What is Digg?23 Comments
- AkatsukiNoTobi, on 10/12/2007, -2/+14Well the Ageia PhysX card is currently a complete waste of money. Not only do a very limited amount of games support that card, but the performance boosts are not worth the +$200 price tag. Quad cores, once optimized, will completely disregard the need for a physics card; this also is the same for dual core processors.
Not only that, but AMD/ATI allow a third GPU with Crossfire to run physics calculations. This third GPU is much faster with the physics calculations and also works with all games. Plus an X1600 Pro costs less than half the price of one of the Ageia PhysX cards.
So unless Ageia prices the card at less than $100 these cards will make a high quality paperweight. - manageMyRights, on 10/12/2007, -0/+7Theres a technical difference between what some might consider 'supercomputing' (big metal) and GNU/Linux clusters.
Most clusters are connected with gigabit switches and each node has its own local storage. This is great for stuff thats 'ridiculously parallel' like looking at different sections of the genome, but you have actually split up the jobs, send them to each node, and collect the results as they finish. There are some great programs that make this easier but its time consuming.
Problems like simulating molecular dynamics requires lots of communication between processors and therefore clusters fail. - radu79, on 10/12/2007, -3/+9That article didn't make any sense whatsoever, and it's purely based on wild speculations.
Besides, AMD and Intel are already into the multicore battle. Intel announcing something like 80 cores in a few years (in production, not just a prototype). - NikoKun, on 10/12/2007, -0/+6Personally... I think if PhysX cards want to have a future... they seriously need to think of other ways to do it... because as an Optional card, most people would skip it, and all the games have to have that feature as optional.
What they SHOULD do... is merge PhysX cards with Graphics cards. Make it all part of 1 card, call it a gaming graphics card or something like that... make it easier for us to buy and figure out... Otherwise, PhysX, although a good idea, wont be much of a deal. - inactive, on 10/12/2007, -1/+6True, I also think PhysX cards won't have much of a future. PhysX cards will be like DVD decoder cards back then. It was beneficial to have it, but eventually when computers got faster, there's no use for it anymore.
- Derrekito, on 10/12/2007, -1/+6With enough cores, will it matter in the long run? Don't get me wrong I love the idea of having a dedicated chip just for phys, but economy wise I raise the question. It's hard enough to drop somewhere around $ 300-400 per card (SLI is a nice feature) then what, drop another butt load of money for a phys? It seems more like multi core is the way of the future, but hey remember RAMBUS?
- trubbleshute, on 10/12/2007, -0/+4Supercomputing, for gaming at least, is dependent upon real time physics and I think we are pretty far away from it right now.
Realistic graphics are here, look at the shots of Crysis. I don't know if a devoted physics processor will be part of the wave of the future, but it will raise awareness. - acdcfanbill, on 10/12/2007, -1/+5generalized computing pipelines for specific gfx/phys seem to be the wave of the future.
- acceptab1euname, on 10/12/2007, -2/+5Call me when I can blast big holes out of the walls like I did in Red Faction. I still can't believe that in games like Battlefield 2, you can shoot a WOODEN fence with a *&#@!^ TANK and the fence will fare just fine.
- Topher06, on 10/12/2007, -0/+3Yeah, so far physics add-on cards are vaporware. While the product might actually exist, what games really deliver an experience to take advantage of it.
I think the future of gaming is in physics personally. I keep talking about the game Little Big Planet because, well, the game rocks. And why it does is the fact that fabric in the game moves like fabric. Pieces and parts move like they should in real life, and the level of interactivity is amazing. All this in a simple sidescroller.
I am not seeing this level of innovation in PC gaming. While PC gaming might boast high frame rates with high resolutions, I get tired of games filled with static scenes, where rocks on the ground don't move, you shoot a window and a little bullet hole is displayed, but the glass doesn't shatter. Or the fact that the game seems lifeless because nothing moves around you, everything is static and still except for the elements the developers bothered to program to move.
Once physics is implemented in a game, simply adding wind to a scene should blow up dust clouds randomly, foot prints get left in the mud, water splashes realistically as you shoot or run through it, windows shatter, concrete explodes into many pieces, laundry hanging on the line will blow in the wind, or get ripped to shreds as they are riddled with bullets. The bottom line is that games will appear more realistic when the subtle things start behaving realistically.
Physics cards are currently underutilized, but by the time developers start taking advantage of them, computer technology will leapfrog their necessity. Intel proposed a viable 80 core CPU, nVidia actually released a 128 core GPU. While these are not consumer level products, how many months will it be before we see 16 - 32 core CPU's or GPU's? Heck, when you can dedicate an entire CPU core to wind physics, one to dust, one to glass, etc, there won't be a need for a dedicated PPU.
About all Ageia could hope for is to develop a software API they can license for use in the gaming market, but I think buying their add-on card is a waste. - j_bellone, on 10/12/2007, -0/+2Call me when the game developers take notice of this technology. Seriously. If Aegia wants to be taken seriously they should be lobbying the likes of Valve, Epic and id to get their technology included in the developer builds of those game engines. Those engines make up a massive number of games on the market currently and will be making up the number of games in the future. This card, even though targeted at gamers in general, will most likely fall onto the FPS gamers laps. Another reason to lobby Valve, Epic and id. When I start seeing AAA titles that support this card I will consider buying it. Until then they are merely another fish in the pond offering "better performance."
Better at a marginal cost for a marginal (less than marginal) number of titles. They should have pushed hard to get their technology inside one of the consoles. Hopefully next time around (if they are still around) they will do so. The only way to push new technology is to get the people designing the technology to look at your hardware. The consumer is not going to buy something that *might* work with games in the future. They should be offering their SDK, examples and what not for FREE to developers. They should be going to developers and giving demonstrations. I don't want to hear this ***** about it being like CELL. Its nothing like CELL. - diggfinity, on 10/12/2007, -0/+2I got one when they first came out only because I was ordering a PC on someone else's dime and figured "hey, what the hell...maybe it'll actually come in handy"... I'm still waiting to use it.
- Starways, on 10/12/2007, -3/+5Intel also said we'd have 10 Ghz chips in 2010. That's not going to happen.
- smedstadc, on 10/12/2007, -0/+2Three cheers for the well informed minority!
- RandomSkratch, on 10/12/2007, -0/+2@acceptab1euname
That's called game design, not limitation due to no extra processing. The reason why they did not include this type of thing in the game is because they designed it for multiplayer (online). Adding in real time environmental destruction (that isn't pre-canned) takes a lot of processing AND this also has to be sent to everyone else on the server.
Now they are attempting to do such a thing with Bad Company (or whatever it's called) but we've yet to see actual game footage. I am going to presume it will use pre-canned destruction. I.e. create full building along with other sub models that include several layers of destruction. - j_bellone, on 10/12/2007, -0/+1DVD decoders were beneficial until the graphics cards vendors started throwing them on the GPU. Deja Vu. Except this time around Aegia is only on a handful of games. We *need* this technology. But we *need* it to be mainstreamed. I for sure am not buying a card that could be obsoleted by the next generation of nVIDIA's or ATI's boards. Especially when I don't see any games that I am going to play that get a significant upgrade from the card. Not to mention the card costs just as much as another GIG of DDR2. Which would give me better performance overall.
- inactive, on 10/12/2007, -1/+2Don't be silly. Nothing useful works for linux
/sarc - kevptim, on 10/12/2007, -1/+1Crysis
- noahhoward, on 10/12/2007, -10/+7I don't know, right now the future of supercomputing is Linux clusters. I know a few companies and I think even the Army uses them.
- SultanTravi, on 10/12/2007, -7/+3Get some cathodes in that thing, and the glow will attract hawt chix like moths to a flame.
- inactive, on 10/12/2007, -11/+4I'm buying one of these so I can say my computer isa SUPACOMPUTA!!!!!!1
The chicks will be stickin like white on rice. - PepeGSay, on 10/12/2007, -7/+010ghz by 2010? hmm.... Top end Quad Cores run at 2.66 ghz. Which is just over 10ghz on a single chip and is more efficient.
- Ratteler, on 10/12/2007, -13/+5Call me when it works for Linux.
Check out the new & improved 
© Digg Inc. 2009
— Content created and posted by Digg users is