I'd like to get a good Benchmark discussion going...

For this we will need a good benchmarking tool and specific data. The result file for the X2 - The Threat rolling demo does not specify all settings used. For example, "Automatic Quality" is not specified in the results. You must state if you used this feature. It can be assumed that this dynamically increases and decreases complexity given frame rate. Therefore, this feature should be DISABLED for benchmarking purposes.

Is the X2 Running Demo truly a good Benchmark?

When the final game ships you will know the answer to this. My first impression is that the rendering engine does not scale well at all. Certain scenes cause significant stutter. The quality of the graphics are not terribly impressive. If you remove the "fake" sepculars you end up with what is very basic geometry and it is not very smooth looking.

I would only use this demo to make an assessment of performance for this particular title (see more below). Unless the engine is reworked or the models and/or complexity of the game is changed, I think we can assume that the rolling demo should give you some idea of the performance you can expect in the final product.

Can we really see how the newest cards stack up against each other with this?

That depends on what you want in a "new" card. The technology in X2 is based on Vertex Shader 1.0. Hardly new technology and the latest cards sport VS 2.0. It may be more fair to think of X2 benchmarking as a way to compare how the latest hardware runs legacy 3D code.

Well who has the fastest card? At least running this Demo...

I really do not think you should use X2- The Threat rolling demo to arrive at any conclusion of this. Only with HEAVY emphasis on "running THIS demo."

RIGHT. Now on to the results. Hold on to your chair because my Geforce 4 ti4600 just outperformed your Geforce FX 5800 Ultra in a MAJOR way.

Scene GeforceFX Geforce4 Geforce4/GeforceFX
5800 Ultra ti4600 Ratio
---------------------------------------------------------------------
0 199.59 167.149 0.84 *
1 34.373 71.063 2.07
2 20.038 51.362 2.56
3 11.749 38.212 3.25
4 18.925 27.79 1.47
5 17.335 25.975 1.50
6 8.962 15.561 1.74
7 28.945 27.93 0.96 *
8 15.354 57.418 3.74
9 12.954 45.787 3.53
10 27.574 41.344 1.50
11 20.182 35.133 1.74
12 16.446 35.619 2.17
13 18.615 34.784 1.87
14 42.424 30.515 0.72 *
15 21.856 47.91 2.19
16 26.714 32.336 1.21
17 35.483 46.699 1.32
18 19.239 52.381 2.72
19 6.611 15.289 2.31
20 23.686 33.506 1.41
21 18.244 39.737 2.18
22 18.231 36.071 1.98
23 26.693 56.11 2.10
24 19.132 47.406 2.48
Average 28.3742 44.52348 1.57

* Denotes instances where the GeforceFX 5800 Ultra outperformed Geforce4 ti4600.

Note that the GeforceFX only scores better in 3 scenes. In some scenes the GF4 outperforms the GF/FX by more than 300% (!!!!).

My scores (Geforce 4) were run with "Automatic Quality" disabled. If this feature is activated the scores go up (!!).

Cellar_Dweller, first of all your are using "software vertex shading". I think you need to check if you had FSAA/AA forced ON while testing. Something is definitely wrong with your results.

You would do well to check your own testing if you want to have a "good Benchmark discussion." However, I think this particular demo can be scratched as a good general benchmark.


Setting & Driver used:
System specific video settings:
Adapter name: NVIDIA GeForce4 Ti 4600 [NVIDIA GeForce4 Ti 4600]
Vertex Shader: Hardware, version 1
Hardware T&L: Yes
Fog capabilities: W-fog Z-fog RangeFog Table(Pixel)Fog VertexFog

Detonator 30.82
"Automatic Quality" DISABLED
"Antialiasing" DISABLED
"Bumpmapping" ENABLED
"Shadows" ENABLED


EDIT:

Can anyone tell me how to fix the table in my post? I had it nicely delimited and it shows as such in the editor. Very hard to read and misaligned in the header.

My point was to use this demo as the benchmark period and nothing else for my little discussion here...

Also to run with all settings on, and in benchmark mode and to see with similar configurations as far as the computer was concerned...

All video card settings in the driver itself set to nothing, no AA, no Anistropic... 1024x768 at 32bit depth...

So of course if you turn things off in this benchmark things go up...

Well obviously this wasn't as easy as I thought it would be... Just following the directions I gave and not digging into it at all and just running the demo with all Shadows, AA etc. that is found in the Demo only ON...

Then post results to compare my card to others... Not everyone had a 5800 Ultra so I wanted to see how it fared...

I have 5 systems one with a 5600 Ultra, 5800 Ultra, Geforce 4 Ti 4600, Geforce 2 Ultra, and a Matrox G200...

I only consider two of those cards comparable for what I was wanted to try to do with this discussion... NOT AN ARGUMENT or anything else...

Just the results from comparable systems with the settings set the same... Thank you...

Well I guess this is discussion over because it's screwed up already...

EDIT:

Can anyone tell me how to fix the table in my post? I had it nicely delimited and it shows as such in the editor. Very hard to read and misaligned in the header.

I think what you might want to try is a table with the data aligned in fixed width cells. There are a number of HTML tools out there that you can use to accomplish that or you can do it by hand (tedius).

For the purposes of this discussion (benchmarks), I've more or less stopped running benchmarks for the simple reason they DON'T always give an accurate indication of what a card is capable of during REAL world performance evaluations. Here's why ---

My son's system is Intel P4 based and he's running an NVidia Ti-4600. Mine is AMD XP based and I'm running an ATi 9700 Pro. My son's video card doesn't support DX9 functions and mine does ... so running any kind of benchmark that utilizes those functions is a waste of time (for comparison purposes).

That said, his real-world gaming performance absolutely kicks butt. Obviously it doesn't stomp a 9700 Pro but the differences (visual and perceptual) are so small as to be imperceivable. Only when you crank up the AA settings are the differences in the hardware evident. His P4 system is highly tuned (as is my AMD system) and optimized but that's not all there is to it.

DX8 games, of which comprises about 99.9% of what's out there at the moment, are not terribly difficult for today's AND yesterday's hardware to handle. It's only when you throw something like the Doom III alpha game (which my son's system won't even run) at today's hardware do we begin to realize what the POTENTIAL of tomorrow's hardware should be.

And too, most of the "glitter" associated with benchmarks has lost it's charm for me (since the NVidia driver debacle) and frankly, I just don't see the point any more. The only TRUE test of a graphics card is how it performs during GAME PLAY with YOUR favorite game. That's the only time you'll see any difference between one card and another. It's what YOU are used to seeing and expecting both visually and perceptually that allows you to justify upgrading to the next (EXPENSIVE) level of hardware.

The Ti-4600 in my son's system used to be in my system. When I switched to the 9700 Pro, I actually missed seeing a few games the way they looked via the Ti-4600 card. Does that mean my vision is impared? I don't think so. Does that mean the 9700 Pro sucks compared to the Ti-4600? Hardly. It just means that I got used to seeing things a certain way and then they changed. It took me a while to get used to the new "look" of the 9700 Pro.

My benchmark score went up, for what it's worth, like I really give a rat's ass. But what I MISSED was the way a few games looked ... and that has NOTHING to do with benchmarks and isn't really measureable.

Benchmarks are guides and measures for evaluating relative performance. Nothing more. They don't tell you how card A is going to look (in a certain game) compared to card B. Those are subjective qualities that only YOUR eyes can tell you. Don't forget that.

Later.

what exaclty is the x2 benchmark?
i have lots of benchmarks and have never heard of this one

Well I guess this is discussion over because it's screwed up already...

LOL ... next time be more specific about not wanting anyone with a mind of their own, a differing point of view or opinion, or anyone over the age of about 12 to add to the discussion (or lack thereof as you apparently want to stipulate). I can just see you as a little kid ... pounding your fists on the table and repeating over and over ... "I WON'T eat my peas! I WON'T eat my peas!!"

WAY later, dude.

Well I guess this is discussion over because it's screwed up already...

childish twat

He had it coming from the start. He's made that he spent so much on the 5800 POS and it didn't perform as well as it's worth. Ah well, that's marketing for ya.

I'm about 4 times older than that...

Oh and yes I did stomp my hands and fists...

Obviously people just can't follow directions???

If someone asks you a question you answer it you don't give the history of the world... NO ONE CARES...

And no I'm not MADE either or even MAD about my 5800 Ultra it's the best card I've ever had and it's fast as hell..

It was a simple test run bench mark, post results...

How hard it is to find good help these days???

gotta give him credit tho.....he tried.
just goes to show you that running a benchmark..especially these days with so many ways to put a system together...isnt the end all be all...that some guys insist they should be.