3D and it was on a $300 MSRP board-- not one of the $500 or
$1,000 coin-op or LBE boards. And, while our team at 3Dfx did the base
level driver and alot of the application code for the demo, it was the
Microsoft D3D folks that really enabled us to COLLECTIVELY pull the
demo off. Since that demo, we've worked with Microsoft to really show
what is possible with Direct3D-- as evidenced by the D3D games demo'd
at E3 on Voodoo Graphics. Included among the demos were Microsoft's
own Monster Truck Madness and Hell Bender, as well as other D3D games
by Viacom (Divide), Ocean/GTE, Psygnosis, Virgin, and others. We also
accompanied Microsoft on their recent WinHEC Asia tour where our VP of
Sales and Marketing demonstrated advanced features of D3D including
mip mapping, filtering, Z-buffering, alpha blending, etc. and
performance beyond anything else in the market (1M tris/sec on a
P5-166). In fact, 3Dfx DEMOS were used to illustrate the CONCEPTS
talked about by Rob Mullis of Rendition (on what is good 3D?

In terms of CAD support for Rendition versus 3Dfx or others-- you
should first understand that most, if not all, CAD programs will
support D3D and/or OpenGL. If they don't now, they will in the future
(more than likely). Certainly, companies like SoftImage, Caligari,
Fractal Designs, Sense8, Datapath, PTC, SDRC, etc., etc., and many
others are already "on it!" And, once again, by working with
Microsoft (the creators of OpenGL and D3D for Win32 and Windows95),
ALL graphics vendors "get" these applications. In most cases, it
really comes down to performance. For texture mapped applications,
Voodoo Graphics DELIVERS the kind of polygonal and fill rate
performance that no other PC graphics chip can even approximate. The
closest competitor is probably Lockheed's Real3D/100 series, which
costs over 2x Voodoo Graphics and has yet to sample. Based on
Lockheed's published specs (versus our actual tests), Voodoo Graphics'
consumer configuration-- like you will be able to buy from Orchid this
summer for $300, outperforms even the Real3D/100. The net of this is
that your games will run faster and your architectural walkthroughs
will be at real time frame rates.

If you are really interested in key endorsements of the technology--
check our our web page (http://www.3dfx.com) and read the SAIC press
release about their selection of Voodoo Graphics for their government
visual simulation applications. Here's a company that previously
purchased Reality Engines and Lockheed Image Generators now using
Voodoo Graphics on PCs! No other PC graphics chip maker can make this
same claim.

Lastly, in terms of tech support. Companies like Dell, Compaq, and
IBM might differ with you on the importance of tech support. Each of
these companies sell essentially the same product that you can buy at
a local screwdriver shop-- but the technical support they offer
customers is typically much better. For alot of consumers and
businesses alike, tech support, whether it's on their computer, their
graphics card, or their car, is very important. While 3Dfx may have
the best technology, it's not enough-- we strive to provide the best
tech support as well.

Gary McTaggart

unread,
Jul 4, 1996, 3:00:00 AM7/4/96
to

In article <01bb676b.57eafc40$5974...@server.accent.net> "Frank Lachance" <fra...@accent.net> writes:

>> Anyhow, I'll take my comments back since both companies are new.
>> I like the technology behind the Verite but the Voodoo is generaly
> faster.


>
>Tell me.. If you have a P166 and +, why do you absoluteely need more
>speed? It's like if you buy a slower gfx card, the P166 will still be
>fluid... it's just that everybody is dazzled by the Voodoo.. Ok, it's
>powerful but not flexible..

It is a misconception that the Voodoo is not flexible. The Voodoo is
designed with flexibility *and* performance in mind. You would not
believe some of the cool stuff that you can do with this chipset
*without* having to program to some weird paradigm or write
microcode. I used to program on an SGI Onyx/Reality Engine 2 machine
heavily (in fact, the "Valley of Ra" was originally developed there),
and I do not feel held back at all by Voodoo Graphics.

I'd like to hear something more concrete about what makes the Voodoo
inflexible. The game developers are findint it to be very flexible,
in fact they are amazed at how easy it is to make their games run on
Voodoo when they previously thought that their games would not run on
an "inflexible" 3D chipset. Take a look at a screenshot of "Prey" by
3D Realms at http://www.3dfx.com/game_dev/game_dev.html or at
www.3drealms.com . This game does some really cool lighting.

Here's a quote from the 3D Realms web page:

". . . We're supporting the 3DFX card, and this is what it'll look
like. It is so cool, we're having a hard time looking at the normal game
now. Check out the card at E3 (the Electronic Entertainment Expo) in
LA. You can actually walk around in Prey, and in hi-res it's a dream!
(In a nutshell, we think the 3DFX card rocks.)"

Gary McTaggart
--
~/.signature

Daron Myrick

unread,
Jul 4, 1996, 3:00:00 AM7/4/96
to

|>The $300 card from Orchid is every bit as powerful as the 3Dfx
|>Obsidian board in my machine. You will get the same 3Dfx developed
|>and tuned drivers with the Orchid board that we use with the
|>Obsidian. The Obsidian is meant to be more expandable for
|>arcade/location based entertainment (LBE) use and also a reference
|>board for OEMs.

Yes, but I want all the expandibility and dual processing capability that the
Obsidian board gives.
Why can't I buy an obsidian board?
I keep asking for the price and I keep getting referred back to the Orchid
card which isn't expandible and doesn't support two Voodoo chips.
I've always gone with the best and I'm sold on the Obsidian as along as it's
as stable and has fast Direct3D drivers for Win95.

Dave Glue

unread,
Jul 5, 1996, 3:00:00 AM7/5/96
to

On Wed, 3 Jul 1996 16:49:58 -0400, "Frank Lachance"
<fra...@accent.net> wrote:

>
>Verite chipset IS NOT A LOW-END LOW-COST SOLUTION... Verite is a powerful
>chipset that can match the Voodoo but with a slower processing of
>poly/s...

By that logic, a 3D Blaster can match the Verite- just with a "slower
processing of polys"- speed is not a minor "feature" of 3D cards, it's
not a "just" situtation. Speed is _everything_ with these devices!

And the Voodoo supports more effects than the Verite. This is not
"slamming" the Verite, it looks to be a great card as well, but the
early indications are that the Voodoo has the best performance- you'll
pay for it though, as well as requiring a 2D accelerator.

>
>BTW, The Verite can do over 180,000 poly/s fully rendered polygones.. the
>Voodoo, 655,000 poly/s .. HMMMMM... SATISTIC?? CANNOT COMPARED THOSE...

You are not making an ounce of sense at this point.

Dave Glue

unread,
Jul 5, 1996, 3:00:00 AM7/5/96
to

On Wed, 03 Jul 1996 19:02:11 -0700, Brian Hook <b...@netcom.com> wrote:

>> The Voodoo has a SOLID setup of polygons. Verite, due to his RISC core, is
>> programmable so that the setup can be expanded if a new method/setup is
>> coming on the market... And RISC core MEANS SPEED AND POWER for those
>> kids who want speed... here, you have power AND speed...
>

>Out of curiousity, why does a RISC core mean speed and power, given that ASICs in general are
>significantly faster by design (you don't need to interpret instructions).

Because he read once that RISC is good, CISC bad. About as detailed
as most computer and videogame magazines get these days. :)


Dave Glue

unread,
Jul 5, 1996, 3:00:00 AM7/5/96
to

On Wed, 3 Jul 1996 16:54:53 -0400, "Frank Lachance"
<fra...@accent.net> wrote:


>The Voodoo has a SOLID setup of polygons. Verite, due to his RISC core, is
>programmable so that the setup can be expanded if a new method/setup is
>coming on the market... And RISC core MEANS SPEED AND POWER for those
>kids who want speed... here, you have power AND speed...

As much speed as the Voodoo? Not from what I've heard, but may be
"enough" to give a good performance boost along with image quality.

My understand of the RISC core in the Verite is that it can help with
some of the overhead of graphics API's and can perform some of the
set-up geometry (but not the actual floating point calcs, more in
terms of scene management). I don't know if it's possible to say, get
sub-pixel accuracy just by "programming" the RISC core. If that was
truly possible, and not have a huge impact on speed, don't you think
Rendition would be working on such a feature and list every possible
effect under the sun for their chipset?


Cuyler Buckwalter

unread,
Jul 5, 1996, 3:00:00 AM7/5/96
to

> As far as can see thus far, there is no other card on the PC
> market this year that will beat the 3Dfx/Orchid cards.

Does the Orchid Righteous card support both 3D and 2D
graphics? If so, what 2D chipset does it use (how will
it compare to other 2D accelerators on the market today)?

Or is the Orchid Righteous just a 3D solution that
requires a separate 2D accelerator? If so, will
it work with most modern 2D accelerators (such as
the Matrox Millenium or Hercules Dynamite 128)?

I will feel silly delaying the purchase of my video
card in order to buy a Righteous card only to find
out I still need to get a 2D video card to go with
it.

Orchid's web site doesn't say too much about its
Righteous card. It's really coming out this month?

Cuyler

Eric Calcagni

unread,
Jul 5, 1996, 3:00:00 AM7/5/96
to

EB wrote:

> 3Dfx's responce was this:
> "Over 1 Million.
> 350k was an estimated figure (on the datasheet/web), before we received chips
> back from the fab.
> Actual performance has been much higher than [even we] anticipated."


>
> As far as can see thus far, there is no other card on the PC market this year

> that will beat the 3Dfx/Orchid cards. I'm terribly anxious to see a live demo
> of it in action. I think I'll have to get one. I've got a neeeed 4 speeeed. :)
> (I sure hope a 3D version of MW2 will be made for the voodoo! :)
>
> -EB
>

I have no idea what software will be bundled with the Voodoo. I'd personally
say that some Egyptian fighting game has a good chance though. ;)

I asked 3DFX what titles would be shipping with the card, but they wouldn't
say anything for obvious reasons. However, I did get this response a while
ago...

(3DFX person:)

"I've been told the list of games being pushed for bundling, although at least one
of them has fallen off since it won't be ready. I've seen some of the ones closest
to completion (actually, two are complete). I don't remember most of the
others. I really can't tell you about them. Take a look at what games are getting
bundled with others, that will give you an indication of which companies and games."

Notice that last sentence. ;)

So, IMHO, I'd say that there might be a pretty good chance of Descent or MW2 being
ported to the Voodoo. If I remember right, one of the Descent programmers
said the Voodoo was an awesome card, so use that if you will.

MW2, I think, has a pretty good shot at it. They already make one for the
Rage, and a Virge version is in the works. I believe a Mystique version is
too (going by what I have read in various newsgroups.) If MW2 isn't
used as part of the bundle, I'd say there is a good shot of it being available
seperately.

- Eric C.

(PS: I'm really hoping id decides to go with the Voodoo. If not, I'll be
glad to wait for the direct 3D version of Quake. Quake NEEDS this card. ;P

Eric Calcagni

unread,
Jul 5, 1996, 3:00:00 AM7/5/96
to

Cuyler Buckwalter wrote:

> Does the Orchid Righteous card support both 3D and 2D
> graphics? If so, what 2D chipset does it use (how will
> it compare to other 2D accelerators on the market today)?
>
> Or is the Orchid Righteous just a 3D solution that
> requires a separate 2D accelerator? If so, will
> it work with most modern 2D accelerators (such as
> the Matrox Millenium or Hercules Dynamite 128)?
>
> I will feel silly delaying the purchase of my video
> card in order to buy a Righteous card only to find
> out I still need to get a 2D video card to go with
> it.
>
> Orchid's web site doesn't say too much about its
> Righteous card. It's really coming out this month?
>
> Cuyler

Yes. You will need a separate 2D card. Go get one now and don't
worry about the Voodoo quite yet.

Orchid's web site, to put it succinctly, sucks. They give NO
information on the Righteous 3D. EVERY other company that
has planned to put out a 3D card has given at least a little
bit of information on it. In fact, I'm beginning to wonder
if Orchid is going to update their website -after- they release
the Righteous 3D. This makes me worry just a tad. I almost
wonder if they even care if they produce a real product.

I will probably go with the Orchid card anyway. Hell, it's
only a video card, but I like the chips it has on it. :) I
looked at Fujitsu's web page, and they mention nothing about their
Voodoo based card either. What's the deal here???

As far as a release date, there are no firm ones. The earliest
date that has been mentioned is the end of this month. However,
my gut instinct tells me that the earliest will be at the end of
August. More than likely, you won't be able to buy one until then.

I hope I'm wrong though. Researching 3D cards for the past
8 months has worn me out. ;) I'm just ready to buy a damn card
and start playing!

- Eric C.

EB

unread,
Jul 6, 1996, 3:00:00 AM7/6/96
to

The 350k number was a preliminary estimated posted on a 'needs-to-be-updated"
Web page:
I wrote to 3Dfx asking about this:
http://www.3Dfx.com/products/obsidian.html says:
"An entry level Obsidian coupled with an Intel Pentium-90 or a PowerPC 603
delivers over 45 Megapixels/sec textured fill rate and 350k textured
triangles/sec. Obsidian provides the fill rate performance required for games
with high depth complexity to break the 30Hz barrier."

BUT!...
http://www.3Dfx.com/products/sys3d.html says:
"With the capability to deliver over one million texture mapped triangles per
second and up to ninety million filtered, Z-buffered, alpha-blended, LOD MIP
mapped, anti-aliased textured pixels per second, Obsidian provides the frame
rate, polygon intensity and depth complexity that your texture mapped 3D games
deserve."

3Dfx's responce was this:
"Over 1 Million.
350k was an estimated figure (on the datasheet/web), before we received chips
back from the fab.
Actual performance has been much higher than [even we] anticipated."

As far as can see thus far, there is no other card on the PC market this year
that will beat the 3Dfx/Orchid cards. I'm terribly anxious to see a live demo
of it in action. I think I'll have to get one. I've got a neeeed 4 speeeed. :)
(I sure hope a 3D version of MW2 will be made for the voodoo! :)

-EB

______________________________

Alex Fernandez

unread,
Jul 6, 1996, 3:00:00 AM7/6/96
to

On Sat, 06 Jul 1996 02:11:31 GMT, Kry...@lightspeed.bc.ca (EB) wrote:
>As far as can see thus far, there is no other card on the PC market this year
>that will beat the 3Dfx/Orchid cards. I'm terribly anxious to see a live demo
>of it in action. I think I'll have to get one. I've got a neeeed 4 speeeed. :)
>(I sure hope a 3D version of MW2 will be made for the voodoo! :)
>
> -EB
>
>______________________________
>ampl...@tamu.edu (Jon) wrote:
>

I'm kinda getting tired reading these posts about which chip is better than the
other. It all comes down to which games are written for which chips. I haven't
seen much in the press about game companies supporting the voodoo chip.
But I have seen press releases about companies supporting the Rendition chip in
their games. I play ICR2, NASCAR, and will play F1GP2. I know that Papyrus will
support the Verite. Because of that support, I will look seriously at the
products based on the Verite. Also, as I understand it, the voodoo does not
have a VESA 2.0 compliant VGA/SVGA core (correct me if i'm wrong) since it
requires a 2D board. The Verite does have a VESA 2.0 compliant VGA/SVGA core
that can sustain the full 133MB/s PCI thruput. The following is a quote from
one of their reps:

>Moreover, we believe we have an extremely fast SVGA / VESA engine. We are
>able to process data as fast as the PCI bus is able to give it to us. This
>can be up to the theoretical maximum of 133 MB/sec. Usually the core logic
>cannot sustain more than 100 MB/sec.

It all comes down to software support for the 3D features in DOS. But in Win95,
it's a different story with Direct X.

**********************************************************************
**********************************************************************

Alex Fernandez This signature file serves no other purpose
"Don't know nothin'!" than to try to avoid those unholy NNTP errors
I keep getting using Netscape 2.0 Mail when I
follow up on articles where there is more
included text than new text. Does everyone
have this problem? Is there a better workaround than this stupid, long
and useless signature file? If so, let me know. I hope this is enough
new text.
**********************************************************************
**********************************************************************


Jon

unread,
Jul 6, 1996, 3:00:00 AM7/6/96
to

On Sat, 06 Jul 1996 05:45:12 GMT, alex...@shore.net (Alex Fernandez)
wrote:

>On Sat, 06 Jul 1996 02:11:31 GMT, Kry...@lightspeed.bc.ca (EB) wrote:
>>As far as can see thus far, there is no other card on the PC market this year
>>that will beat the 3Dfx/Orchid cards. I'm terribly anxious to see a live demo
>>of it in action. I think I'll have to get one. I've got a neeeed 4 speeeed. :)
>>(I sure hope a 3D version of MW2 will be made for the voodoo! :)
>>
>> -EB
>>
>>______________________________
>>ampl...@tamu.edu (Jon) wrote:

(hey, cut and paste correctly if you do it, I didn't write anything
here!)

>I'm kinda getting tired reading these posts about which chip is better than the
>other. It all comes down to which games are written for which chips. I haven't
>seen much in the press about game companies supporting the voodoo chip.
>But I have seen press releases about companies supporting the Rendition chip in
>their games. I play ICR2, NASCAR, and will play F1GP2. I know that Papyrus will
>support the Verite. Because of that support, I will look seriously at the
>products based on the Verite. Also, as I understand it, the voodoo does not
>have a VESA 2.0 compliant VGA/SVGA core (correct me if i'm wrong) since it
>requires a 2D board. The Verite does have a VESA 2.0 compliant VGA/SVGA core
>that can sustain the full 133MB/s PCI thruput. The following is a quote from
>one of their reps:

You must not realize that future games will not need particular
support since the hardware company will just write Direct3D support.
Until then, yes, creative labs has a good foot in getting DOS based
supporting their card. But this isn't the end all reason to get this
or that card, since this isn't the near future usage!

Yes, the voodoo needs a 2d card.

>It all comes down to software support for the 3D features in DOS. But in Win95,
>it's a different story with Direct X.

No, it doesn't, and yes, it is, and that's why the former isn't true!

Jon

unread,
Jul 6, 1996, 3:00:00 AM7/6/96
to

On Fri, 05 Jul 1996 22:56:36 -0700, Eric Calcagni
<calc...@3-cities.com> wrote:
>I hope I'm wrong though. Researching 3D cards for the past
>8 months has worn me out. ;) I'm just ready to buy a damn card
>and start playing!

Here, here!!! Same. :)

ALX

unread,
Jul 6, 1996, 3:00:00 AM7/6/96
to

In article <31ddfa97...@news.shore.net>, alex...@shore.net (Alex
Fernandez) wrote:

> I'm kinda getting tired reading these posts about which chip is better
> than the other. It all comes down to which games are written for
> which chips. I haven't seen much in the press about game companies
> supporting the voodoo chip.

Why not visit their web page? 3DFX has listed quite a number of
prominent developers including quotes from company programmers/
representatives.



> But I have seen press releases about companies supporting the
> Rendition chip in their games. I play ICR2, NASCAR, and will
> play F1GP2. I know that Papyrus will support the Verite. Because
> of that support, I will look seriously at the products based on
> the Verite.

Once again, visit 3DFX's web page. They have Papyrus' support
as well:

"We've been working with a number of 3D graphics hardware
technologies, and we find developing for the 3Dfx Interactive
Voodoo Graphics environment to be very intuitive, which has made
our game development easier. Their emphasis on high-quality
development tools and responsive support gives us a competitive
edge that will let us get our games to market faster."
-- David Kaemmer, chief technical officer, Papyrus Design Group

> The Verite does have a VESA 2.0 compliant VGA/SVGA core

True. It depends on your needs and preferences though:

If you're seeking all-out, blazing fast 3D performance and can live
with your current 2D card, go with the Voodoo. If you're looking
for an all-around/replacement card, from what I've been reading thus
far, the Verite based boards are for you. Of course, an even better
combo (more expensive of course), is a Voodoo/ET6000 combo, or if
compatible, a Voodoo/Millenium combo.

> It all comes down to software support for the 3D features in DOS.

Again, visit their web site. I think you'll be quite surprised
with the number of developers.

->ALX<-


ALX

unread,
Jul 6, 1996, 3:00:00 AM7/6/96
to

In article <31DE00...@3-cities.com>, Eric Calcagni <calc...@3-cities.com>
wrote:

> Orchid's web site, to put it succinctly, sucks. They give NO
> information on the Righteous 3D. EVERY other company that

Well, that's not entirely true. There's a short section on the
Righteous 3D within the press-release dealing with the Micronics/3DFX
combo.

> bit of information on it. In fact, I'm beginning to wonder
> if Orchid is going to update their website -after- they release
> the Righteous 3D. This makes me worry just a tad. I almost

I'm just glad 3DFX is actually developing all the drivers. At
this point, it seems 3DFX will handle driver updates, so the lack
of info on Orchid's site doesn't seriously bother me.

> I will probably go with the Orchid card anyway. Hell, it's
> only a video card, but I like the chips it has on it. :) I
> looked at Fujitsu's web page, and they mention nothing about their
> Voodoo based card either. What's the deal here???

It looks like Fujitsu's been dropped. 3DFX has no mention of
them either. I remember though, someone in this newsgroup a few
days ago, saying that another *major* manufacturer would be
producing Voodoo based boards. He didn't give any clues though.

->ALX<-


Eric Calcagni

unread,
Jul 6, 1996, 3:00:00 AM7/6/96
to

ALX wrote:
>
> It looks like Fujitsu's been dropped. 3DFX has no mention of
> them either. I remember though, someone in this newsgroup a few
> days ago, saying that another *major* manufacturer would be
> producing Voodoo based boards. He didn't give any clues though.
>
> ->ALX<-

Heh heh. That's because that person was me. ;) I think
people misunderstood me and thought that I knew the answers.
I don't. I'm just a standard consumer like you guys. What
I did was pass along what a 3DFX rep told me. And he didn't
tell me anything except what I posted to you guys.

Why do you say Fujitsu's been dropped? Where did you get that
info? It may well be true, I just haven't heard of it.
Was it in a press release that I missed or something?

- Eric C.

Scott Pedersen

unread,
Jul 7, 1996, 3:00:00 AM7/7/96
to

Eric Calcagni wrote:

> In fact, I have gotten better support from 3DFX than from any company
> I have ever purchased anything from - and I don't own any 3DFX stuff
> yet.
>
> I know I'll eventually have to deal with technical support from
> the companies that actually make the video cards that support
> have the voodoo on board, and I'm not expecting much from the current
> offerings. But I am so impressed by 3DFX's complete and sound answers
> to my questions that they have my business.
>

> Plus, the voodoo will probably be the hottest thing out there (next to
> PowerVR and Read 3D). ;)
>
> So if you guys are looking for a company that seems to care about
> this 3D "revolution" we are in, I'd suggest you ask them a few
> questions. Damn nice people.
>
> Later - Eric C.

Uhm, I did ask them a few questions Eric, and unlike yourself they chose
not to reply to my email. I don't quite know why? Where did you send
your email to? If it was in...@3Dfx.com, then I'm stumpted, maybe they
just don't like Aussies?.......(only joken', its just I'm just a little
surpissed- having read all the 'positive' email correspondance posts)

PS: Maybe you could help me out (seeing as they listen to you- or a
least make an effort to). I was interested in knowing what packins are
planned for the cards (i presume these have been decided on since the
card is supposed to go on sale this month), and secondly; how much are
the Voodoo graphics cards (specifically the 'Orchid' 4 meg EDO ram
version) going to cost. Oh and if any out there knows; Do Orchid
actually distribute and sell video cards in Australia (god I hope
so....)?

Dave Glue

unread,
Jul 7, 1996, 3:00:00 AM7/7/96
to

On Sat, 06 Jul 1996 05:45:12 GMT, alex...@shore.net (Alex Fernandez)
wrote:


>I'm kinda getting tired reading these posts about which chip is better than the
>other. It all comes down to which games are written for which chips. I haven't
>seen much in the press about game companies supporting the voodoo chip.

Do a search for "Dimension 3D", a web page dedicated specifically to
3D cards, it is extremely comprehensive. Look at the DOS titles
planned for the Voodoo. I'd say support is not a problem.

Dmitriy Veremeev

unread,
Jul 7, 1996, 3:00:00 AM7/7/96
to

gar...@3dfx.com (Gary McTaggart) wrote:

>In article <4r7s6d$c...@news.onramp.net> dar...@onramp.net (Daron Myrick) writes:
>
>> |>Theoretically: If both products are released at approximately the same
>> |>time, yet the 3DFX delivers twice the performance of the Verite, are
>> |>you going to go with Rendition because they've been around longer? I
>> |>really doubt it.
>>
>> I correct myself by stating that the Rendition I was thinking of was not the
>> Rendition company that makes the Verite.
>> As far as which company is newer, Rendition is but only by less than a year.
>> I'm sold on the Voodoo after visiting their site.
>> Too bad we have to buy an Orchid card for Voodoo power since I always hated
>> Orchid.
>> I'm tempted to buy the Voodoo card make by 3DFX themselves since at least I
>> know I won't be getting support and drivers from a company that's one step up
>> from Diamond.
>> Hopefully the card will be in the $500 to $600 range so that I can afford it
>> though it's still quite steep.
>

>The $300 card from Orchid is every bit as powerful as the 3Dfx
>Obsidian board in my machine. You will get the same 3Dfx developed
>and tuned drivers with the Orchid board that we use with the
>Obsidian. The Obsidian is meant to be more expandable for
>arcade/location based entertainment (LBE) use and also a reference
>board for OEMs.

Well, but will there be available Add-on card version of the VooDoo?
Especially for Matrox Millenium (as for one of the fasters gfx cards)?

>
>Gary McTaggart
>3Dfx Interactive
>
>--
>~/.signature

BYE.
Yours,
LodeRunner (Dmitriy Veremeev). E-mail: mini...@rinet.ru

Walter Meyer

unread,
Jul 7, 1996, 3:00:00 AM7/7/96
to

dav...@interlog.com (Dave Glue) writes:

The URL is:

http://www.excalibur.net/~3d/

It's a good site.

Walter

ALX

unread,
Jul 7, 1996, 3:00:00 AM7/7/96
to

In article <31DF19...@3-cities.com>, Eric Calcagni <calc...@3-cities.com>
wrote:

> Why do you say Fujitsu's been dropped? Where did you get that


> info? It may well be true, I just haven't heard of it.
> Was it in a press release that I missed or something?

I'm basing it on 3DFX's web site. A while back, under the products
section, they had both the Orchid and Fujitsu products listed. Now
the Fujitsu name is no longer there.

->ALX<-


C. Aardvark

unread,
Jul 8, 1996, 3:00:00 AM7/8/96
to

Brian Hook (b...@netcom.com) wrote:
: > The Voodoo has a SOLID setup of polygons. Verite, due to his RISC core, is
: > programmable so that the setup can be expanded if a new method/setup is
: > coming on the market... And RISC core MEANS SPEED AND POWER for those
: > kids who want speed... here, you have power AND speed...

: Out of curiousity, why does a RISC core mean speed and power, given that ASICs in general are

: significantly faster by design (you don't need to interpret instructions).

: Also, please check the performance of the Verite and the Voodoo Graphics and tell me, once
: again, how the RISC score is faster and more powerful.

: Brian

Here Here! Good show.. just because something is implemented in RISC
does not guarantee it will be fast. High chip frequencies are easier
to obtain in a RISC architecture due to their (relatively) simpler
design vs. CISC (for example). What some seem to forget is the fact
that it takes a boatload of MHz in a RISC chip to outstrip some
hardwired solution (fixed and not programmable). Those companies
devoted to RISC are hoping to push speeds higher while they lower
costs.. fair enough, it's obviously possible. Others will choose a
more hardwired or hardcoded solution.. equally capable and possible.
..sometimes more costly. Chip realestate is not cheap.. it takes a
lot of space to implement a lot of things in hardware vs. microcode...

Only impartial benchmarks will prove which is best.. but this
belief that only RISC chips will be worth a flip soley because they
are RISC is not logical. It's just as bad as people comparing
the clock fequencies of totally unrelated CPU's these days.

Oh well.. there's my useless addition to try and stop the RISC-worship :)

-James (who thinks RISC is great in the right applications!)

Daron Myrick

unread,
Jul 8, 1996, 3:00:00 AM7/8/96
to

RISC is just that, RISC (Reduced Instruction Set).
On the other hand, there is faster than RISC, which is optimized hardware that
uses thousands to millions of logic circuits in the form of a single chip,
that uses no instructions at all.
For instance, hardware level multipliers have been used for years, which can
multiply SEVERAL integers or floating point numbers in one clock cycle
compared to a single clock cycle or several clock cycles for a RISC chip to
CALCULATE a multiply.
Same with line drawing, since the line drawing logic isn't done through a
instruction processor on your accelerator card, it's done through optimized
hardware in the form of logic circuits which are much faster.
The drawback is that the lack of a processor is there is no hardware
expandibility, so the object is to put as much and more on the chip as is
needed at the time, and have some fast routines for a system CPU to perform
when graphics expansions are needed.

Brad Konopik

unread,
Jul 13, 1996, 3:00:00 AM7/13/96
to

In article <31D4DF...@3-cities.com> calc...@3-cities.com writes:
>Ok. I've been researching heavily into 3D cards in the past several
>months. I have seen all the specs, all the gee-wiz items and
>useless polygon counts. After all of this, I have decided I am going
>to get the Voodoo chipset. They seem to have a winning chipset
>on their hands (I hope :) But I just wanted to add...
>
>3DFX has won my over by their wonderful technical support online.
>Every e-mail I have sent, they have answered with a personal answer.
>EVERY one.

yeah -- but wait til they have product out and 1,000,000 customers to
support ... then see how fast they turn your email around :-)

--
internet: kon...@austin.ibm.com | Brad Konopik
uunet: uunet!austin.ibm.com!konopik | IBM AIX Technical Support
ibmvnet: KONOPIK at AUSTIN | Austin, Texas

Brad Konopik

unread,
Jul 13, 1996, 3:00:00 AM7/13/96
to

In article <4rcram$l...@news.onramp.net> dar...@onramp.net (Daron Myrick) writes:
>In article <31D9B6...@netcom.com>, b...@netcom.com wrote:
>|>*AHEM*, read a computer architecture text book when you get a chance and come
>|> back and remind me
>|>what PIO means (in the context of my post).
>
>PIO as in Peripheral Input/Output device, or a simple chip that takes the
>place of several logic chips, giving I/O capability including handshaking and
>port to bit level priority processing.
>I remember PIO chips very well back from the early 1980's.
>Actualy there were 8085 and 8080 PIO chips along with Z-80 PIO chips too, long
>before the 8086 and 8088 and the PC in general.
>Ahhh, brings back mammories of the good old days :-) .

Ahhh, yes ... fond ol' mammories...back when Motorola 6845 CRT Controllers
were the things we programmed.

>Have no Idea what PIO stands for today, though I know what it does and how it
>works.

PIO in my office means Programmed I/O -- as in simple In/Out instructions
to the IO channel -- as in NOT DMA.

0 new messages