Next 3DMark

Miscellaneous Forums/General Discussion/Next 3DMark

Informatic interview links if you want to know what can be done these days with sm3.0:


http://www.beyond3d.com/interviews/fm05/p2/

If you missed part 1, you can read it here:

http://www.beyond3d.com/interviews/fm05/p1/


We will introduce a plethora of new and improved effects in the next 3DMark. HDR is maybe one of the coolest new effects that we have done. We know that there are some games out already using HDR, but we wanted our HDR to “go to eleven” (one louder). We are using a very complex and advanced post-processing effect for our HDR rendering which makes it look absolutely amazing, and it really works like a charm. You should be able to see some of it in the screen shot we have provided for this interview. Of course we also use tone mapping, which is essential in order to make HDR look correct. But we didn’t stop there. We also wanted to create some other mind-blowing effects such as subsurface scattering effect in the shaders, HDR refraction & reflection water shaders with depth fog, heterogeneous fog, realistic sky model with cloud blending, Strauss lighting model etc. A lot of new things we haven’t had in previous 3DMarks. Of course we also made a lot of work with the shadows, so we will be introducing an even better approach to dynamic soft shadows for every object on screen than we have in 3DMark05.




The current implementation in the next 3DMark is using shadow maps, but we are using a different technique than we did in 3DMark05. In 3DMark05 we used something called LiSPSM (Light Space Perspective Shadow Maps), whereas in the next 3DMark we use something called CSM (Cascaded Shadow Maps). In the next 3DMark we use 5 pieces of 2048x2048 shadow maps (in 3DMark05 we only use 2). The two approaches are great, but CSM seems to be even better and more robust than LiSPSM.

To make those shadows smooth, we use 16 samples of a randomly rotated kernel for every pixel. I am not sure if anyone has done this before in any games/apps, but it works great and you really get smooth edged shadows for every pixel. Of course we would like to use even higher numbers of samples, but when we did some research in using higher amounts, it really killed the performance. 16 seem to be the sweet spot for today’s high-end graphics card, for which the next 3DMark is targeted. This smoothing technique wasn’t possible to do in SM2.0.



There's also nice collection of screenshots from the previous (and coming) 3DMarks:

http://www.beyond3d.com/interviews/fm05/p2/index.php?p=06





holy crap... that screenshot of the coming 3DMark is sweet..
I can't wait to see when actual games are of that quality..
when I come to think of it.. I just want to see games with the current quality that's in the current 3DMark.. :)

when I come to think of it.. I just want to see games with the current quality that's in the current 3DMark.. :)
Why? I have a pretty decent system, and the 3DMark scenes run somewhere between 10 and 20 fps. Add to that the fact that 3DMark can get away with "cheating", in places where a real game wouldn't be able to, I can't imagine they'd run very well at all.

looking forward to see it! but i guess it wont be too smooth on my aging 9700pro

Soon the difference between offline and realtime rendering will be nigh.

Wait... I'm sure I've read that somewhere before, Dejavu!

That screenie is pretty sweet. </obviousness>

I somehow think the next 3DMark won't run too smooth on my 7800GTX. That is kind of the point of a benchmark, to push the hardware and be ready for the next gen. It will probably run almost smooth on the next evolution of graphics cards, a single card solution that is as fast as SLI 7800's and has better quality rendering (HDR + anti-aliasing).

Why? I have a pretty decent system, and the 3DMark scenes run somewhere between 10 and 20 fps. Add to that the fact that 3DMark can get away with "cheating", in places where a real game wouldn't be able to, I can't imagine they'd run very well at all.

I meant that I'm looking forward to seeing these games, when the hardware is also capable of running it smoothly..

I call it 3D@#!*, I mean it's the worse excuse for a program that I can think of, it's got nothing to do with game gfx rendering, just a glorified penis size test on gfx hardware.

Yeah, that's what people with GeForce MX440's always say...


(jk)


just a glorified penis size test on gfx hardware.



Exactly, it's a benchmark. Just like F1 is a glorified penis size test for drivers and cars. Anyways, at least it's a chance for me to do something bit more advanced gfx than my Blitz3D tech stuff...

And in reality 3DMark tech IS what games are using, we just use it year or two before any games do - and of course our load is heavier than what games usually use... but just look at 3DMark03 or 05 and you can see that they used stuff that games now use too, and for example 03 goes pretty fast with todays hw although it hardly did when we published it! :)

It also kind of shapes the optimisations of tommorrow: cards are built that strive to get the best framerates on carmacks and futuremark's stuff.

So in a sense, I would say the work mustang does has an impact on the hardware I will be purchasing in the future.

Mustang, work in more 2D stuff ;] nVidia's 2D support is going fast downhill in 3D mode ;]


Mustang, work in more 2D stuff ;] nVidia's 2D support is going fast downhill in 3D mode ;]



And soon we have Windows Vista and EVERYTHING is 3D... :) 2D is dying. Just yesterday we thought how cool all those side scroller games were...

Side scrollers are still better.

Sure, but most of them coming out now use 3D hardware. "better" has nothing to do with the rendering.

Like it or not, 2D is out. The latest shootemups for sale on the pc (side scrolling or vertical scrolling) are 3D in rendering.

If you don't know what I'm talking about then I suggest you look up jets n guns.

But I'm a whiny old dinosaur. =]

me too... but i know that the big one is headed my way... and i aint talkin about king kong...

vista may very well be the extinction event that causes all of us dinosaurs to eventually dissapear...

... and it may take a few of those mammals along with it

we must be able to adapt... or we will become extinct...


--Mike

Extinguish me then. At least I'll die happy. =]

No, wait, I'm whiny and miserable. Dammit.

Is there anything, some kind of effect, that can be done in 2D that cannot be done in 3D?

Complete destruction of underlying pixels. =]

True. Imagine trying to do Spheres of Chaos in 3D. :O

Err, Spheres of Chaos can be done in 3D. Well, using 3D rendering anyway and just keeping it a 2D game.

One of my projects is a SOC clone. Not being done with 3D rendering though, it has to be said. Though, the idea does have some merit.

Err, Spheres of Chaos can be done in 3D. Well, using 3D rendering anyway and just keeping it a 2D game.
You mean you can do the 1-pixel-per-particle effect in 3D? There's thousands of them. I'll believe it when I see it. :)

I plan to do it in BMax - would be easier if I could kill off those particles, drawing them on the landscape gfx when I wanted, but thanks to lack of Imagebuffers, I either have to keep them alive (>|<indefinately) or lower the spray count.

big10p: My ultimate plan (apart from world domination, still working on that) is for my SOC clone to work fine from 640x480 to 1920x1200. Adapt to arbitrary aspect ratios, have as many particles as possible on screen without dropping frames and include as many psychedelic effects as possible.

At 640x480, particles will have to be 1 pixel only. At 1920x1200, each particle will have to be almost 3x3 pixels to be the same size. That's where the 3D hardware can help out. Doing particles by pixel plotting or 2D image drawing is a heck of a lot slower than asking the 3D hardware to update all the sprites (or whatever system might be used).

As I said though, for now I'll just stick with 2D, so that I can at least get the thing working before I start adding bells and whistles.