Early Righteous reviews: comments

54 views
Skip to first unread message

Dave Glue

unread,
Oct 4, 1996, 3:00:00 AM10/4/96
to

Basically, if you've learned one thing from these early reviews: processor
speed _still_ matters a great deal. Something I've been telling people
over and over, but even I was a litlte surprised at these results.

The Monster Truck/Hellbender comments on a P100 were quite disappointing-
cripes, it doesn't sound as if they're beating the Virge that much with
these two! After numerous discussions with 3DFX employees, I am a little
underwhelmed- we were led to believe a P90-100 would be enough to get 25+
fps on _everything_. And heck, it's not like most of the games were
graphical powerhouses. If the Voodoo can't run them all at 30fps, it's a
sad comment on PC 3D (hardware and software) and the PC architecture in
general. It will be interesting to see how the 3D Blaster fares. Very
early I know, but I was a little surprised with the Voodoo's much-touted
power, I wouldn't think you would need a P166 just get a consistent 30fps
on everything.

A helpful equation to keep in mind? Take any frame rates on a system a rep
from a 3D company gives you, and HALVE them. That's your _real_ frame rate.
:)

Erick Cid

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

In article <01bbb240$811c7ec0$dc98d4c7@dave>, dav...@interlog.com says...

>
>A helpful equation to keep in mind? Take any frame rates on a system a rep
>from a 3D company gives you, and HALVE them. That's your _real_ frame rate.
>:)
>
>

Respectfully, lets wait till we here some others reviews before we "Rush to
judgement ;)" Anyway, I think its been hinted about that a 133 would probably
be the realistic minumum for full use of a 3d accelerator. When I get mine
tomorrow (doubtful) or Monday, I will post *very* complete and detailed
benchmarks. My machine is a P-200, so I assume Ill be getting full frames from
these games. If there is slowdown on this puppy, then I think its time to
complain.

Erick


--
--------------------------------------------------
Erick Cid
el...@interport.net
http://www.interport.net/~elcid
--------------------------------------------------


Dave Glue

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to


Erick Cid <el...@interport.net> wrote in article
<534bpb$1...@park.interport.net>...

> Respectfully, lets wait till we here some others reviews before we "Rush
to
> judgement ;)" Anyway, I think its been hinted about that a 133 would
probably
> be the realistic minumum for full use of a 3d accelerator. When I get
mine
> tomorrow (doubtful) or Monday, I will post *very* complete and detailed
> benchmarks. My machine is a P-200, so I assume Ill be getting full frames
from
> these games. If there is slowdown on this puppy, then I think its time to

> complain.

Actually, looking back at that review the user had a Neptune chipset, and
his 256k cache was probably not pipeline burst, not to mention non-EDO
memory. That's a P100, but likely quite a slow P100 by today's
comparisons. I imagine a good P133 EDO PB cache system should perform
significantly better, more so than just the clock speed increase would
indicate. Anyone with a decent P133+ system, post your review when you get
it!

BP

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

On 5 Oct 1996 00:57:15 GMT, el...@interport.net (Erick Cid) wrote:

>In article <01bbb240$811c7ec0$dc98d4c7@dave>, dav...@interlog.com says...
>
>>
>>A helpful equation to keep in mind? Take any frame rates on a system a rep
>>from a 3D company gives you, and HALVE them. That's your _real_ frame rate.
>>:)
>>
>>
>

>Respectfully, lets wait till we here some others reviews before we "Rush to
>judgement ;)" Anyway, I think its been hinted about that a 133 would probably
>be the realistic minumum for full use of a 3d accelerator. When I get mine
>tomorrow (doubtful) or Monday, I will post *very* complete and detailed
>benchmarks. My machine is a P-200, so I assume Ill be getting full frames from
>these games. If there is slowdown on this puppy, then I think its time to
>complain.
>

>Erick

OK I don't get it.
Why do they call them accelerators if you need a p200 to run game and
full speed ?

It seems like they try to much and achieve to little.
I want a card that has great graphics and makes a p100 run like its a
200.

Why not just use 256 colors ? It not like a games artist actually uses
32,000 color when they draw. I bet the r3d would fly in 640x480 256
mode and still look great.

I think I'll wait for a true accelerators one that accelerates the
486/100 in my closet or wait until that killer flight sim that I just
can't stand not to have appears.

Right now they don't seem to accelerate anything but how fast money
drains out of our wallets.

Oh the hell with it I'm going to order one now ;-)

Jon

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

six...@worldnet.att.net (BP) wrote:

>On 5 Oct 1996 00:57:15 GMT, el...@interport.net (Erick Cid) wrote:
>
>>In article <01bbb240$811c7ec0$dc98d4c7@dave>, dav...@interlog.com says...
>>
>>>
>>>A helpful equation to keep in mind? Take any frame rates on a system a rep
>>>from a 3D company gives you, and HALVE them. That's your _real_ frame rate.
>

>OK I don't get it.
>Why do they call them accelerators if you need a p200 to run game and
>full speed ?

Because the 3D cards offer other features besides raw speed, they look
much better at the same resolution. Realize that resolution isn't
everything, which has been the case so far with PC graphics mostly.

>
>It seems like they try to much and achieve to little.
>I want a card that has great graphics and makes a p100 run like its a
>200.

A p200 could NOT get but around 2fps with all the same options enabled
where the OR3D gets 60-70fps on a P166 with all options enabled.
Again, it's about looks, not just raw resolution as well.

>
>Why not just use 256 colors ? It not like a games artist actually uses
>32,000 color when they draw. I bet the r3d would fly in 640x480 256
>mode and still look great.

There is a drastic difference between the two, get the DirectX2.0
demos and play around with the color modes, and you'll see what I
mean.

>
>I think I'll wait for a true accelerators one that accelerates the
>486/100 in my closet or wait until that killer flight sim that I just
>can't stand not to have appears.

It's actually best if a 3d card is "scalable", meaning the graphics
performance scales with the CPU, because the CPU does and will always
share the task of rendering complex graphics.

>
>Right now they don't seem to accelerate anything but how fast money
>drains out of our wallets.

:), is true somewhat.

>
>Oh the hell with it I'm going to order one now ;-)

I'm debating that myself.


Smoke Crack and Worship Satan

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

In article <01bbb288$8ea205e0$dc98d4c7@dave>, "Dave Glue" <dav...@interlog.com>
wrote:

>
>
> Actually, looking back at that review the user had a Neptune chipset, and
> his 256k cache was probably not pipeline burst, not to mention non-EDO
> memory. That's a P100, but likely quite a slow P100 by today's
> comparisons. I imagine a good P133 EDO PB cache system should perform
> significantly better, more so than just the clock speed increase would
> indicate. Anyone with a decent P133+ system, post your review when you get
> it!
>

Dave, you know my home system. I rushed over to work after reading these
messages at 1 AM, and sure enough the package was sitting on my chair.
I've been out of town so I had no idea, but the timing was perfect!

P5-133, Triton-I chipset, 512k PBC, Matrox Millenium 2MB.

I have the full version of Monster Truck Madness, so I want to comment
on that first. Without the 3Dfx card, 640x400 mode is unplayable on
my machine. With the 3Dfx card, at 640x400x16 bit mode with all
3D features on it looks really close to 30FPS all around. I don't
know how to get a frame rate counter, but it is definitely smooth.
I'd say 25-30fps. It beats 640x400x8 handily on my machine. And it
looks really good. There are some strange visual artifacts on the
horizon, but overall the game looks great and runs *fast*.

Descent 2: Quartzon is also much faster than software at 640x400x16.
I can't get the "framerate" code to work, so I don't know the exact
fps but it is over 30fps in most rooms so far.

Conclusion: this card is obviously faster than software.

However, there *is* a good Win95 benchmark with FPS on the CD.

1. Copy the r3ddemo\wiz folder from the CD to your hard drive.
2. Make sure you mark all the files you copied as read/write or
the demo will not run. To do this, just select all the files
right click them, select properties, and turn OFF the read
only flag.
3. Double-click on WIZ.EXE, wait til it starts up and you are
in a 3D room.
4. Press CTRL+P.
5. it will say "INIT", wait, then show you a FPS rate.

This gives you a good framerate counter which we can compare
to see how much effect processor has on this card.

I get 29.9 FPS from the starting point. Interestingly, CTRL+M turns
mipmapping on/off, and CTRL+B turns bilinear filtering on/off. Neither
one has any real effect on the frame rate.


Eric Calcagni

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

Dave Glue wrote:
>
> Actually, looking back at that review the user had a Neptune chipset, and
> his 256k cache was probably not pipeline burst, not to mention non-EDO
> memory. That's a P100, but likely quite a slow P100 by today's
> comparisons. I imagine a good P133 EDO PB cache system should perform
> significantly better, more so than just the clock speed increase would
> indicate. Anyone with a decent P133+ system, post your review when you get
> it!

Yes. You are correct. My system is on the very low end of the
performance spectrum when compared to other P100's (Remember I'm an
overclocked P90. The cache is definitely not pipeline). I purchased my
system over two years ago when a P90 was the fastest system around. So,
by today's standards, I have a pretty pathetic system as far as a
Pentium system is concerned.

Anyway, I thought that my system would be a good system to review the
R3D for that very reason. After reading my review, there is no reason
to believe that their system will perform any worse. Of course, from my
perspective, that sucks. ;)

- Eric Calcagni

Eric Calcagni

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

Smoke Crack and Worship Satan wrote:
>
> Descent 2: Quartzon is also much faster than software at 640x400x16.
> I can't get the "framerate" code to work, so I don't know the exact
> fps but it is over 30fps in most rooms so far.
>

Type "framerate" in the middle of playing a game.

- Eric C.

Eric Calcagni

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

BP wrote:
>
> OK I don't get it.
> Why do they call them accelerators if you need a p200 to run game and
> full speed ?
>
> It seems like they try to much and achieve to little.
> I want a card that has great graphics and makes a p100 run like its a
> 200.
>
> Why not just use 256 colors ? It not like a games artist actually
uses
> 32,000 color when they draw. I bet the r3d would fly in 640x480 256
> mode and still look great.
>

I can tell you right now that my P100 runs faster than a non-accelerated
200 mhz Pentium Pro at 640x480x16 on every demo that came with the
card. Plus, the graphics look better. The $300.00 upgrade was
definitely worth it to me.

Now all I need to do is save up and GET a PPro to really make this thing
fly like an SGI. ;)

- Eric Calcagni.

Scott Mathers

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

Eric Calcagni <calc...@mindspring.com> wrote:

Close. It is "frametime" No "" of course :)

Samuel S. Paik

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

>Why do they call them accelerators if you need a p200 to run game and
>full speed ?

It's a good question. Generally, a 3D accelerator requires some minimal
performance out of the CPU before you get useful frame rates out of them.
On the other hand, it would require a _very_ fast CPU before you could
get the same image quality that you would out of most 3D accelerators.
I would estimate you would need a 2 GFLOP+ CPU to achieve the same image
quality and frame rate you would get out of a 3Dfx Voodoo Graphics chip
set.

Voodoo Graphics was targeted for people with a Pentium with a reasonable
bus, say a P100 Triton based system. I ran stuff on a P75 with a Neptune
PCI chip set for awhile, and Voodoo was always faster than software.

>Why not just use 256 colors ? It not like a games artist actually uses
>32,000 color when they draw. I bet the r3d would fly in 640x480 256
>mode and still look great.

There are many good reasone why 3D accelerators don't support 8bpp modes.

>wait until that killer flight sim that I just
>can't stand not to have appears.

EF2000? Jane's Longbow? Back to Baghdad? Falcon 4.0?

Sam Paik

--
408-749-8798 / pa...@webnexus.com
I speak for xyne KS since I AM xyne KS.

Tom Riegsecker

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

Erick Cid wrote:
>
> In article <01bbb240$811c7ec0$dc98d4c7@dave>, dav...@interlog.com says...
>
> >
> >A helpful equation to keep in mind? Take any frame rates on a system a rep
> >from a 3D company gives you, and HALVE them. That's your _real_ frame rate.
> >:)
> >
> >
>
> Respectfully, lets wait till we here some others reviews before we "Rush to
> judgement ;)" Anyway, I think its been hinted about that a 133 would probably
> be the realistic minumum for full use of a 3d accelerator. When I get mine
> tomorrow (doubtful) or Monday, I will post *very* complete and detailed
> benchmarks. My machine is a P-200, so I assume Ill be getting full frames from
> these games. If there is slowdown on this puppy, then I think its time to
> complain.
>
> Erick
>
> --
> --------------------------------------------------
> Erick Cid
> el...@interport.net
> http://www.interport.net/~elcid
> --------------------------------------------------

I have a P133, 32MB EDO, PB Cache...everything seems to run at 30+ fps
(Descent 2, R3D Demos, MTM, etc).

Jeremy

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

Eric Calcagni <calc...@mindspring.com> wrote:

I used to have an Intel Plato (Neptune ?) motherboard with a P90 and
256K async cache (pre Triton). I upgraded it to an ASUS P55TP4XE
motherboard (Triton 1) with 256 k PB cache and that alone made my 3D
Bench score go from 83 to 100 (20% increase). When I then swapped the
P90 for a P133 my score went from 100 to 125 (another 20%). This
seemed to make a big difference to SVGA like Screamer were unplayable
in SVGA before the upgrade.

Anyway the point is that the motherboard alone can make a big
difference.

I can't wait to get my Righteous 3D but us poor suckers in the UK have
to wait until 14 October for it to be released :-(

Jeremy

Condor

unread,
Oct 5, 1996, 3:00:00 AM10/5/96
to

On Sat, 05 Oct 1996 08:05:36 GMT, six...@worldnet.att.net (BP) wrote:

> On 5 Oct 1996 00:57:15 GMT, el...@interport.net (Erick Cid) wrote:
>
> >In article <01bbb240$811c7ec0$dc98d4c7@dave>, dav...@interlog.com
says...
> >
> >>
> >>A helpful equation to keep in mind? Take any frame rates on a
system a rep
> >>from a 3D company gives you, and HALVE them. That's your _real_
frame rate.
> >>:)
> >>
> >>
> >
> >Respectfully, lets wait till we here some others reviews before we
"Rush to
> >judgement ;)" Anyway, I think its been hinted about that a 133
would probably
> >be the realistic minumum for full use of a 3d accelerator. When I
get mine
> >tomorrow (doubtful) or Monday, I will post *very* complete and
detailed
> >benchmarks. My machine is a P-200, so I assume Ill be getting full
frames from
> >these games. If there is slowdown on this puppy, then I think its
time to
> >complain.
> >
> >Erick
>

> OK I don't get it.

> Why do they call them accelerators if you need a p200 to run game
and
> full speed ?
>

Unaccelerated video may run say 10fps. Accelerated video will set to
to say 15fps. This is based on the same clock.

> It seems like they try to much and achieve to little.
> I want a card that has great graphics and makes a p100 run like its
a
> 200.

Why don't you design one such? No hard feelings.

>
> Why not just use 256 colors ? It not like a games artist actually
uses
> 32,000 color when they draw. I bet the r3d would fly in 640x480 256
> mode and still look great.

True. Programmers or rather, graphics artistes rarely uses 32K colors.
They may only use 200 colors but think, they are using shades ouside
the 256 colors boundary. Only at 32K colors can they get the shade
they want and you want. Get the idea?

>
> I think I'll wait for a true accelerators one that accelerates the

> 486/100 in my closet or wait until that killer flight sim that I


just
> can't stand not to have appears.
>

> Right now they don't seem to accelerate anything but how fast money
> drains out of our wallets.

It's just that hardware advances does not parallel software
requirements.

>
> Oh the hell with it I'm going to order one now ;-)
>
>
>
>
>
> >
> >

Colin Barnowe

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to


Kai <kai...@msn.com> wrote in article <00001c05...@msn.com>...


> >Now all I need to do is save up and GET a PPro to really make this thing
> >fly like an SGI. ;)
> >- Eric Calcagni.
>

> I just did that today - went from a P166/512PB to a PPro200/256, and
> due to the 3dfx's coprocessor nature, there was a significant but not
> huge improvement. The Righteous Direct3D Control Panel test/demo
> went from ~47 to ~58 fps, while regular 2D graphics (Millenium)
> nearly doubled in SVGA due to posted writes.

Quick question: are you using fastvid? If not, try it and get back to us?

Gregory Seid

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

Tiitus Tamminen wrote:
> BP (six...@worldnet.att.net) wrote:
> : Why do they call them accelerators if you need a p200 to run game and
> : full speed ?
> The problem is, these cards are designed so that your CPU still has to do
> part of the work. If you have a slow CPU, no matter how fast 3D accelerator
> you have, the accelerator has to wait for your slow CPU to feed it data.
> That means power go unused.

It's the lack of a little slab of silicon known as the geometry processor.
As far as I know, not a single user-level 3D card for PCs supports any type
of geometry processing, thus leaving your CPU to do all that work. There is,
however, the GLINT Delta (from 3DLabs) which does a great deal of geometry
processing. However, it's incredibly expensive. Still, I'd hedge my bets
that by the middle or end of next year, we should see 3DLabs' 'low end'
Permedia paired up with the Delta for about US$400. One more advantage to
the Delta is that it's scalable: you can add more than one Permedia (or
GLINT 500TX, depending upon which card you bought) for greatly increased
performance.

-----
Gregory Seid
gm...@lehigh.edu
http://www.lehigh.edu/~gms2/

Dave Glue

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to


Kai <kai...@msn.com> wrote in article <00001c05...@msn.com>...

> CPU _does_ make a difference with a 3D accelerator. I just upgraded from
> a P166/512PB to a PPro200/256 and my Righteous Direct3D test in their
> Control Panel page went from 47 to 58 fps...

Your Direct3D test should be much higher than that. See the threads on
maximizing the refresh rate.

Dave Glue

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to


Gregory Seid <gm...@lehigh.edu> wrote in article
<3257E1...@lehigh.edu>...

> It's the lack of a little slab of silicon known as the geometry
processor.
> As far as I know, not a single user-level 3D card for PCs supports any
type
> of geometry processing, thus leaving your CPU to do all that work. There
is,
> however, the GLINT Delta (from 3DLabs) which does a great deal of
geometry
> processing.

From what a rep from 3DFX told us her, it's simply geometry setup portion-
something the Rendition and 3DFX already do. Geometry set-up is but one
part of the pre-rendering stage. it's not a true geometry processor.

>However, it's incredibly expensive.

Actually no, the Delta version is only $100 more than the non-Delta
version.


Andreas Schildbach

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

"Colin Barnowe" <cbar...@darkwing.uoregon.edu> hat geschrieben:

>
>
> Kai <kai...@msn.com> wrote in article <00001c05...@msn.com>...

> > >Now all I need to do is save up and GET a PPro to really make this thing
> > >fly like an SGI. ;)
> > >- Eric Calcagni.
> >
> > I just did that today - went from a P166/512PB to a PPro200/256, and
> > due to the 3dfx's coprocessor nature, there was a significant but not
> > huge improvement. The Righteous Direct3D Control Panel test/demo
> > went from ~47 to ~58 fps, while regular 2D graphics (Millenium)
> > nearly doubled in SVGA due to posted writes.
>
> Quick question: are you using fastvid? If not, try it and get back to us?

To use Fastvid on the Voodoo is not that easy. If you don't supply any
parameters, Fastvid searchs for the LFB (linear frame buffer) of your
2D card. It will completely ignore your Voodoo.

What you have to do is to make out the memory range of the Voodoo and
set Fastvid parameters accordingly.

It is still unknown if Voodoo is compatible to Write Combining.

I don't think Write Combining will have that much of effect on the
Voodoo because after the textures have been loaded up there is not
much bandwidth used, unlike 3D games on 2D cards.

- Andreas

---

Andreas "Goonie" Schildbach
schi...@informatik.tu-muenchen.de
Ride more roller-coasters!

Gary Tarolli

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to dan...@micronics.com

I first want to say how rewarding it is to read all your
reviews after having worked on the design of Voodoo Graphics
(the chipset on the Orchid Righteous 3D board) for over two years.
I am one of the founders of 3Dfx and one of our goals was
to deliver the highest quality graphics possible to the PC gamer.
It was and still is a very risky proposition because of the cost
sensitivity of the marketplace. But your reviews help convince
me that we did the right thing.

I thought I would share with you a little bit about what is
inside the 3Dfx Voodoo Graphics chipset. There are 2 chips
on the graphics board. Each is a custom designed ASIC containing
approximately 1 million transistors. Although this number of
transistors is on the order of a 486, it is a lot more powerful.
Why? Because the logic is dedicated to graphics and there’s a
lot of logic to boot. For example, bilinear filtering of
texture maps requires reading four 16-bit texels per pixel (that’s
400 Mbytes/sec at 50 Mpixels/sec) and then computing the equation
red_result=r0*w0+r1*w1+r2*w2+r3*w3 where r0:3 are the four red
values and w0:3 are the four weights based on the where the pixel
center lies with respect to the four texels. This is performed
for each color channel (red, green, blue, alpha) resulting
in 16 multiples and 12 additions or 28 operations per pixel.
At 50 Mpixels per second that is 1,400 Mops/sec. The way this
is designed in hardware is you literally place 16 multipliers
and 12 adders on the chip and hook them together. And this is
only a small part of one chip. There are literally dozens of
multipliers and dozens of adders on each of the two chips dedicated
only to graphics. Each chip performs around 4,000 million actual
operations per second, of which around one third are integer
multiplies. These are real operations performed - if you were to
try to do these on a CPU (or a DSP) you must also do things like
load/store instructions and conditions. In my estimation it would take about
a 10,000 Mip computer (peak) to do the same thing that one of our
chips does. This is about 20 of the fastest P5-200 or P6-200 chips
per one of our chips. Not exactly cost-effective. So if you want
to brag, you can say your graphics card has approximately the same
compute power as 40 P5-200 chips. Of course, these numbers are more
fun than they are meaningful. What is meaningful in graphics is
what you see on the screen.

Now of course, if you were writing a software renderer for a game,
you wouldn’t attempt to perform the same calculations we perform on
our chip on a general purpose CPU. You would take shortcuts, like
using 8-bit color with lookup tables for blending, or performing
perspective correction every ‘n’ pixels. The image quality will
depend on how many shortcuts you take and how clever you are.
Voodoo Graphics takes no shortcuts and was designed to give you
the highest quality image possible within the constraint of 2 chips.
As your reviews have shown, it is evident that you can see the
difference in quality and performance.

Now I am sure the subject of triangle setup and geometry calculations
will come up sooner or later in this newsgroup. Let me make a
preemptive strike and answer your questions before you ask them.
There is no geometry acceleration on the board, where geometry is
defined as geometric transformation and lighting. In the Wizard’s
tower demo you see lighting being applied to texture maps through
the use of a ‘lighting map’, which is another texture map that
contains the results of off-line radiosity calculations. This is
not traditional lighting in the OpenGL sense, but is nonetheless a
very powerful method of performing static lighting. It is becoming
more popular with games and personally, I think its great! It
requires bilinear filtering AND high fill rates, both of which
the R3D card has.

Now back to triangle setup. The Voodoo Graphics chipset performs
about 2/3 of triangle setup in hardware. When designing Voodoo
Graphics I carefully studied exactly what triangle setup our design
required and we placed things that were hard for a Pentium to
perform in hardware and left things easy for a Pentium to perform
out of the hardware design. Our triangle engine is also very
efficient in that it requires less setup than most (I worked 9 years
at SGI and have a lot of triangle engine experience). The net
result is that the 1/3 of triangle setup we perform on the Pentium
is not many cycles at all. That is why our triangle numbers are so high.
With an efficient design, you can afford to use the Pentium to perform
some of your triangle setup. With an inefficient design, you cannot.
I know this is a very controversial subject, so I will stop right here.

I hope this answers some of your questions in advance. Thanks for
buying the board and I hope you enjoy it. As for the flight sims,
I am waiting for one too, and am anxiously awaiting to see what our
chips can do. I wrote the original SGI flight simulator and hopefully,
I won’t have to write another one :-}

Smoke Crack and Worship Satan

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

In article <3257E1...@lehigh.edu>, Gregory Seid <gm...@lehigh.edu> wrote:
> Tiitus Tamminen wrote:
> > BP (six...@worldnet.att.net) wrote:
> > : Why do they call them accelerators if you need a p200 to run game and
> > : full speed ?
> > The problem is, these cards are designed so that your CPU still has to do
> > part of the work. If you have a slow CPU, no matter how fast 3D accelerator
> > you have, the accelerator has to wait for your slow CPU to feed it data.
> > That means power go unused.
>
> It's the lack of a little slab of silicon known as the geometry processor.
> As far as I know, not a single user-level 3D card for PCs supports any type
> of geometry processing, thus leaving your CPU to do all that work. There is,
> however, the GLINT Delta (from 3DLabs) which does a great deal of geometry
> processing. However, it's incredibly expensive. Still, I'd hedge my bets
> that by the middle or end of next year, we should see 3DLabs' 'low end'
> Permedia paired up with the Delta for about US$400. One more advantage to
> the Delta is that it's scalable: you can add more than one Permedia (or
> GLINT 500TX, depending upon which card you bought) for greatly increased
> performance.
>

Did you read my post comparing 3dfx-enhanced Descent 2 framerates on
a P5-100, P5-133, and P5-166?

The frame rate at the start of level 1 without moving the ship was
30.0 on all those machines. The non-accelerated versions came in
at 14.5, 17.25, and 18.75 respectively.

Dave Glue

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to


Gary Tarolli <tar...@3dfx.com> wrote in article
<325839...@3dfx.com>...

> I hope this answers some of your questions in advance. Thanks for
> buying the board and I hope you enjoy it. As for the flight sims,
> I am waiting for one too, and am anxiously awaiting to see what our
> chips can do. I wrote the original SGI flight simulator and hopefully,
> I won’t have to write another one :-}

Greatly appreciated Gary. Mucho thanks!

Jon Oden

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

Colin Barnowe wrote:
>
> Kai <kai...@msn.com> wrote in article <00001c05...@msn.com>...
> > >Now all I need to do is save up and GET a PPro to really make this thing
> > >fly like an SGI. ;)
> > >- Eric Calcagni.
> >
> > I just did that today - went from a P166/512PB to a PPro200/256, and
> > due to the 3dfx's coprocessor nature, there was a significant but not
> > huge improvement. The Righteous Direct3D Control Panel test/demo
> > went from ~47 to ~58 fps, while regular 2D graphics (Millenium)
> > nearly doubled in SVGA due to posted writes.
>
> Quick question: are you using fastvid? If not, try it and get back to us?

I'm guessing that the direct 3D tunnel demo is not very processor
intensive and relies greatly on the 3D card
maybe a real world app would see a more dramatic increase with the
rightous involved?

Actually.. now that I think about it.. that shouldnt matter.. it might
be the way the tunnel test was coded..
I'm gonna hit send anyway.. hehehe

ALX

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

In article <325839...@3dfx.com>, Gary Tarolli <tar...@3dfx.com> wrote:

> I hope this answers some of your questions in advance.

It sure does, and thanks for the post. It's not often a company
head honcho actually takes the time to write such an extensive
thank you note on Usenet.

->ALX<-


William Ball

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

For Monster Truck Madness typing "frameit" (no quotes) brings up a
small frame rate indicator in the lower right corner.

Bill


Kai

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

>Now all I need to do is save up and GET a PPro to really make this thing
>fly like an SGI. ;)
>- Eric Calcagni.

I just did that today - went from a P166/512PB to a PPro200/256, and
due to the 3dfx's coprocessor nature, there was a significant but not
huge improvement. The Righteous Direct3D Control Panel test/demo
went from ~47 to ~58 fps, while regular 2D graphics (Millenium)
nearly doubled in SVGA due to posted writes.

Kai

Kai

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

>> Anyway, I think its been hinted about that a 133 would probably
>> be the realistic minumum for full use of a 3d accelerator. When I get mine
>> tomorrow (doubtful) or Monday, I will post *very* complete and detailed
>> benchmarks. My machine is a P-200, so I assume Ill be getting full
frames from
>> these games. If there is slowdown on this puppy, then I think its time to
>> complain.
> I have a P133, 32MB EDO, PB Cache...everything seems to run at 30+ fps
> (Descent 2, R3D Demos, MTM, etc).

CPU _does_ make a difference with a 3D accelerator. I just upgraded from


a P166/512PB to a PPro200/256 and my Righteous Direct3D test in their
Control Panel page went from 47 to 58 fps...

Kai

John L. Hardy IV

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

Kai, your final sentence piques my curiosity. The system you upgraded to I
have currently with a Matrox Mill. 4m.. I hope you were indicating that the
2d doubling you received was going from the p166 to the Pro200 and not a
speed increase from the addition of the 3dfx. <g>
--
John IV
jo...@gte.net / jha...@adobe.com
Adobe Systems: Internet/Persuasion SQA Team, Seattle, Wa

Kai <kai...@msn.com> wrote in article <00001c05...@msn.com>...

Andreas Schildbach

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

kai...@msn.com (Kai ) hat geschrieben:

Could you please post your results for Descent II and all the other
benchmarks on PPro 200? And if you can remember, please also post all
scores you got on the P166.

I'd really like to know if PPro 200 helps the Righteous 3D in reality.

Tony Marshall

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

jez...@frimley.demon.co.uk (Jeremy) wrote:

>
>I used to have an Intel Plato (Neptune ?) motherboard with a P90 and
>256K async cache (pre Triton). I upgraded it to an ASUS P55TP4XE
>motherboard (Triton 1) with 256 k PB cache and that alone made my 3D
>Bench score go from 83 to 100 (20% increase). When I then swapped the
>P90 for a P133 my score went from 100 to 125 (another 20%). This
>seemed to make a big difference to SVGA like Screamer were unplayable
>in SVGA before the upgrade.
>

I suppose you know all about 3D bench. It's ludicrously inaccurate (my
P120 also gets 125), but I take your point..

>Anyway the point is that the motherboard alone can make a big
>difference.
>
>I can't wait to get my Righteous 3D but us poor suckers in the UK have
>to wait until 14 October for it to be released :-(
>

Thanks for that. Oct14th eh - just enough time to let the dust settle
and see if it really lives up to the hype..

Tony Marshall | <tangerine dream.yello.mike oldfield
Carlisle, England | neuronium.depechemode.psb's.oasis.jmj
to...@zigzag.u-net.com| ashra.stranglers.vangelis.floyd.goa >
----------------------|--------------------------------------
My Mystery Tune Page http://www.personal.u-net.com/~zigzag

Tiitus Tamminen

unread,
Oct 6, 1996, 3:00:00 AM10/6/96
to

BP (six...@worldnet.att.net) wrote:

: Why do they call them accelerators if you need a p200 to run game and
: full speed ?

:
: It seems like they try to much and achieve to little.

: I want a card that has great graphics and makes a p100 run like its a
: 200.

The problem is, these cards are designed so that your CPU still has to do

part of the work. If you have a slow CPU, no matter how fast 3D accelerator
you have, the accelerator has to wait for your slow CPU to feed it data.
That means power go unused.

: Why not just use 256 colors ? It not like a games artist actually uses


: 32,000 color when they draw.

You are forgetting the shading of the graphics. Just look how dithered
those 8bit graphics look.

: I bet the r3d would fly in 640x480 256


: mode and still look great.

I think it would STILL have to wait for your slow CPU to do its part
of the work, so actually you might not get any more speed for your games.

: I think I'll wait for a true accelerators one that accelerates the
: 486/100 in my closet

Don\t expect that kind of accelerator to be that cheap. It should have
extra silicon to relieve your CPU completely from the task.


Jon

unread,
Oct 7, 1996, 3:00:00 AM10/7/96
to

Gary Tarolli <tar...@3dfx.com> wrote:

>I first want to say how rewarding it is to read all your
>reviews after having worked on the design of Voodoo Graphics
>(the chipset on the Orchid Righteous 3D board) for over two years.
>I am one of the founders of 3Dfx and one of our goals was
>to deliver the highest quality graphics possible to the PC gamer.
>It was and still is a very risky proposition because of the cost
>sensitivity of the marketplace. But your reviews help convince
>me that we did the right thing.

I personally held off getting the Orchid card with your chipset since
everything I purchase ends up no more than a year later becoming my
younger brother's. He loves games, but also likes Internet stuff, and
3d internet stuff. I'm a physics major and would also like 3d in a
window that's fast, for rendering 3dstudio or maple plots(as future
support is available).

Are there plans to integrate your chipset with a 2d chipset to make a
board that does this? Thanks.

>I hope this answers some of your questions in advance. Thanks for
>buying the board and I hope you enjoy it. As for the flight sims,
>I am waiting for one too, and am anxiously awaiting to see what our
>chips can do. I wrote the original SGI flight simulator and hopefully,
>I won’t have to write another one :-}

Ya, thanks. Even though this fact doesn't necessarily benefit you
directly, I'd like to say I'm glad you put an excellent chipset out;
it's pushing the 3d market beyond the current weak line of
chipsets...Giving us comsumers a choice.

Dave Glue

unread,
Oct 7, 1996, 3:00:00 AM10/7/96
to


Jon <ampl...@tamu.edu> wrote in article
<3258696b...@news.tamu.edu>...


> Are there plans to integrate your chipset with a 2d chipset to make a
> board that does this? Thanks.

Yes. The press release has been on their web site since it's inception.
Many moons ago 3DFX announced a partnership with Alliance semiconductor for
that very purpose.

BP

unread,
Oct 7, 1996, 3:00:00 AM10/7/96
to

On Sun, 06 Oct 1996 18:58:15 -0400, Gary Tarolli <tar...@3dfx.com>
wrote:

>I first want to say how rewarding it is to read all your
>reviews after having worked on the design of Voodoo Graphics

Snip Lots of cool Tech stuff


>waiting to see what our
>chips can do. I wrote the original SGI flight simulator and hopefully,
>I won’t have to write another one :-}

I hate you !!!

Detailed Tech stuff alway makes me want to go buy something I can't
afford.

Where's my wife's credit card ?..............Oh Honey.


Andy Hanson

unread,
Oct 7, 1996, 3:00:00 AM10/7/96
to

BP wrote:
>
> On Sun, 06 Oct 1996 18:58:15 -0400, Gary Tarolli <tar...@3dfx.com>
> wrote:
>
> >I first want to say how rewarding it is to read all your
> >reviews after having worked on the design of Voodoo Graphics
> Snip Lots of cool Tech stuff
> >waiting to see what our
> >chips can do. I wrote the original SGI flight simulator and hopefully,
> >I won’t have to write another one :-}
>
> I hate you !!!
>
> Detailed Tech stuff alway makes me want to go buy something I can't
> afford.
>
> Where's my wife's credit card ?..............Oh Honey.

It's not that bad, According to Chris Runges website (which is very cool
by the way -http://www.nd.edu/~crunge/3dfx/index.htm), Computer Quick is
selling it for $257(http://www.cqk.com)

Andy

Per Christian Krabsetsve

unread,
Oct 7, 1996, 3:00:00 AM10/7/96
to

Samuel S. Paik (pa...@webnexus.com) wrote:
: >Why do they call them accelerators if you need a p200 to run game and
: >full speed ?
: It's a good question. Generally, a 3D accelerator requires some minimal

: performance out of the CPU before you get useful frame rates out of them.
: On the other hand, it would require a _very_ fast CPU before you could
: get the same image quality that you would out of most 3D accelerators.
: I would estimate you would need a 2 GFLOP+ CPU to achieve the same image
: quality and frame rate you would get out of a 3Dfx Voodoo Graphics chip
: set.
I agree. The CPU still has to do the rotations and movement of the vectors,
it also has to check collisions between vectorobjects and a lot of stuff :)
When the accelerator does the drawing/shading/+++ it gives the programmer
more cputime for AI and calculations of the objects.

: >Why not just use 256 colors ? It not like a games artist actually uses

: >32,000 color when they draw. I bet the r3d would fly in 640x480 256


: >mode and still look great.

: There are many good reasone why 3D accelerators don't support 8bpp modes.
If you have a picture of a piece of wood and want to map it to vector wall,
you may want to shade it and have light/dark parts of the wall, Try having
lots of different btimaps with different colors and you'll soon see the
advantage of hicolor. Try to merge two 256 color bmps in photoshop or
anything equal, and you'll see a difference, try 5-6 different pictures and
the result is not so good, because they have to share a 256 color palette.
256 colors would destry much of the point of the accelerators, because there
wouldn't be enough colors for shades of the different colors and dithering
would suck.

-PerCK

Unknown

unread,
Oct 7, 1996, 3:00:00 AM10/7/96
to

On Sun, 06 Oct 1996 18:58:15 -0400, Gary Tarolli <tar...@3dfx.com>
wrote:

>I first want to say how rewarding it is to read all your


>reviews after having worked on the design of Voodoo Graphics
>(the chipset on the Orchid Righteous 3D board) for over two years.
>I am one of the founders of 3Dfx and one of our goals was
>to deliver the highest quality graphics possible to the PC gamer.
>It was and still is a very risky proposition because of the cost
>sensitivity of the marketplace. But your reviews help convince
>me that we did the right thing.
>

well, they did convince me, too.
now i only have to wait for a 3dfx board to arrive in germany :)

holger

-
I'm sorry for this, but as I'm getting more and more of this junk email
i'm forced to do this. My real E-Mail address is can be found below.

Ash nazg durbatuluk, ash nazg gimbatul, ash nazg | Holger Mueller
thrakatuluk agh burzum-ishi krimpatul. | ava...@teuto.de

Tony Marshall

unread,
Oct 7, 1996, 3:00:00 AM10/7/96
to

sha...@cc.hut.fi (Tiitus Tamminen) wrote:

>BP (six...@worldnet.att.net) wrote:
>
>: Why not just use 256 colors ? It not like a games artist actually uses
>: 32,000 color when they draw.
>

>You are forgetting the shading of the graphics. Just look how dithered
>those 8bit graphics look.
>

Bilinear filtering will use up a lot of colours too..

Samuel S. Paik

unread,
Oct 7, 1996, 3:00:00 AM10/7/96
to

>To use Fastvid on the Voodoo is not that easy. If you don't supply any
>parameters, Fastvid searchs for the LFB (linear frame buffer) of your
>2D card. It will completely ignore your Voodoo.
>
>It is still unknown if Voodoo is compatible to Write Combining.

Do not use Fastvid on Voodoo Graphics. The drivers already enable
some PCI enhancements on the Pentium Pro. Write Combining will not
help as there are no byte writes at all.

Sam Paik

--
408-749-8798 / pa...@webnexus.com
I speak for xyne KS since I AM xyne KS.

Tommy McClain

unread,
Oct 8, 1996, 3:00:00 AM10/8/96
to

Gary,

Thanks for the information and your presence.

Tommy McClain

Dimension 3D
http://www.Dimension3D.com/


J. Tige Richardson

unread,
Oct 8, 1996, 3:00:00 AM10/8/96
to

Is it OK to continue to use Fastvid for your 2D card when you DO have the
Righteous 3D? I am considering the Righteous 3D but I want to make sure I
can continue to use Fastvid.

Thanks for any help you can provide.

Tige

ma...@micronics.com

unread,
Oct 8, 1996, 3:00:00 AM10/8/96
to

Hi Tige.

I would hazard a guess that as long as you do not let FASTVID muck around
with the memory space used by the Orchid Righteous 3D (Voodoo graphics chip
set) you can continue to use it for your 2D card.

Sincerely,

Marc Warden, Orchid Technology (ma...@micronics.com)


Matthias Striegl

unread,
Oct 9, 1996, 3:00:00 AM10/9/96
to

mrb...@msn.com (William Ball) writes:

>Bill

Typing "frames" also does it ...

(Just for info, hmm....)

Matze

Tiitus Tamminen

unread,
Oct 11, 1996, 3:00:00 AM10/11/96
to

Gregory Seid <gm...@lehigh.edu> wrote:

>It's the lack of a little slab of silicon known as the geometry
>processor. As far as I know, not a single user-level 3D card for PCs
>supports any type of geometry processing, thus leaving your CPU to do
>all that work.

I don't think _all_ tasks should be taken away from the CPU. Only those
tasks with which it has severe difficulties.

Think if your PC was full of different co-processors, DSPs etc. for all
the tasks related to a game, and only 5% of the power of your CPU would
be used. Do you think that would be a cost-effective solution? Wouldn't
you feel bad already having a fast Pentium/PPro, the power of which
would never be used?

Now, I must admit I'm not quite sure how suitable a general-purpose CPU
like PPro is for geometry calculations (transformations etc.) compared
to a chip designed specifically for that task, maybe someone would shed
some light on this? I understand the current 3D accelerators are taking
away tasks that really are _not suitable_ for a Pentium/PPro. Is
geometry calculations also one of these tasks (that are not suitable
for a PPro, thus another extra chip should take care of it)? Just asking
those who are programming for both.


Reply all
Reply to author
Forward
0 new messages