JC Keynote talks consoles

[nelg]

I think Carmacks point about Physics is this. If you want to do it 'all' you had better do it all 'correctly' and that is going to take more power than these consoles will have. Tumbling boxes is one thing, completely destructable environments where everything acts correctly is a huge huge problem

That's how I interpreted it also.

 

[Tsmit42]

I think Carmacks point about Physics is this. If you want to do it 'all' you had better do it all 'correctly' and that is going to take more power than these consoles will have. Tumbling boxes is one thing, completely destructable environments where everything acts correctly is a huge huge problem

But it's alright for the graphics? JC is just about pretty graphics and that is why Doom 3 sucked so bad. It was nice to look at for a while, but once you start playing it you knew all the game had was pretty lighting. A.I. and physics/animation is what will drive the next generation of games. If not, all we are getting is the same games with a new look. Graphics are at a point where they outpace AI and physics tremendously and it is time for them to catch up, even if you can't do 100% what you want, 50% is better than 20%(just made up numbers, but you get the point).

I would take the graphics of today for the next 5 years if AI and physical simulation improves dramatically. I am tried of seeing pretty pictures moving around like Frankenstein, having 20/1000 vision, and the memory of a 120-year-old.

 

[Titanio]

Well last i check the rsx and xenos weren't on the market yet . By november at i will have thier tri card r520 on the market and by the time rsx comes out in marchish of 2006 i'm sure both ati and nvidia will have put out refreshes or modified cards , if not be reading thier new chips like the r600 and whatever replaces the g70 .

I believe he said *current*. Maybe he meant to say "at the time of their release", in which case the gap on the graphics side would obviously be closer.

don't forget current highend cpus are now dual core . a dual core athlon 64 may very well come pretty close to a x360 cpu and a cell in realworld tasks . We really haven't seen any games specificly targeting dual core cpus

In some tasks I'm sure it'd be better, but in others, not better. In others I think you could see a much bigger difference in the opposite direction, and if those are the tasks consuming more execution time, that's a better optimisation. It's still in the realm of the theoretical for most of us here of course. But I wonder what would happen if you tried to take code optimised for Cell and run it on a dual core Intel or AMD In some way that's what he's proposing doing, because it'd be easier from a dev point of view (although I don't know how optimised his code for one platform would be in that instance), but he kind of glosses over the performance issue there once you come to PCs..but I guess he can depend on PC CPUs making up some ground by the time his next game ships.

isn't the ppu add in card for pcs supposed to do 50000 pieces ?

Not sure, but to clarify, Epic never put any figure on what cell would be doing in that regard.

edit - I think the initial claim for Physx was 32k rigid bodies, 40-50k particles.

 

[Gabrobot]

What about a game like Motorstorm. I'm not saying that it is 100% real right now. That's not my point. My point is wouldn't physics play a huge part in the mudd since in that video. Wouldn't 1000s of pieces of mud have to be calculated in real-time to get it to stick to the side of cars, trucks, 4 wheelers, and dirt bikes the way it does?

Look at the way it streaks on the windshield when the wipers are turned on. If SCEE can get this done right the way they displayed at E3 wouldn't that put Carmack's theory about physics at a 0?

...This is hilarious. They don't calculate "1000s of pieces of mud", they use particle effects (which may have physics, but they're group physics). The mud sticking to the cars are decals, and the mud streaking is simply an effect, not physics. These are graphical features, which are actually what Carmack is advocating...you've actually proved his point.

WOW, I've been reading this forum for almost 2yrs and never felt compelled to write...until now. I don't think I've ever heard JC try to spin anything, but that feeling ended yesterday. How he can possibly believe physics won't have a MAJOR affect on gameplay is beyond my comprehension. I lust after games that contain fully destructable envoroments.

Imagine a pirate game where the cannon balls leave gaping holes in the sides of the ship. Water might rush in if the void is low enough, or if up high it could be a point of entry for the enemy, or maybe a "sniper" position.

You need to read. Carmack is saying current hardware can't simulate physics that would have a major affect on gameplay. His point is that what you're saying is impossible on current hardware.

About AI he is compeletely off way, people do notice when a AI is scrippted, if not at first time second or third they will (unless they are very stupid), this is the kind of thing that really add replay vakue (it is the reason why Halo is famous IMO, and even Halo ido a have a AI that should be very poor in comparition to what they should be to do if AI already is a prymary focus).

Er, just about every major game released recently uses scripted AI. Scripted AI is not when a monster jumps out from a corner...that's a trigger in the level that plays an animation. Scripts are an actual programming language that the AI is programmed in, and generally you have certain actions (like ducking) which are called into action based on what's happening. So instead of coming up with a really advanced AI simulation that has the characters actually visually processing their environment and figuring out how and what to move (this is like what Trespasser tried to do) you have paths that the characters can use to navigate and in certainly situations they use certain scripted actions like ducking and flanking and so forth. To the user there's not much difference (except that the scripted AI is going to usually look more polished and smooth), but one runs at 2 fps and takes a huge amount of work to develop and the other runs at 60 fps and can be quickly and easily set up and directed. It's not particularly hard to figure out which is the better way.

 

[Bjorn]

He did say they were about as powerful as current high end PCs, yes. He later said they were all in the same "ballpark-ish". I'm not sure what perspective he was coming from with the former at least, but yeah, it certainly is one of the odder comments (and frankly isn't true - there's no GPU on the PC market as powerful as RSX or Xenos, ..

So we know for a fact that the RSX and Xenos are faster then the GTX 7800 ? I know that transistors aren't everything but the GTX has what, 60 million more transistors then the Xenos. And i seriosly doubt that the RSX has that many more then the GTX. So don't expect any miracles from any of them. Faster ? perhaps, but not by much. And you would think that Carmack knows a bit more about this then most people here, after all, he has both dev kits to fiddle with.

 

[jvd]

I believe he said *current*. Maybe he meant to say "at the time of their release", in which case the gap on the graphics side would obviously be closer.

Of course whats current for us as end users and him as a developer are two diffrent things . Aside from that I think something for late summer release should be considered current. THe r520 should be released next month .

n some tasks I'm sure it'd be better, but in others, not bette

well also remember that pc cpus are asked to do diffrent things and have certian tasks off loaded to other areas .

Not sure, but to clarify, Epic never put any figure on what cell would be doing in that regard.

mckmass seems to indicate that they have .

Remember the cell and xenon will be asked to do many more tasks than just physics . So a dedicated chip should perform much faster and of course take some load off the cpus .



I believe in the genration that happens in 2010 or so we will see instead of massive cpus like we have in this gen , more dedicated hardware again. We will see perhaps a ppu + sound chip intergrated . I believe this is the onl way to truely advance a.i and physics

 

[Titanio]

You need to read. Carmack is saying current hardware can't simulate physics that would have a major affect on gameplay. His point is that what you're saying is impossible on current hardware.

He's saying you couldn't "simulate the world" - Half Life 2 was doing physics on P4 that had a "major effect" on gameplay IMO. I think it's a sliding scale, and he doesn't think you should go overboard..past a certain point it'd be a whole lot of work.

So we know for a fact that the RSX and Xenos are faster then the GTX 7800 ? I know that transistors aren't everything but the GTX has what, 60 million more transistors then the Xenos. And i seriosly doubt that the RSX has that many more then the GTX. So don't expect any miracles from any of them. Faster ? perhaps, but not by much. And you would think that Carmack knows a bit more about this then most people here, after all, he has both dev kits to fiddle with.

RSX is very likely simply the G70 but clocked faster. In that instance it'd obviously be more powerful than the current highend (G70). Also consider the systems they're going into - cpu-to-gpu bandwidth is many times that on the PC. That's something for example Nvidia keeps pointing at with RSX, so I presume it's a positively differentiating factor.

Remember the cell and xenon will be asked to do many more tasks than just physics . So a dedicated chip should perform much faster and of course take some load off the cpus .

This is true of course, though it is interesting to wonder how something like Cell working wholly on physics would compare to PhysX. CPUs also have some advantages. For what it's worth, there had been noises that Cell in particular could outperform PhysX some time ago, though that may well have been under the consideration of the entire chip working on that task.

 

[Bjorn]

he spoke on that as well, thats just like the grass swaying in the wind, or more rock rolling down a hill, yeah it looks good but it doesn't really add anything to game play, even if the mud wasn't there the game would be the same

Better graphics doesn't add to the game play either if you're using that logic.

 

[Bjorn]

RSX is very likely simply the G70 but clocked faster. In that instance it'd obviously be more powerful than the current highend (G70). Also consider the systems they're going into - cpu-to-gpu bandwidth is many times that on the PC. That's something for example Nvidia keeps pointing at with RSX, so I presume it's a positively differentiating factor.

There might be some things that will run much faster on the consoles then on the PC thanks to the GPU-CPU bandwidth advantage. But that doesn't make the Xenos and the RSX faster then current PC gpu's. Well, they might be a bit faster, but i consider it a moot point if it's not more then 10-20% which i seriously doubt. And of course, Carmack seemed to be of the opinion that they were'nt faster at all.

 

[jvd]

This is true of course, though it is interesting to wonder how something like Cell working wholly on physics would compare to PhysX. CPUs also have some advantages. For what it's worth, there had been noises that Cell in particular could outperform PhysX some time ago, though that may well have been under the consideration of the entire chip working on that task.

that could be true though i dunno. However when looking at price to make (not to sell since sony will make money off the software and physX has to make it off the hardware ) it may very well be that the ppu offers performance on par with a cell chip but at much cheaper costs and of course will only be asked to do physics .

Of course we don't know how they will go about it but depending on costs we may even see high end ppus with two chips and 256megs of ram if the market can handle it .

The future of pcs is very interesting .

Multi core chips
x86-64 coming into full force
eax 5 with sound cards having thier own 64 megs of ram
2 and 3 video cards
ppus

Lots of interesting stuff

 

[pc999]

...Er, just about every major game released recently uses scripted AI. (...). To the user there's not much difference (except that the scripted AI is going to usually look more polished and smooth), but one runs at 2 fps and takes a huge amount of work to develop and the other runs at 60 fps and can be quickly and easily set up and directed. It's not particularly hard to figure out which is the better way.

The problem is that user notice the diference, not at the 1-2 time but most of people pass the game because they already learned the AI, that is one of the reason why normally as levels pass aperar or more or new enemys in most games (e.g. D3) it is a way of making players have a new thing to do till the end (not the only reason, players will always want theem anyway), a really tatical AI would/should percive what we are doing then not fall always in the same error.

Anyway the performance argument is always a good one, but if there is ways to do I really think that they should do, and a lot people of think that there is much room too improve it.

You have a nice thread here http://www.beyond3d.com/forum/showthread.php?t=21946

About peple dont want good AI because it would be too hard, I think that is a bad argument, because today we have AI that is hard to killbecause they need to be take 5+ shoots in the chest, they will just need to tone that and aim (etc...) down, IMO everyone would prefer defeat AI because they are smarter, than because they do better a routine.

BTW thanks for the info about scripted AI :smile:

 

[inefficient]

A great example of a game in which you basically have full control over the world and has an accurate physics simulation is SilentStorm2. If you have played that game you would have an idea of problems a game world like this brings.

In HL2 you only had physical control over things the designers wanted you to have control over. In SS2 you had control over pretty much everything. And that resulted in some pretty random outcomes. For example you could use explosives to blow arbitrary holes in ceilings /floors/walls. You could use firepower from guns or explosives to blow away locks from doors or just blow the whole door away. You could collapse structures.

All that freedom was a really nice thing... when it worked. But it was not uncommon to totally screw yourself so that you had to reload an earlier save. Most involved unintentionally blowing up critical areas so that you could no longer get to areas you wanted to go. Or “burning your bridges" in a way that left you stranded in a certain area.

A lot of these individual problems might have been solved if they did more QA testing and carefully limiting the player in certain area. But because of all the variables introduced by giving the player so much freedom it was probably just not feasible for the small PC dev that created that game.

It was still a fun little game. I just don't think mainstream gamers will ever be able to accept a game where there where so many accidental game over scenarios.

 

[Titanio]

Todd Hollenshead, CEO of iD, was asked about the Multiprocessor/Cell comments..

http://www.gameinformer.com/News/Story/200508/N05.0813.1950.27388.htm

GI: Now, this is something I took out of John Carmack’s speech yesterday, but he seemed not very hot on the whole multi-processor situation.

TH: No, no he’s not. He gave a keynote at GDC two years ago where he basically said that his philosophy was, make the most powerful processor and the most powerful GPU and that would give him the best ability to make the game look great.

GI: The one thing that I kind of got out of it yesterday, and I kind of could be off base, but he seemed happier with Microsoft, and not too thrilled with Sony and the Cell technology. Do you think that’s accurate?

TH: I think that’s accurate. I’m not really one to put words in John’s mouth but what I do know is we’ve had 360 stuff for a while and it was relatively easy to bring our internal project onto 360. We just got the PS3 stuff recently and it was relatively more difficult to bring that up on the PS3. So John’s first impressions are, “360 great, PS3 – pain in my ass.†I think the more we work with it - I don’t know if John himself will be doing the primary PS3 work or not - but we’ll have to see about that stuff. Also, 360 is further along in their process as well. They have more final hardware, and they have better drivers, and the SDK has been more refined and revised. I think the PS3 stuff, in all fairness to Sony is a little bit more raw. I think we’ll have to wait and see, but I don’t think that’s going to ultimately change the way we’re going to approach developing on PS3. He knows from a technology horsepower standpoint that it’ll do everything that we want it to do, so we’re committed to it.

I think they're on a lot more even keel with regard to all this stuff than reports and commentary suggests. I think it's also good to re-echo Carmack's comments about the CPU stuff being "just quibbles", and how he's really very content with everyone's hardware.

 

[EpicZero]

Todd Hollenshead, CEO of iD, was asked about the Multiprocessor/Cell comments..

http://www.gameinformer.com/News/Story/200508/N05.0813.1950.27388.htm



I think they're on a lot more even keel with regard to all this stuff than reports and commentary suggests. I think it's also good to re-echo Carmack's comments about the CPU stuff being "just quibbles", and how he's really very content with everyone's hardware.

So I take it from this, he really doesn't like the PS3 too much.

 

[Acert93]

So I take it from this, he really doesn't like the PS3 too much.

Well, GI (and most others) took JC's comments as saying he, "seemed happier with Microsoft, and not too thrilled with Sony and the Cell technology". Tood agreed with this and IHO summerized it in general as:

So John’s first impressions are, “360 great, PS3 – pain in my ass.â€￾

But just as importantly, PS3 being a pain does not mean iD is going to ignore the PS3:

He knows from a technology horsepower standpoint that it’ll do everything that we want it to do, so we’re committed to it.

While Todd does note how it is still early on for Sony--and these are first impressions-- but he does report what we have known for a while: MS has some pretty good tools and is looking top be in the drivers seat in that area when XNA fully comes 'online' in 2006.

What iD has noted is basically a generalization, a summary, of MS and Sony and their respective philosophies: A software company and a consumer hardware company. Neither are neglecting software or hardware, but in their respective console products you get a feel for how their primary business interests gives them an advantage in those areas. Sony has been able to leverage cutting edge industrial designs, optical media, and connections with media (gaming and movie) outlets. MS has experience in developer tools and PC contacts.

Similar products with minor twists that appeal to different audiances.

 

[wco81]

Is Carmack still a big into OpenGL as opposed to DirectX?

If he was, wouldn't he prefer the PS3 instead of the X360?

 

[Gabrobot]

He's saying you couldn't "simulate the world" - Half Life 2 was doing physics on P4 that had a "major effect" on gameplay IMO. I think it's a sliding scale, and he doesn't think you should go overboard..past a certain point it'd be a whole lot of work.

Well yeah, it's all kind of relative...what I was thinking was Hl2's physics effect on gameplay is pretty minor compared with simulating the whole world or something. So when I said major, I was thinking really major.

The problem is that user notice the diference, not at the 1-2 time but most of people pass the game because they already learned the AI, that is one of the reason why normally as levels pass aperar or more or new enemys in most games (e.g. D3) it is a way of making players have a new thing to do till the end (not the only reason, players will always want theem anyway), a really tatical AI would/should percive what we are doing then not fall always in the same error.

Well that's what's great about scripted AI...there's a lot of stuff that can be improved and added. You could set up a pretty advanced system where there are tons of context based scripted actions (like interacting with the environment, i.e. shooting through glass, knocking down doors, ect.) combined with some extra programming to mix up their strategies a bit and even some basic learning, and things would be pretty impressive. The thing is you can pour tons of effort into scripted AI and have things turn out beautifully, plus you still have some control over how things happen. The quality really just has to do with how much you want to put into it (so in Doom 3's case, they just didn't do a lot with the AI).

 

[function]

Is Carmack still a big into OpenGL as opposed to DirectX?

If he was, wouldn't he prefer the PS3 instead of the X360?

That's supposing that what he sees as the advantages of the Xbox 360 platform didn't outweigh his liking for OpenGL ...

 

[EpicZero]

Well, GI (and most others) took JC's comments as saying he, "seemed happier with Microsoft, and not too thrilled with Sony and the Cell technology". Tood agreed with this and IHO summerized it in general as:



But just as importantly, PS3 being a pain does not mean iD is going to ignore the PS3:



While Todd does note how it is still early on for Sony--and these are first impressions-- but he does report what we have known for a while: MS has some pretty good tools and is looking top be in the drivers seat in that area when XNA fully comes 'online' in 2006.

What iD has noted is basically a generalization, a summary, of MS and Sony and their respective philosophies: A software company and a consumer hardware company. Neither are neglecting software or hardware, but in their respective console products you get a feel for how their primary business interests gives them an advantage in those areas. Sony has been able to leverage cutting edge industrial designs, optical media, and connections with media (gaming and movie) outlets. MS has experience in developer tools and PC contacts.

Similar products with minor twists that appeal to different audiances.

Ok, thank you Acert. But one more question would be, why does Carmack seem to differ with Rein/Sweeney?

 

[jvd]

Ok, thank you Acert. But one more question would be, why does Carmack seem to differ with Rein/Sweeney?

I don't think they do . Sweeney is a nvidia boy . He loves nvidia . However he is able to say both systems are very powerfull in all his interviews he has made it a pont ot bring up the xbox 360 even when not asked .

I think the systems are going to be so close in power that its really the development tools that will set them apart.