Doom3 graphics vs. Battle of Proxycon (GT2)

[Neutrality]

New design.

-Neutrality-

 

[Nick[FM]]

BTW, is 3DMark03 engine a new design or based on MAX-FX?

Here's a quote from our 3DMark03 WP:

3D Engine. Previous versions of 3DMark used the MAX-FX 3D engine. The trend in 3D engines is moving towards very lightweight DirectX wrappers as more and more work is transitioned to the API and graphics card. For 3DMark03, we have based the benchmark directly on top of the DirectX 9.0 3D platform. This allows the benchmark to be independent of specific technologies embedded in a single 3D engine implementation. We have developed a set of lightweight wrapper routines on top of DirectX to aid code re-use.

There!

 

[Katsa]

Sorry, I should have been more specific. I was refering to the comment and discussion around here that 3dMark extrudes all vertices on a model rather than just the edges to create a shadow volume.

Looking at Doom3, and looking at 3DMark, I think the 3DMark characters have higher polygon count, making the situation a bit more complicated so that a direct comparison of methods may not be fully thruthful.

If Doom3 has low poly (relatively speaking, of course) characters, silhouette finding with CPU may make more sense. You do need to transfer the vertices over the AGP bus also, so there is a poly count after this method is not efficient. Further, the extrusion could be considered to be a more straightforward method.

Don't all current NVidia shadow volume demos use extruding vertices to create the shadow volume, and none use the Doom3 technique?

 

[Jare]

Just a question... What about memory consumption? FM says that GT2 consumes way more than 64Mb of video mem (don't know about AGP memory), while I don't think Doom3 goes that high. Are the speed differences so acute on 128Mb vram cards?

 

[Saem]

the 3D strippers might come in "handy"

Yeah, DoAXXX.

 

[Reverend]

My first real post about 3DMark03!

DOF and "bloom" via PS_1_1 in 3DMark = messy blur = not good

I have had a few runs of 3DMark03 and never once was I impressed by such post-processing effects, try as hard as I might. It's more of blurry CGI-like cutscenes more than anything else.

I am, most of all, disappointed in 3DMark03's choice/implementation of showing off PS_2_0.

BUT every one of the Game Tests still "look good". Personally, I was most impressed with Troll's Lair.

As an aside, I'm conducting an interview with Markus Maki (some of you may know who he is). While I definitely stayed away from talking about 3DMark/Futuremark (he's not as involved with 3DMark as before), there are questions there regarding the relationship between/amongst benchmarks, games and video cards that, hopefully, will prompt some interesting answers from Markus.

 

[K.I.L.E.R]

I somehow think that 3Drealms are shooting for the "most busy detail" award in Duke Nukem 4 Ever. The result shall be seen someday.

Yes we will all be in wheel chairs by then.

Don't you mean dead?