Nvidia going in aerospace industry...

digitalwanderer said:
You make him out to be some neutral party though and he's not, he's always been and probably always will be nVidia biased.

I like the Carmack, but I don't worship him. I think his game engines are fabulous and his games suck.

(I didn't much care of D3, but I love the D3 engine. :) )
I never intended to make Carmack out to be a neutral party. My point was that he seems to care about the industry as a whole, and just because he doesn't echo all of the current marketing BS like other industry figures is not a reason to dismiss him as irrelevant. He strikes me as someone who is honest and will tell it like it is without sugarcoating. I would think more people on this Forum would appreciate that. I guess not.

And as far as I can tell, most of his negative comments have been about next gen cpus and multi cpu programming, so I'm not sure what you're getting at with your "nVidia bias" comment.
 
Gabrobot said:
Lost what? What has physics really done for gameplay? So far physics are used mostly for graphical enhancements, not as serious gameplay additions.
In all of ID's games seeing as every object apparently has the same mass. While not a FPS what do you think this game would be like without decent physics? http://www.chroniclogic.com/ . I guess your not a fan of prey ( which is based on the D3 engine ) either seeing as the walking on walls and the changing of gravity are just graphical enhancements.

Caramck simply thinks that putting extra development into supporting physics hardware is pointless if it doesn't do anything for the gameplay.
No carmack is saying simply putting extra effort into developing physics is pointless if it doesn't do anything for the gameplay. As he quite clearly said "physics and physics hardware" and he thinks it is pointless.

Why did carmack even bother switching to opengl? he could have done everything on his fast serial processors anyway opengl ( and in the future for him D3D ) has simply been graphical enhancements. Frankly if carmack thinks intel ect can keep up the increase of IPC for ever he is joking himself as what we are getting X^2 die increase with shrinkage but every if/else in code adds 2^X complexity die shrinkage is going to hit a wall very soon anyway. ( However 3d processors will fix that )
 
Last edited by a moderator:
Skrying said:
Are you blind? Nvidia HAS been successful in having id nearly tailor their engine around Nvidia hardware in the past, present and future. Its not secret at all. I'm not sure how you can say such a thing without knowing you're basically skipping over the facts.
Actually many people think it's the opposite and that Nvidia architected their chip to be fast at stencil shadows because they knew many like Carmack would use them. If true it's not a bad decision on their part.
 
Skrying said:
Are you blind? Nvidia HAS been successful in having id nearly tailor their engine around Nvidia hardware in the past, present and future. Its not secret at all. I'm not sure how you can say such a thing without knowing you're basically skipping over the facts.
Pretty sure 3dcgi nailed it:

Carmack: "Oh boy, I can use a z-only pass for shadowing!"
NV: "Doom 3 in 2003? Double z output on NV30! Hooray! We win Doom 3 benchmarks!"

Of course, history didn't turn out like that (despite what the [H] D3 benchmarks from the NV35 launch would have you believe :v: ), but that description seems to be the general consensus. Remember, D3 was designed around the NV10 featureset, so using shaders wasn't an option. There are also things like UltraShadow (that silly NV35 OGL extension designed explcitly for Doom 3 which may or may not do anything) that show that NV catered to Carmack, not the other way around.
 
bloodbob said:
In all of ID's games seeing as every object apparently has the same mass.


Why don't you go in and take a look to see how it actually works. I'll give a hint, Doom 3's objects have mass values you can set (along with density, bounciness and friction). Doom 3 as a game made crap use of the physics, but that was entirely the fault of who ever set the values, not Jan Paul Van Waveren.

bloodbob said:
While not a FPS what do you think this game would be like without decent physics? http://www.chroniclogic.com/ .

There's no question that physics can have a huge impact on gameplay (I'm a Trespasser fan, or rather, a fan of what Trespasser was supposed to be), but what will physics hardware do to help (at least right now)? The fact is you can't make it do anything more then stick ten zillion objects bouncing around on screen. Why is it worth going to the trouble of adding hardware physics support when you can already bounce objects around and do ragdolls and what not with the software physics? (You just can't do as many of them) Hardware physics won't give you the kind of improvement that hardware graphics gave.

bloodbob said:
I guess your not a fan of prey ( which is based on the D3 engine ) either seeing as the walking on walls and the changing of gravity are just graphical enhancements.

Prey doesn't use physics hardware. And besides which, the two examples you gave could have been done with Quake's physics.

bloodbob said:
No carmack is saying simply putting extra effort into developing physics is pointless if it doesn't do anything for the gameplay. As he quite clearly said "physics and physics hardware" and he thinks it is pointless.

They already have a physics engine. (Which means there's not a lot of fundamental developing that needs to be done...it already has proper path based collision detection, they beef up the calculations and turn off their silly cut switches that turn the physics off after a few seconds and they're right up to date for their next engine.) I'd like to point out that Doom 3's physics were actually quite a big deal back when it was announced, since no one since Trespasser had done anything like that.

bloodbob said:
Why did carmack even bother switching to opengl? he could have done everything on his fast serial processors anyway opengl ( and in the future for him D3D ) has simply been graphical enhancements. Frankly if carmack thinks intel ect can keep up the increase of IPC for ever he is joking himself as what we are getting X^2 die increase with shrinkage but every if/else in code adds 2^X complexity die shrinkage is going to hit a wall very soon anyway. ( However 3d processors will fix that )

No, as I said the difference made by graphical hardware is much larger than that of physics hardware. And you can scale graphics quite easily without directly effecting gameplay, whereas physics (if they are actually implemented into gameplay) would have a huge effect, thus making it a large design challenge. Carmack already talked about this, why should I have to tell you this? If you're going to have such a strong opinion, at least stay half way informed.

digitalwanderer said:
You make him out to be some neutral party though and he's not, he's always been and probably always will be nVidia biased.

He didn't seem particularly reluctant to run Doom 3 on a Radeon 9700 back in 2002. After that he tended to prefer NVIDIA hardware, but on the other hand he had perfectly valid reasons. It's interesting to note that NVIDIA has always been the one pushing itself as the card for Doom 3, ect. However, Doom 3 is not a TWIMTBP game, and I think that's an interesting point to be made. In fact, when Doom 3 came out Carmack said that either highend card was fine, despite NVIDIA being somewhat ahead in benchmarks, which if anything came across like he was saying the difference didn't matter that much and ATI was still quite fine. Again, I'm having a hard time seeing a bias. Honestly, the only time I can think of him really bashing a card, was the Geforce 4MX, because he thought it a dishonest way to market essentially a Geforce 2 MX card.

digitalwanderer said:
I like the Carmack, but I don't worship him. I think his game engines are fabulous and his games suck.

Doom 3 wasn't his game. ;)

Doom RPG and Orcs and Elves are the only games he's done major gameplay design on since Quake 3.

Skrying said:
Are you blind? Nvidia HAS been successful in having id nearly tailor their engine around Nvidia hardware in the past, present and future. Its not secret at all. I'm not sure how you can say such a thing without knowing you're basically skipping over the facts.

As was said before, it was more a case of NVIDIA tailoring their hardware around Doom 3 then vice versa. And no wonder, given Carmack picking ATI's card to show Doom 3 off on at E3 2002. Since then, as was pointed out, NVIDIA has simply had the better hardware for what Carmack was trying to do. At least until the Xbox 360 came along, and now he finds it easier to test things on it because it won't get broken by driver updates. And in case you missed it somewhere along the way, the Xbox 360 is ATI powered.

Skrying said:
The engine sucks because it was a huge investment and has basically done nothing. I struggle to name 4 games that use it. Of the three I know two are id franchises. Previous efforts from id had MUCH better success, because the engine could do everything. The Doom 3 engine has produced games that so far all look the same. Quake Wars being the lone exception, which isnt out yet, that is extremely heavily modified to reach a certain goal.

Skrying said:
I think Doom 3 as a game was terrible. I think the engine was a commericial failure. I love Prey and it uses the Doom 3 engine, still doesnt chance my thoughts that as a whole the engine didnt do near as well, commericially, as id wanted it to.

How do you know how id intended the Doom 3 engine to do commercially? Consider these facts: They stated several times (mostly back in 2002-2003) that they make far more on game sales then on engine licenses (why do you think Epic, which does rely quite a bit on engine licenses, do so much engine hyping and support in order to win the shit loads of licensees? Because they need that many to actually make that kind of focus pay off). They didn't openly offer the Doom 3 license for sale until after the game was released (Prey being an exception, because 3DRealms is an old buddy of id Software), and about a year and a half ago, they started encouraging prospective licensees not to license the Doom 3 engine, and instead steered them towards their new engine technology (which presumably they are openly licensing before their next game is released).


I don't consider Carmack a god or any such nonsense, but I find all the Carmack bashing to be rather unreasonable. I also find it ironic that it's often bashers who seem to think he is some kind of deity who is responsible for everything id Software does, and all the game design choices. The fact is he does technology stuff, mostly graphics related. It's people like Jan Paul Van Waveren who is in charge of the physics, and it's the game designers who are responsible for the gameplay (and the way that the physics objects are set up).
 
Gabrobot said:
Doom 3 wasn't his game.

I don't think I'd go too far with such claims. There have been plenty of past examples where Carmack has influenced overall game design simply because of his position at id.
 
Well, I suspect NV saw a bit of a two-fer here, PR warm 'n fuzzies-wise. They've been making a push in the aerospace professional market anyway:

http://www.send2press.com/newswire/2006-05-0516-002.shtml

Good for them, as far as I'm concerned. Made me feel a little warmer and fuzzier about them. :smile:

Edit: And, oh yeah, re Carmack's future influence. . .well, could be he's a has-been. But then this is a community that has always heavily discounted the past and present in favor of the immediate future. Which makes Carmack and Newell big yawners right now, and Epic the cock of the walk. I'll be much more impressed if everyone is still oohing and aahing over Epic say 6 months after UT2007 is released.
 
John Reynolds said:
I don't think I'd go too far with such claims. There have been plenty of past examples where Carmack has influenced overall game design simply because of his position at id.

Well, these are his words:

PCGamer Interview With John Carmack said:
PCG: How much say do you have in what goes into an id game?

JC: Internally, Quake III was viewed as my game. It was a game I wanted to play. It was the id game that I probably spent the most time playing and enjoyed the most, but it was actually one of our less successful titles. The focused minimalism is appealing for me, but probably isn’t the best direction for a top-flight commercial game company.

So I don’t have much say any more. Post Quake III, I’ve stepped back a bit from the design side on the PC space because I’m really not representative of what most of our market is now. I did realize that my very simplified game-design ethic isn’t really what the market is demanding. I’m no longer in there vetoing things being added to the game design. That’s probably one of the reasons why [Doom 3] may have taken longer to ship, but was a richer game for it.

The only thing I can think of that he really had anything to do with gameplay wise in Doom 3, was using advanced GUI's with the mouse instead of a use button. Perhaps he had more to do with the gameplay early on, but then it should be noted that the Alpha indicated a more interesting gameplay direction then the final game actually used. (Zombies getting back up, and being able to stand on them to keep them from getting back up)
 
Err, the NV35 numbers from [H] were absolutely correct, I got those framerates with it back then as well.
 
There was also a question about what he thinks about AMD buying ATI. He said that he preferred even three major players in GPU arena, and two upto this point is OK. He said that he would definitely not like to see a single player (which would be Nvidia of course if ATI leaves high-end market) which dictates everything.

In addition, he also went ahead and said that he does not like IHVs changing a particular behavior of the GPU behind their back (he gave offloading vertex processing to CPU in driver as an example, but I guess shader optimization is another area), and both companies are guilty in this respect. He even went ahead and said something like "bad bad IHVs".
 
_xxx_ said:
Err, the NV35 numbers from [H] were absolutely correct, I got those framerates with it back then as well.
Except D3 came out 15 months after the [H] benchmarks. You know, after the NV40 came out and all.
 
I can only assume that Carmack will be using leftover FlowFX coolers from NV30 as his rocket's propulsion system from now on. :p

(I'm sorry, I couldn't resist, it's been sooo long since I've last been able to make a FlowFX gag!)
 
_xxx_ said:
And still I was able to play it with the NV35. I don't quite get what you're trying to say.
I'm saying that the much-lauded ability of the NV3x to play Doom 3 was pretty much irrelevant by the time it came out since it had already been replaced by the NV40. That NV35 was supposedly (at the time) faster than R3x0 in Doom 3 never really played the role that NVIDIA wanted it to play.
 
If he can't see the significant advancements that physics and physics hardware will bring to gaming, then he really has lost it.

Carmack has stated he's in favor of simpler gameplay styles, though Doom 3 had a fairly complex physics engine in spite of that.

Yet, for some reason, some of you think he is hopelessly out of touch. Why, exactly?

Because EA and Ubisoft drive the market today, not ID. He's no longer the standard which everyone else follows, he's just the one oddball with enough clout to still be around.

Mark Rein/Epic and Nvidia with the Unreal Tournament series

Nvidia sponsers their games, but I wouldn't say they're nvidia optimized. They're really more AMD ******s.

Frankly if carmack thinks intel ect can keep up the increase of IPC for ever he is joking himself as what we are getting X^2 die increase with shrinkage but every if/else in code adds 2^X complexity die shrinkage is going to hit a wall very soon anyway.

The Core 2 Duo showed quite a bit more IPC can be had, and without sacrificing clock speed, and half of its chip is another core. They probably could have increased IPC even more if they had gone single core. (but Intel needed to design a native multicore chip eventually) Anyhow, if designs like the Cell can provide any improvement to games, then a design with a powerful OOE core combined with a powerful on-die math coprocessor is probably even better. Core 2 Solo + PhysX processor maybe?

In addition, he also went ahead and said that he does not like IHVs changing a particular behavior of the GPU behind their back (he gave offloading vertex processing to CPU in driver as an example,

So basically he reprimanded 3dfx, PowerVR, and Intel there, and that's it.
 
Back
Top