Previous
NVIDIA, CUDA and PhysX Article
PC Article by Kieron Gillen
Page 1 of 2.
Page 2 
3D card manufacturers shouldn't take this the wrong way, but it takes a lot to make us crawl out of the communal Eurogamer bed (yes, all the Eurogamer writers share a single large bed - we do it for frugality and communality, which remain our watchwords) and go to a hardware presentation. There's a nagging fear someone may talk maths at us and we'd come home clutching the local equivalent of magic beans. And then we'll be laughed at by our fellow writers and made to sleep in the chilly end where the covers are thin and Tom left dubious stains. That's no fun at all.
Then again, there's some things you can't help but go and have a gawk at. So when an invite claims, "All too often new hardware brings with it a small performance increase - maybe a 5-10 percent over the previous fastest thing. Wouldn't it be far more exciting to see a speed increase of x20 or even x100... well, we'll be happy to show just that on Friday," you have wander along. Even though you suspect it may be a trap and they're going to attack you with ill-shaped blades, you have to find out what on earth they're talking about.
As we suspected, it wasn't quite what we were hoping for. Sure, there are programs which gain a x100 increase via the methods NVIDIA talks about on this particular Friday, but unless you're working in economics or astrophysics modelling, it's not exactly that relevant. However, something more quietly astounding was explained. Mainly, that despite the fact that no-one you know bought a PhysX card, if you're a PC gamer with a relatively recent NVIDIA card, you've already got one. Or, at least, you will soon. Spooks.
Get him!
The primary idea NVIDIA was trying to push was Optimised PC - the approach discussed in Rob Fahey's interview with Roy Taylor the other day. The idea being that the traditional PC approach where you buy the fastest PC processor you can doesn't actually lend the best results, at least in most situations. If you spent more on - predictably - a GPU-driven 3D card, for an increasing number of areas, you're going to get much higher performance. If the program is using the GPU in a meaningful way, anyway. NVIDIA highlights areas like image-processing and HD video-encoding, as well as - natch! - games. You lose in single-threaded activities - like, say, just booting up a program - but they argue a small loss in opening a Word Document is less noticeable than frames in games or similar.
Where it starts getting interesting is NVIDIA's development language, CUDA. The problem with all the threading programming methods is that it's radically different to single-threading (and, yes, we're getting into, "Why would anyone care about this but a programmer?" territory, but its background for the key point later). It's hard to do, and CUDA is basically a way to make things more accessible.
NVIDIA claims anyone experienced in C or C++ will be able to get a grip on it (i.e. not us, but the aforementioned programmers). This means that anyone who codes in CUDA can program the GPU to do pretty much whatever they like; it's by turning the 3D card into a bank of processors that the financial analysts and the astrophysics guys are getting such impressive results. And impressive savings, as it's a lot cheaper to do it this way.
Want to comment on this article? Log in, or register!
Comments: 1-23 of 23 in total
|
0
This sounds pretty good.. except for the fact that I've never bought a GPU that has had any power to spare for physics 6 months after purchase. You'll still need two cards. The only difference is that now you don't have to buy that old PhysX card, you can just buy TWO Nvidia cards, which is exactly what they want.
|
|
john_silence
20/06/08 @ 09:26
0
+1, mrkreku!
Also, that CUDA thing, it does work with Havok too, right, since it's only a matter of coding, not hardware? Apparently ATI worked with Havok - in spite of the fact Havok was bought by Intel a while ago: ATI devoted part of the shader units on the new Radeon 4800's (I'm plugging a 4850 into my computer this afternoon!) to process physics, using the Havok engine. A bit the same approach, but already integrated into the hardware. Which would make this old news of a sort - not that it's uninteresting, but maybe too one-sided, a bit of research on ATI's new architecture is missing from the article, perhaps ;) Anyway Nvidia have missed the train on the next generation - it will be my first ATI card after 3 successive Nvidia ones. The new GTX was already labelled "too hot, to late" and called a "dinosaur" in the first reviews that cropped up. It's interesting to hear they're upgrading already existing material like the great 8800GT's, however. It goes to show what could be made of all these GPU's if their power was properly tapped, and how far ahead of consoles they could take PC gaming.
Edited 1 times,
most recently on 20/06/08 @ 10:27
|
|
0
This is why I hate that "Just upgrade" mentality of the PC gamers. It's nonsense as no-one is using the existing GPUs to their fullest anyway. The new Geforce 280 and 260 GTX units are supposed to be upto 3 times faster than the existing 9800 but, give almost identical benchmarks because they can't be fed data fast enough by the CPU. This is exactly the same situation with the new ATI cards.
There is simply no reason to buy them. You'd need to load up the cards with fast memory and then shift all game geometry and textures to the card at the start of the level. Then you only need to be telling the card where the camera is looking and where the player has moved to. Minimal geometry changes per frame in other words. |
|
0
CUDA is amazing. In GIS, Manifold supports CUDA and processing of large satellite imagery has dropped from about 2 minutes per tile about 3 seconds. That's quite a leap. It would be cool for 3d max and maya pick up on this - i'm pretty sure they don't support this yet.
And from what i know, financial analysts rarely use anything other than Excel & VBA. Bless. |
|
0
I don't understand the maths or the coding stuff, but anything that improves gaming on my exisiting 2x8800GT SLI setup will be very welcome indeed.
It's a very good move from nVidia to offer this on 8-series cards rather than forcing everyone to upgrade. Having the user base out there from day one gives this a good chance of success. I can't really see that many people buying the new 260/280 GTX cards until the price has fallen substantially. |
|
0
@penhalion - According to Guru3D who benchmarked NVIDIA's new cards, the new GeForce 280 card is slower in almost all the games they tested than the GeForce 9800 GX2. Granted the latter is a SLI model based on the GeForce 8 series on one convenient card so it only uses a single slow but it is also cheaper. The new NVIDIA cards don't support DX 10.1 either. I was actually quite disappointed when I read the results; not quite what I expected for such an expensive card based on newer technology than the GeForce 8 series.
|
|
0
Speak of the devil: ATI released the 4850 and it's good!
http://www.anandtech.com/video/showdoc.a... Now wait for the 4870! Woo, I'll finally get some use out of my Crossfire motherboard :D |
|
0
How is this different from ATI Assymetric Physics processing?
http://ati.amd.com/technology/crossfire/... You'll be able to buy a new card and dedicate whichever you want to do phys. proccessing. And it's supposed to be supported for all cards from x1600 and up. |
|
subtlesnake
22/06/08 @ 16:30
0
"This is why I hate that "Just upgrade" mentality of the PC gamers. It's nonsense as no-one is using the existing GPUs to their fullest anyway. The new Geforce 280 and 260 GTX units are supposed to be upto 3 times faster than the existing 9800 but, give almost identical benchmarks because they can't be fed data fast enough by the CPU. This is exactly the same situation with the new ATI cards."
The cards are only CPU limited at low resolutions, and extremely high framerates when the CPU isn't fast enough to run the game at 200 FPS, or whatever the GPU would deliver were it not CPU limited. At 1600*1200 or greater though, the GTX 280 gives an extra 50% - 80% more performance, compared with the 9800 GTX - which isn't bad scaling given the GTX 280 only has 90% more math power and 25% more texturing power to start with. Obviously you're only going to need one of the cards if you run at high resolutions - because you have a large TFT say - but that's the way it always is in the beginning, and games like Crysis or Assassins' Creed or World in Conflict/Lost Planet DX 10 are in need of the extra horsepower. |
Comments: 1-23 of 23 in total
Want to comment on this article? Log in, or register!
