TR Forums

http://www.xbitlabs.com/news/multimedia ... 34045.html

Holy ****, thats more than my graphics card cost me!

That's just stupid! No way am I going to pay this much for something that only 3 or 4 games will use. GFX: you know what you're getting, this just doesn't seem to make any sense and the games will have to make some gameplay feachure which uses this and is fun. Watching objects fall or water simulations isn't worth $250-$300!

I see still birth at those prices!

Will games designed for PhysX work on a computer without a PhysX card??
They bloody well better

mikehodges2 wrote:

Will games designed for PhysX work on a computer without a PhysX card??
They bloody well better

Yeah they'll work. The API this card recognizes, NovoDeX, is actually used by a few games right now, including Half Life 2. So with the card, the oly difference you'd see is something like, if you spawned like 4000 dudes, and then surrounded them with toxic barrels, then blew them all to ****, your cpu would grind to halt, displaying them at like 2 fps. With the physics card, this would happen in real time. **** like that. Without the card, this is done in software on the cpu. So initially theres no change.....but in future games, things will be writte to have more realistic water, for example, that actually flows around objects realistically, etc, or buildings that blow up and spray the wood/glass/stone, etc out in the correct directions, etc. So it'll be a while before we see games that are truly real-physics based, but adoption of this card will make it easier. Of course, I ain't paying that for one...theyd need to be around $100 for me to consider it...

/me thinks about Gary's Mod using this card...

Pax-UX wrote:

That's just stupid! No way am I going to pay this much for something that only 3 or 4 games will use.

I wonder if the gaming market would be what it is today if everyone said that about the 3dfx Voodoo when it came out.. oh, wait, didn't it cost more than that?

danazar wrote:

I wonder if the gaming market would be what it is today if everyone said that about the 3dfx Voodoo when it came out.. oh, wait, didn't it cost more than that?

We've been over this and over this every time there's a news post about this thing. The two situations aren't at all analogous.

UberGerbil wrote:

danazar wrote:

I wonder if the gaming market would be what it is today if everyone said that about the 3dfx Voodoo when it came out.. oh, wait, didn't it cost more than that?

We've been over this and over this every time there's a news post about this thing. The two situations aren't at all analogous.

Sure they are....it's a new product that will supposedly "revolutionize games." Noone wants to get their feet wet until they see what it can do and is cost effective. So we shall wait and see. Like I said before, EVERYONE will have one of these in like 3 years, trust me.

I wrote this in the chat:

Too much.

I don't think this will have the impact on gaming that 3d cards did. It might have about the same effect, or maybe a bit more, than 3d positional audio had.

Price it at between $100 and $150, it might sell. Build it into motherboards or video cards for a slight additional cost, it will sell. Hell, have Creative buy it and add it to their sound cards, it will sell. But at $250 to $300, they will gather dust on the shelf with all but the hardest of hardcore gamers ignoring it.

UberGerbil wrote:

danazar wrote:

I wonder if the gaming market would be what it is today if everyone said that about the 3dfx Voodoo when it came out.. oh, wait, didn't it cost more than that?

We've been over this and over this every time there's a news post about this thing. The two situations aren't at all analogous.

I agree. The difference between 2d sprites and 3d was enourmous. The difference between 3d rendered by the cpu and 3d rendered by a video card was enormous.

This is a lot more subtle. Additionally, while games may be written to take advantage of some of the benefits, most games are still going to be limited to the lowest common hardware denominator. So what that means is that while a regular wall might blow into 50 or so chunks when struck with a missile, a computer with this card may get hundreds of chunks, but other than the sheer number, nothing will be new. It would be very, very difficult to harness the power of this card fully without writing off the largest section of the market.

The true benefit of this card will be to totally rewrite the rules of physics in games. This would allow companies to design completely destructable (or changeable) environments. But when 95% of the market isn't going to spring for the $250-$300 for the card, then the designers will have to write for that lowest common denominator. The lowest commond denominator will not have the ability to use totally changeable environments. So what you get is more fragments, but no real improvement in physics.

It's difficult for me to articulate this. Everyone is expecting this to be revolutionary for games. I argue that until the part reaches a fairly high level of market penetration, there will be no real fundamental change in games. This creates negative market feedback because if there is no really fundamental change in games, then no one will buy the cards at the proposed pricing levels. But if that technology is built in at little additional cost, then it has the chance to be ubiquitous. If that were to happen, then you would see a fundamental change in gaming.

I don't know what the licensing model is for this either. If they are licensing at both the manufacturer's and at the game designer, then there may be some resistance to using this technology unless the terms are extremely favorable. MS is likely to come in with their own API and hardware spec if the price is too high.

I'm a fan of this technology, but I don't agree with their business model.

As big as 3D acceleration? Probably not...

The next big thing? IMO, quite possibly...

People love physics (in games) the more complicated the better... HL2 was just the tip of the iceberg...

Or think of it this way... Right now your CPU is doing all the math, with this card say it suddenly has some free time on it's hands... What might that do to your framerate? Suddenly your $250 investment helped your next gen SLI setup (or whatever highend GPU you've got) gets it's 75% performace boost without requiring you to turn on insane levels of AA and AF to justify what you spent on it... And the physics in the game of choice are that much more impressive... I could go for that...

Maybe you want your new dual core CPU to have it's new found power eaten up by fancy physics (when games go multithreaded) but I'd rather one handles AI and whatever else it does while the other takes care of figuring out what I just told my virtual squadmates to do and translates it into commands they care understand, or whatever else developers decide they want to use it for.

Of course the thing could be crap, but I say we wait an see before we bash it...

I thought half-life 2 used Havok?

Anyway, it looks like they're basically giving away the API to as many developers as possible and hoping to make all their money by selling the accelerator cards, which is interesting. A lot will depend how flexible the API is and how well it's accelerated.

madlemming wrote:

I thought half-life 2 used Havok?

Just using it as an example...

drfish wrote:

As big as 3D acceleration? Probably not...

The next big thing? IMO, quite possibly...

People love physics (in games) the more complicated the better... HL2 was just the tip of the iceberg...

Or think of it this way... Right now your CPU is doing all the math, with this card say it suddenly has some free time on it's hands... What might that do to your framerate? Suddenly your $250 investment helped your next gen SLI setup (or whatever highend GPU you've got) gets it's 75% performace boost without requiring you to turn on insane levels of AA and AF to justify what you spent on it... And the physics in the game of choice are that much more impressive... I could go for that...

Let me clarify: What you are proposing is merely shifting computations from one board to another. That's fine, but when only 5% of your market has the device, then all the programmers will be able to do is give you more particles. They will not be able to fundamentally change the physics of the game because people without the card will not be able to play it. Now if your base is 90% of the market, you can make REAL changes to the physics, like completely destructable environments, because you'd only be giving up 10% of your market versus 95%.

Until this technology reaches a large portion of the market, we're only talking incremental changes in gaming, not something worth $300 to the average gamer.

I agree with Thresher, and I think you're hitting the nail on the head when you talk about destructible environments (I used to call it "deformable objects", but it's all good). These cards will make destructible environments possible, and I believe they were developed with exactly that in mind. However, game developers aren't going to code for such complex physics unless there is demand. That means that the physics hardware needs to be widely available and generally affordable.

Keep in mind that we're still very early in this process. The card isn't even available yet. Just because it debuts with a $250 price tag doesn't mean it will stay there indefinitely.

Well I read somewhere that AGEIA's physics middleware is being developed for PS3... Is it quite possible that in the future, console's will have better physics in games than 'lowest-denominator' PC-games?

Let me try another one...

Forget the fact that this thing can do wiz-bang stuff, only think about physics in current titles...

Say that physics in a given game uses 25% of the CPU (low? high?) if you can take the 25% off your CPU and have it done elsewhere it's like getting a faster CPU since it can now do other things...

So you got two systems then, both with the same über GPUs, one has a 3000+ in it, the other a 4000+ a cost difference of $350... Could you theoretically spend $200 on a physics card and get the same performance out of both systems but save $150 doing so? Maybe. Could the system with the physics card do things the one without can't? Absolutely!

Maybe you don't need super crazy object counts to make this thing worthwhile... Mainstream and highend can both benifit?

Call me crazy, but I just like having options...

drfish wrote:

madlemming wrote:

I thought half-life 2 used Havok?

Just using it as an example...

I may be wrong about that....maybe HL2 uses HAvok, and it was Ut2007 that uses Novodex....in any event something that';s out now or will be out soon already uses it.

Someone from AGEIA wrote:

Ageia's PPU is an important step in the right direction as it can take the current limit of 30-40 bodies of today's high-end CPUs to a maximum of 40,000. One can truly see, when looking at those kind of numbers, just why a PPU is the right thing for gaming and should be a technology we all keep an eye on.

They're gonna be optimised for physics processing, the way that gpu's are optimised for gfx rendering. If you run 3dMark05, you'll notice the CPU can only pull 2 - 6 (ish) fps...whereas a GPU can do nearer 30.

I think in few years we will actually have realistic acting water (not photo-realistic though) - HL2's water looks good, but walking through it doesnt create ripples, and the edges are straight lines at best

When an anticipated title comes out (HL3 perhaps?) and the water is near perfect, when using a PPU, won't you get one of these? I probably will...but then again, i'm not cheap.

drfish wrote:

Let me try another one...

Forget the fact that this thing can do wiz-bang stuff, only think about physics in current titles...

Say that physics in a given game uses 25% of the CPU (low? high?) if you can take the 25% off your CPU and have it done elsewhere it's like getting a faster CPU since it can now do other things...

So you got two systems then, both with the same über GPUs, one has a 3000+ in it, the other a 4000+ a cost difference of $350... Could you theoretically spend $200 on a physics card and get the same performance out of both systems but save $150 doing so? Maybe. Could the system with the physics card do things the one without can't? Absolutely!

Maybe you don't need super crazy object counts to make this thing worthwhile... Mainstream and highend can both benifit?

Call me crazy, but I just like having options...

Most people aren't in a situation where they would benefit by having that 25% offloaded elsewhere. A very small segment, specifically FPS games, could benefit from it, but that's a very small percentage of a very small percentage. I don't have one of these cards, yet my old AMD XP2500 with a Radeon 9800 could run anything I threw at it. And it did this pretty darn well.

Very few game design companies will design their game solely for the leading edge of technology. But this card really doesn't add $250+ worth of stuff for the average gamer until games are designed specifically for it. At best, it's an incremental change. It may increase some people's framerates a bit, but the average gamer just isn't going to spend that much on an incremental change, that's good money towards a faster computer.

Think of it this way:

For $300 you could buy a new nVidia 6800 (almost enough for a GT!), you could buy a 17" LCD (or 19" if if you catch Dell at the right time), 2 gigs of memory, 2 Raptor 32GB drives, 3 Soundblaster Audigy 2 ZS audio cards, several DVD burners, or a really nice set of speakers, all of which would add more value to the average gamer than this card. Add to that prospect of multicore processing and the card becomes even less of a value since code could be used to take advantage of that second core just like it would this coprocessor.

Now if games were written specifically to the advantages of this card, I would turn that argument around. But because market saturation will be very limited at this price, that situation just isn't going to happen.

Thresher wrote:

Most people aren't in a situation where they would benefit by having that 25% offloaded elsewhere. A very small segment, specifically FPS games, could benefit from it, but that's a very small percentage of a very small percentage. I don't have one of these cards, yet my old AMD XP2500 with a Radeon 9800 could run anything I threw at it. And it did this pretty darn well.

I'm not trying to sway your opinion (well, ok, maybe I am), just shooting from the hip and tossing out ideas... I'm not ready to go out and buy one of these things myself... But let me suggest one more thing, low end systems may have more to gain...

If physics in a game takes 25% of a A64 4000+ how much more does it take on a lesser system? How do physics calculations scale? I don't know... Maybe instead of jumping ship on your Athlon XP system (new mobo, CPU = $$$) you can snag a 6800GT and one of these cards? Then take it with you when you go dual core or whatever... Far fetched? Could be, but like I said, I like options...

I hope this card takes off but I agree at this price point it probably won't. Unfortunately, most gamers put more of a premium on visuals rather than gameplay and this card can have a huge effect on gameplay once games start being written around it.

It'd be cool to take down a building with a few well placed bombs on the right support structures.

I take it none of you are reading my previous post. Am i completely wrong??

Once games that fully utilise the PPU come out, most "serious" gamers will buy one as eagerly as you would a GPU. Wait and see...

So far the only killer app is UT2k7! So until that comes I can't see much happing for the card... how are they going to demo it? Show you their home made movies which aren't real games? Or games that have the tech? In that case what games?

The success of this card will be linked to initally Unreal 3 and UT2k7. But when these games come out it will be time to spend $500 on new GFX card... then if you have some cash maybe a PPU. But it would be like $800 for a hardcore game for just one game!

My money is currently on the new consoles which look like, dollar for dollar, they are going to kick ass compaired to a PC.

This is stupid. Another couple hundred bucks for games.

This product will fail.

If this product does not fail, by some sick twist of fate, and instead gets very commonly used in modern games. I will give up PC gaming altogether.

Zenith wrote:

This is stupid. Another couple hundred bucks for games.

This product will fail.

If this product does not fail, by some sick twist of fate, and instead gets very commonly used in modern games. I will give up PC gaming altogether.

Miserable sod. Is that what you want? The first DVD burners that came out cost, what, $1000? Now you can pick one up for $50. You seem like a classic geek to hate a piece of hardware already...when it isn't even available yet, there are no demo's, and you have NO idea what these things will be capable of.

Lucky Jack Aubrey wrote:

Keep in mind that we're still very early in this process. The card isn't even available yet. Just because it debuts with a $250 price tag doesn't mean it will stay there indefinitely.

...what he said. I really hope these cards take off - it could make games so much more versatile, and realistic!

BE MORE OPEN MINDED!!!

Think back to the days when a 3D card was an addon to the 2D card you had already... the 3D accelerators (Voodoo I anyone?) couldn't do 2D at all so you not only had to have your, oh, I dunno, Matrox G200 or whatever but a Voodoo as well.

Things converge! It wouldn't surprise me in 5 years if PPU and GPU were combined on the same piece of hardware and dubbed an EPU (Environment Processing Unit) or something.

Of course by then motherboard hardware might be so cheap that they will bung the GPU on the motherboard. After all, these days you don't need (for example) a seperate sound card, network card, disk controller, peripheral bus controller, etc etc. It's all on the motherboard because they have become commodities.

Yes I know a lot of those things are essential to the operation of a modern computer, but my IBM PS/2 had a separate disk controller, and no sound or network, yet it was jaw-droppingly top of the range when it was released.

So I guess the PPU will become an essential part of a computer system in the future, just like proper 3D acceleration has (or certainly will be with Avalon/Longhorn/WGF/DX10 or whatever it's called right now.)

Oh and as for the price, well there's two ways to look at that.

1) remember how much the first Soundblasters cost? Several arms and legs, and probably other appendages, children, and pets. Yet now EVERYONE has sound hardware.

2) If you can afford a X850 XTPE 512MB OC whatever, or a 6800Ultra-OC 512MB, then you can afford a PPU card.

Not going to predict sucess or failure for this card but if it were built into a mainstream card (think 6800 gt short bus edition or whatever) and it was sold for $900 alot of the infinite money people would buy it.

mikehodges2 wrote:

Miserable sod.
(snip)

That's very close to a personal attack. Let's keep the discussion civil.

Zenith wrote:

This is stupid...

This product will fail.

If this product does not fail, by some sick twist of fate...I will give up PC gaming altogether.

Civil? Sure...