This is not OT, it's directly relevant to this newsgroup.

> Just noticed this thread about the performance difference in Far Cry if
the
> Nvidia 6800 Device ID is set to ATI R300.
> [url]http://www.driverheaven.net/showthread.php?s=&threadid=44253.[/url] Haven't had
a
> chance to read it yet in depth, and I do hope Nvidia isn't cheating.
Again.

"the Radeon 9800XT is very close in [FarCry] performance to the
6800Ultra at the same IQ levels. Only 7fps in average framerates separate
the cards which is a phenomenal result for the Radeon."

What does this mean? My guess is that FarCry uses a mix of PS1.1 & PS2
shaders for the GeforceFX, and these settings are being carried over
unchanged to the 6800u, while the Radeon has always used PS2.0 alone. If
true, present FarCry 6800u benchmarks are pretty much worthless, especially
with the reported 6800u artifacts.

rms

rms wrote:
> This is not OT, it's directly relevant to this newsgroup.
> "the Radeon 9800XT is very close in [FarCry] performance to the
> 6800Ultra at the same IQ levels. Only 7fps in average framerates
> separate the cards which is a phenomenal result for the Radeon."
>
> What does this mean? My guess is that FarCry uses a mix of PS1.1
> & PS2 shaders for the GeforceFX, and these settings are being carried
> over unchanged to the 6800u, while the Radeon has always used PS2.0
> alone. If true, present FarCry 6800u benchmarks are pretty much
> worthless, especially with the reported 6800u artifacts.
>
> rms

I said OT only cause this involves NVIDIA moreso then ATI, at least it
sounded good in my head =). What disappoints me is I was kinda *hoping*
nvidia would perform that well (~60fps) in Far Cry. Instead they only get
~47fps, which isn't that much better then the 9800XT(~41). Now I really
can't wait to see what ATI will offer as their next-gen.

Khabs,
Mark

"The Mighty MF" <sether01@hotmail.com> wrote in message
news:108ve3qa7fncvf7@corp.supernews.com...

> I said OT only cause this involves NVIDIA moreso then ATI, at least it
> sounded good in my head =). What disappoints me is I

Well, since you're cross-posting to three different news groups, one of
which is Nvidia..... to the Nvidia news readers, it is relevant ;)

"rms" <rsquires@flashREMOVE.net> wrote:
> This is not OT, it's directly relevant to this newsgroup.
>
> > Just noticed this thread about the performance difference in Far Cry if
> > the
> > Nvidia 6800 Device ID is set to ATI R300.
> > [url]http://www.driverheaven.net/showthread.php?s=&threadid=44253.[/url] Haven't had
> > a chance to read it yet in depth, and I do hope Nvidia isn't cheating.
> > Again.
>
> "the Radeon 9800XT is very close in [FarCry] performance to the
> 6800Ultra at the same IQ levels. Only 7fps in average framerates separate
> the cards which is a phenomenal result for the Radeon."
>
> What does this mean? My guess is that FarCry uses a mix of PS1.1 & PS2
> shaders for the GeforceFX, and these settings are being carried over
> unchanged to the 6800u, while the Radeon has always used PS2.0 alone. If
> true, present FarCry 6800u benchmarks are pretty much worthless, especially
> with the reported 6800u artifacts.
>
> rms

To me, the real question is, what exactly can be achieved with PS2.0
and PS3.0 over PS1.1? I just didn't see a whole lot of differences
between my old GeForce 4200 and the newer 5900XT? Will the fish
shine? Do the birds change from fancy parrots to dirty seagulls when
using an old card? If the comparisons done in Geneva two weeks ago
were done in low quality PS1.1 against the highest quality PS3.0, then
that's really the wrong way to market the next generation card. If
the thugs have shining snot dribbling from the noses with PS3.0, maybe
I don't really want to see it!

Nada wrote:
> To me, the real question is, what exactly can be achieved with PS2.0
> and PS3.0 over PS1.1? I just didn't see a whole lot of differences
> between my old GeForce 4200 and the newer 5900XT? Will the fish
> shine? Do the birds change from fancy parrots to dirty seagulls when
> using an old card? If the comparisons done in Geneva two weeks ago
> were done in low quality PS1.1 against the highest quality PS3.0, then
> that's really the wrong way to market the next generation card. If
> the thugs have shining snot dribbling from the noses with PS3.0, maybe
> I don't really want to see it!

Good question. From what I understand, Farcry is only using PS2.0 for
lights/lighting effects. This comes from an article recently posted at
FiringSquad: [url]http://www.firingsquad.com/hardware/far_cry_nvidia/.[/url] It's an
interesting read if you haven't checked it out yet. I believe it also
mentions the PS3.0 vs PS1.1 comparision.

> > To me, the real question is, what exactly can be achieved with PS2.0
> > and PS3.0 over PS1.1?
> Good question.

The biggest new feature seems to be Displacement Mapping, which could
make surfaces look more real. Bump-mapped surfaces look flat close up, but
DM should improve on that.

Other than that PS3 seems to mostly be about internal improvements.

rms

On Thu, 29 Apr 2004 13:49:11 GMT, "rms" <rsquires@flashREMOVE.net>
wrote:

>> > To me, the real question is, what exactly can be achieved with PS2.0
>> > and PS3.0 over PS1.1?
>> Good question.
>
> The biggest new feature seems to be Displacement Mapping, which could
>make surfaces look more real. Bump-mapped surfaces look flat close up, but
>DM should improve on that.
>
> Other than that PS3 seems to mostly be about internal improvements.

Displacement mapping isn't really tied to Shader Model 3.0 (and is in
no way related to Pixel Shader 3.0)

In SM3.0, texture lookup can be done directly by the "Vertex Shader"
unit. In theory, VS3.0 do it in parallel with its other tasks thus
resulting in speed improvements.

ATI cards have had displacement mapping since the 9700 but they
certainly can't do it using VS3.0 unless the new ATI cards support
SM3.0.

The few pictures that nVidia released (from Far Cry) don't actually
use displacement mapping. They have offset/parallax mapping in there,
using pixel shaders, to give a depth perception on the stone walls. It
can be done easily by cards such as GeForceFx or Radeon 9500+ cards
and above and it gives you much better results than regular bump
mapping.

By the way, "Displacement Mapping" is a technique, where a flat
surface is tweaked - that is, certain portions of it are raised or
lowered automatically based on a map in a texture. Unlike, bump
mapping or offset mapping, there is no trick involved there. The 3D
surface is really modified, instead of just giving viewers a depth
perception. Designers can then set up a flat surface in 3D with a
displacement map and the 3D card can then add details to that surface
in real time.

I can't think of a game that has used real displacement mapping
though.
--
Noman

> Displacement mapping isn't really tied to Shader Model 3.0 (and is in
> no way related to Pixel Shader 3.0)
> ATI cards have had displacement mapping since the 9700 but they
> certainly can't do it using VS3.0 unless the new ATI cards support
> SM3.0.

If in any doubt, check out that Ferrari demo for radeon9700. It has
displacement mapping on wheels. Or maybe check out ATI website for more info
on how dx9 works. DM is also discussed there.

"The Mighty MF" <sether01@hotmail.com> wrote in
news:1091u1dicdcf8b0@corp.supernews.com:

> Nada wrote:
>> To me, the real question is, what exactly can be achieved with PS2.0
>> and PS3.0 over PS1.1? I just didn't see a whole lot of differences
>> between my old GeForce 4200 and the newer 5900XT? Will the fish
>> shine? Do the birds change from fancy parrots to dirty seagulls when
>> using an old card? If the comparisons done in Geneva two weeks ago
>> were done in low quality PS1.1 against the highest quality PS3.0,
>> then that's really the wrong way to market the next generation card.
>> If the thugs have shining snot dribbling from the noses with PS3.0,
>> maybe I don't really want to see it!
>
> Good question. From what I understand, Farcry is only using PS2.0 for
> lights/lighting effects. This comes from an article recently posted
> at FiringSquad: [url]http://www.firingsquad.com/hardware/far_cry_nvidia/.[/url]
> It's an interesting read if you haven't checked it out yet. I believe
> it also mentions the PS3.0 vs PS1.1 comparision.

I'm still wondering what the real differences will be between PS2.0 and
3.0 will be.

According to ATI, I get the impression they are saying they look the
same.

Listening to what nVidia has said in one interview about the 6800, what
I'm hearing is not that it looks any different, but that it has more
branching / decision making ability, and so makes it easier for the
developers, and in theory, allows the shader code to be smarter / faster.
So not necessarily better looking, but faster.

Of course then I would expect ATI to claim they can be just as fast with
PS2.0. Although I don't know if they can claim to have the same
flexiblity.

So is it a question of Image Quality? Speed? Or ease of use and
flexibility for the developer?

In the past, developers have definately had complaints about both
vendors, from driver issues to code paths. They've had to make
exceptions and write vendor specific paths. Most recently, it seemed ATI
was the easier card to develop for. They had buggy drivers but in the
end required less in the way of specific optimization for a given code
path. Maybe nVidia has learned and intends to pull off a win this time
around.

I'm still waiting for Doom 3 to come out before I buy a new system. The
wait is killing me.

On 29 Apr 2004 01:20:32 -0700, [email]nada_says@hotmail.com[/email] (Nada) wrote:

>> What does this mean? My guess is that FarCry uses a mix of PS1.1 & PS2
>> shaders for the GeforceFX, and these settings are being carried over
>> unchanged to the 6800u, while the Radeon has always used PS2.0 alone. If
>> true, present FarCry 6800u benchmarks are pretty much worthless, especially
>> with the reported 6800u artifacts.
>>
>> rms
>
>To me, the real question is, what exactly can be achieved with PS2.0
>and PS3.0 over PS1.1? I just didn't see a whole lot of differences
>between my old GeForce 4200 and the newer 5900XT? Will the fish
>shine? Do the birds change from fancy parrots to dirty seagulls when
>using an old card? If the comparisons done in Geneva two weeks ago
>were done in low quality PS1.1 against the highest quality PS3.0, then
>that's really the wrong way to market the next generation card. If
>the thugs have shining snot dribbling from the noses with PS3.0, maybe
>I don't really want to see it!

In the game like FarCry - My 9800 looks a lot nicer than my friends
Ti4200 (mine is on a shelf now, 5900 replaced it) - the PS and other
effects do come to play.

BTW: FarCry runs better with 1GB ram over 512mb on the PC.


--
Remember when real men used Real computers!?
When 512K of video RAM was a lot!

Death to Palladium & WPA!!

We are going to have to wait for both the Geforce 6800 and X800 to
start shipping and for DirectX9.0c to become available before we can
see some reputable benchmark results to know what the real answers
are. Even then, we have the Geforce 6800 Ultra coming out first,
followed by the Geforce 6800 Pro a few weeks later and the X800 coming
out first followed by the X800XT a month after that, so there is going
to be a time gap there before we have both sets of the highest end
cards to be compared with each other. ATI thinks they have the edge in
core and clock speed and they are saying that is going to make up for
some admitted feature ommissions and Nvidia is saying they have the
additional features that are going to make up for a slightly slower
core and clock speed. Right now, we have about two sites that have
done some sketchy benchmarks on pre-production Geforce 6800 cards with
basically beta drivers with beta DirectX9.0c with about 2 dozen sites
linking to those tests and making all kinds of conclusive assumptions
from those results. How can anybody come up with any meaningful
conclusions from that? It does seem like 3Dmark needs to get cooking
on 3Dmark04 and Aquamark needs an 04 update as well.

Wouldn't this effect be the result of Crytek's (the game developer's) doing, not NVIDIA's?