Slashdot Mirror
Zooming in on the GeForce 3
Couple of more in depth hands on looks at the GeForce 3 popped up lately including Sharky's
coverage (with DX8 screenshots) and
AnandTech's take. Same basic story. Good card, ahead of its time, overpriced, nothing will take advantage of its best features. I bet in 12 months we think differently.
Who's going to be the first to say the screenshots have been edited in Photoshop? (or gimp) :)
Baz
Actually by making it pretty easy to put "flashy" effects in, this card might allow the developer to spend more time working on the other parts of the game.
Perhaps, but not likely. Almost every effect is a special case hack that is designed to look good, at the expense of generality. This means that the engines for games (Quake 2/3 is a perfect example) become very specialized. If you want to base a second game off of the same engine, you get exactly the same set of effects as before, because the design is not very extensible.
My reference to simpler games of yesteryear goes farther though... try to get a group of people together for a netparty, for example... how many people do you think will be able to run stuff that requires a Geforce 1, much less a brand new Geforce 3? Second quesion is: why would you want to?
It seems like the technology trend is to push the envelope to the next step (GF3 is a logical progression from GF2), and then software has to play catchup (no, not ketchup :). It seems that, especially with the first few games for a technology, games focus on doing things because they "can", not because they "should".
Instead of designing games with a clear focus, plot, and motivation, games tend to get dilluted to being flashy silly things. Sure they look cool and have nifty features, but are they really FUN to play? Where is the replay value and interactivity with other human players?
The problem with, for example, vertex shaders and the other DX8 features is that they are not a simple extension of a uniform graphics architecture.
Instead of being able to designing a simple graphics engine that supports a variety of features, and then enhancing it as the game progresses, they have to "Target" a "technology" (for example DX7 or DX8). Of course, they two are essentially completely different APIs, and they are mutually exclusive. This means that you get to rewrite your engine every time you want to support a new "technology".
In a lot of ways, I'm really happy that cards (specifically the radeon and GF3) are moving towards programmability and higher quality visuals. It's quite a different race than pushing fillrates and making up stats on textured poly's a second. Maybe when things are fully generalized, we won't NEED an API like DX that gets constantly mutated, torturing developers...
Oh wait, that's called OpenGL. (sorry, couldn't help it. ;)
-Chris
With 128bpp, they may be talking about different buffers.
The high-end SGI workstations in '93 had an effective 140bpp video memory. (I thought I recalled 142, but this is from my hazy recollections.)
8+8+8bpp RGB front buffer.
8+8+8bpp RGB back buffer.
32bpp Z front buffer.
32bpp Z back buffer.
24bpp Windowing buffer.
4 bpp (rle compressed) per-pixel video mode selector.
I'd like to see more of that (plus today's dedicated memory for texture, vertex, transform, lightmaps, etc.)
As for color bit depth, 8bps (RGB 24bpp) is the most you'll see on most CRTs. You won't see 32bpp onscreen, usually the other 8 bits are just dword alignment for speed or an alpha channel for video source weaving.
However, the human eye is quite capable of seeing more colors in other situations; Hollywood typically does 16bps (RGB 48bpp) on their special effects, because they don't like to see 1"x1" jaggies or dithering on the 30' screen.
[
Transform and Lighting is nothing new - all 3D programs do transform - rotations, scales, translations, skews, projection, etc.
Take OpenGL as an example. The "T" (in "T&L") functions are glRotatef, glTranslatef, glScalef, and glMulMatrix. Before there was hardware T&L, people don't use these functions often - they write their own. And it was amazing that even a very simple unoptimized matrix transform code performs better than these gl functions most of the time.
What hardware T&L does (in terms of OpenGL) is to accelerate these functions in hardware - formerly, the OpenGL library does inefficient software transform. Now they'll just blast the arguments to some chip registers and let it do the rest. And it is fast, not only because it reduces bandwidth use (intra-chip communication is fast), but it also releases CPU cycles for other uses, which inevitably will have a positive impact on performance.
So, in short, if developers ditch their own matrix libraries and use the ones provided by the graphics API, they're already making use of hardware T&L. And, yes, unfortunately, hardware T&L only has things to do with frame rates - there's no other advantage than frame rate that hardware T&L provides.
Just remember - ALL effects are archeavable with software. The more you offload from the CPU to the GPU, the more CPU cycles you can save for physics, AI, and graphic effects that the hardware does not do yet. So, even hardware that "only" increases framerate sounds good enough for me.
Another thing to note is that alot of the really cool effects this thing can do are only noticeable in moving images. Bump mapping, for example, can be faked using regular texturing if nothing is moving. You really only see the difference when you see it in action.
------
Yes, it cost $500, but I paid over $2000 for a far-inferior 3D board just a few years ago.
It looks like Hercules will be the lead board vendor on this round. Creative is dropping out of graphics boards.
Carmack has written that all developers should get one of these boards as soon as possible. Gamers may want to wait.
Average eye cannot discern any difference between anything beyond 70 fps. And now that we have hit the plateau, its quite obvious that image quality is next. Even the GeForce2 Ultra hits a measly 25 fps when you run it on 32 bit, 1600 * 1200 splendour with 4x FSAA enabled. This is whats gonna keep the vendors busy for the next one year or more, bringing ultra realism to the graphics. Halo, Unreal II would all be capitalizing on that issue.
Gameplay would be next. AI would improve tremendously, storylines would improve, though sometimes you just dont care a f@#$ about the story and just wanna jump in and let that Chaingun rip.
My only worry would be that DirectX8 is fast becoming the API of choice among the developers (except Carmack who claims to only use OpenGL till kingdom come). And considering that Nvidia has now an unfair advantage over other cards, since they developed DirectX 8 along with M$, well my guess is as good as yours.
However a couple of weeks back, there was much stirring among the gamers when Kyro II kind of beat Ultra in Tile Based Rendering capabilities. I would welcome someone like that anytime.
Rapid Nirvana