Carmack again on PS3-360

[Rangers]

http://www.gameinformer.com/News/Story/200701/N07.0109.1737.15034.htm?Page=3

GI: After Enemy Territories: Quake Wars ships with the highly modified Doom 3 engine and the MegaTexture support, is it time for you guys to move on from that engine?

Carmack: Yes, the in-house development project that we’ve been working on is all new technology. It still has some roots in the Doom 3 technology, but almost everything is new in there. We’re still not talking about exactly what the project is, but it’s a new IP, it’s diverting a little bit from the standard id formula and it’s not just a first-person shooter. Technically, it’s build around an advancement over the MegaTexture technology from Quake Wars. Where that was applied just to the terrain, the version of the new technology applies it into everything, so we can have that level of rich detail on all the surfaces on the entire world. That’s the push that we’re making with graphics technology. The gameplay is somewhat different from anything that we’ve of done before. The company is pursuing Wolfenstein, Doom and Quake franchises with other partner developers and all, but we’re trying to develop a brand-new franchise with this new one. Hopefully, we’ll be talking about that sometime this year, and we’ll be able to go ahead and come out of our own little cone of silence about it.

New game sounds rad. I thought it was supposed to be Wolfenstein?

 

[archangelmorph]

I don't see where he's coming from with his bitterness regarding the multi-core architectures of the next-gen consoles..

It's like he's implying that the added complexity of the hardware makes developers lives harder in all cases which in my oppinion isn't true..

On what basis is it harder?

Harder compared to PC.? (well sure! consoles have always been that way to be fair! it's a part of the unfamiliarity of the h/w + toolsets which makes getting up to speed with the platform a little trickier.. But then it is a learning process.. And once you've cracked it, you don't ever have to "learn" again.)

Or maybe hard compared to last gen..? (well to be fair it will only be harder if as a developer you choose to go out of your way to try to fully squeeze the power out of the system to maximise your engines performance.. If you wanted to build an xbox game on the xbox 360 then i'm sure it wouldn't be "harder" or run "slower" than on the xbox, even if you decided to only utilise one core, no multi-threading and basic GPu functionality.. It would probably scorch an xbox version of the game in performance purely because the processor speeds (CPU + GPU) have been drastically increased over the xbox) so I don't see how developing on the new hardware could ever be counter-productive..?)

And if one decided to fully utilise the hardware to do things that could never be done on xbox and PS2 and to build a game engine with all the next-gen bells and whistles then that's purely a developer decision and one that shouldn't expect you to learn a bit more about how the new hardware works to get there.. Heck even in the PC space developers had to learn new techology when shader programming was first introduced, to be able to get the new performance out of the hardware.. Sure some effects could still have been done using the fixed function pipeline but as a developer you have to expect to change your way of working if you want to increase product quality, it's just the nature of new hardware and the people who want to maximise it's use..

If he seriously considers it too much effort putting in the extra hard yards just to get the kind of performance from next gen multi-core architectures which we are seeing more of then nobodies stopping him from not bothering to be fair.. Heck the PS2 still has the biggest gaming install base of any hardware platform in the world AND its STILL growing so i'm sure there are enough code libraries available for him and his team to cheaply and efficiently develop for the platform and still make bucketloads of cash..

I understand Carmack is a PC kind of guy but how can you complain like he's doing about the direction of change in technology when:

a) It's been inevitable for years to be fair & not just in the gaming hardware space.. Why do you think we've had massively parallel supercomputers for so bloody long?
b) The effort required to make the transition is (like I said before) more of a one stop-shop.. You build your core technology libraries and code base and then you re-use (improving on them later only if you feel the need).. Isn't that one of the fundamental ideal for practical coding in enterprise?
c) In the end it will only increase the technical calibre of industry programmers as a whole and make more intelligent, learned and experienced people, even more intelligent, learned and expreienced.. I really don't see what's wrong with that?

Props to you Carmack for being a genius and all but maybe there's some views you might consider thinking over or keeping to yourself in future..

 

[liolio]

i find interesting that carmack thinks that MS should subsidizing for live in the end.

Carmack: I really haven’t even given it that much consideration. I think that they are providing a decent service for that. We’re very happy with our live downloads and stuff like that, and there’s value there, but I would expect that Microsoft would end up subsidizing that for their platform, but I don’t really know.

 

[Arwin]

I don't see where he's coming from with his bitterness regarding the multi-core architectures of the next-gen consoles..

It's more bitterness regarding multi-core development in general.

GI: At QuakeCon two years ago, you were very adamant during your keynote about not being too thrilled about developing for multi-core systems. Not just specifically with PCs, but also the PlayStation 3 and Xbox 360.

He doesn't like it - it is difficult to manage, especially in a team and when most people do not have experience with it yet:

I do sweat about the fragility of what we do with the large-scale software stuff with multiple programmers developing on things, and adding multi-core development makes it much scarier and much worse in that regard.

He also says his life would have been much easier if we'd just had some nice boost in clock. So there, he just doesn't like multi-core development per se, and therefore his comments should be taken from that perspective.

 

[archangelmorph]

It's more bitterness regarding multi-core development in general.



He doesn't like it - it is difficult to manage, especially in a team and when most people do not have experience with it yet:

That's understandable yes, but as I said, It's a learning process & once you've cracked it.. I'm sure the same level of challenge would have been there when his team first had to get to grips with shader programming (or at least some degree of it).. Again it's just another platform/temporary-challenge which programmers are expected to face as technology (in hardware and software) advances (which it always does.. Hence the inevitability..)

He also says his life would have been much easier if we'd just had some nice boost in clock. So there, he just doesn't like multi-core development per se, and therefore his comments should be taken from that perspective.

But when it's quite clear when you look at how clock speed increase has begun to level off over the past few years (considering how relentless semi-conductor companies like Intel and AMD have been in the past with regards to trying to push them faster and faster) that this would not be the case.. I'm sure if such companies could effectively/efficiently/economically increase clock speeds further in order to give the same level of performance increase which could be achieved by adding another core, i'm pretty sure they would have by now since that's what they've been doing prior.. So why it is to hard to see that his prefered ideal in terms of the direction of hardware advancement is, was and probably wont ever happen..?

I see how your trying to defend him but by saying "my life would be easier if they didn't make hardware better this way.." is rather a pointless statement to make in my view..
I'm sure his life would have been made easier if the hardware NEVER changed.. We just sat here with Geforce 8800 GTX until the end of time..
But you, I and Mr Carmack know full well that this will never happen..

Hardware will advance..
Firmware will change..
And as a result software development paradigms will shift..

It's the nature of technology and so what's the point in complaining about it?

He should just suck in his gut and get on with it.. especially considering the fact that two years down the line it's going to be a moot point..

(Especially when in 10 yrs time, his "old" scalable multi-threaded engine built for quad-core PCs runs approximately 1000x faster on the modern 50-core CPUs of the era without the need for code modification)

 

[Arwin]

I see how your trying to defend him

Oh, that's a mistake - I definitely am not. Í'm just trying to clarify what he actually said and why. I am a big fan of the Cell myself and while I admire what JC did in the past, I am a much bigger fan of the console cowboys that really take hardware to unforseen limits.

 

[archangelmorph]

Oh, that's a mistake - I definitely am not. Ã￾'m just trying to clarify what he actually said and why. I am a big fan of the Cell myself and while I admire what JC did in the past, I am a much bigger fan of the console cowboys that really take hardware to unforseen limits.

Sorry my mistake..

 

[pjbliverpool]

He seems to be quite down on the benefits of DX10 and Vista in general. Can any other devs comment on that? Is MS just trying to blind us with hype about the huge improvements both for developers and end users?

 

[inefficient]

Solo core lover

The Wii is calling you Carmack!

It's just a clock speed improvement over its predecessor. Just what you wanted.

 

[archangelmorph]

The Wii is calling you Carmack!

It's just a clock speed improvement over its predecessor. Just what you wanted.

 

[Rangers]

That's understandable yes, but as I said, It's a learning process & once you've cracked it.. I'm sure the same level of challenge would have been there when his team first had to get to grips with shader programming (or at least some degree of it).. Again it's just another platform/temporary-challenge which programmers are expected to face as technology (in hardware and software) advances (which it always does.. Hence the inevitability..)


But when it's quite clear when you look at how clock speed increase has begun to level off over the past few years (considering how relentless semi-conductor companies like Intel and AMD have been in the past with regards to trying to push them faster and faster) that this would not be the case.. I'm sure if such companies could effectively/efficiently/economically increase clock speeds further in order to give the same level of performance increase which could be achieved by adding another core, i'm pretty sure they would have by now since that's what they've been doing prior.. So why it is to hard to see that his prefered ideal in terms of the direction of hardware advancement is, was and probably wont ever happen..?

I see how your trying to defend him but by saying "my life would be easier if they didn't make hardware better this way.." is rather a pointless statement to make in my view..
I'm sure his life would have been made easier if the hardware NEVER changed.. We just sat here with Geforce 8800 GTX until the end of time..
But you, I and Mr Carmack know full well that this will never happen..

Hardware will advance..
Firmware will change..
And as a result software development paradigms will shift..

It's the nature of technology and so what's the point in complaining about it?

He should just suck in his gut and get on with it.. especially considering the fact that two years down the line it's going to be a moot point..

(Especially when in 10 yrs time, his "old" scalable multi-threaded engine built for quad-core PCs runs approximately 1000x faster on the modern 50-core CPUs of the era without the need for code modification)

So where's your game? I'd like to try out your amazing game.

That aside, saying he'd rather have faster single core CPU's is just a no-brainer.

And Carmack makes better looking games than those console Cowboys all day, lest I pull up some PS2 pics to prove it. Sure, there's tradeoffs all around, he has more power to work with, but there's something to be said for working on the bleeding edge PC hardware too. A certain eliteness to it.

 

[Arwin]

So where's your game? I'd like to try out your amazing game.

That aside, saying he'd rather have faster single core CPU's is just a no-brainer.

And Carmack makes better looking games than those console Cowboys all day, lest I pull up some PS2 pics to prove it. Sure, there's tradeoffs all around, he has more power to work with, but there's something to be said for working on the bleeding edge PC hardware too. A certain eliteness to it.

Ah, but there's the contradiction. There is something to be said for working on the bleeding edge hardware (PC or otherwise), yes, but Carmack very clearly doesn't care about that anymore and hasn't for a few years now.

I guess it's just hard to come back down from rocket science.

 

[archangelmorph]

So where's your game? I'd like to try out your amazing game.

Oh sorry!

I didn't kno I needed to be a member of the elite of the industries software engineering specialists with a vast library of games/technical masterpieces under my belt before I could comment on a couple of JC's statements which relate to an entirely seperate topic altogether..?

So I guess when JC makes a comment on which is the best toilet cleaner to use I should shut my mouth too huh?

Seriously it's a bit immature to carry around this kind of mentality; the "Oh you can't challenge his views! He's John Carmack! The God of all gaming in which we bow down and worship his feet" belief system was something I thought was reserved only for children and fan***s but I guess I was wrong..

 

[Ben-Nice]

Is he basically saying

1) I don't like multicore. It's hard and a hassle.
2) Since it's hard and a hassle and MS has better tools to make it less hard and less of a hassle I prefer the X360 over the PS3.
3) I will back, support, pull for etc. The company that makes my job the easiest.

 

[scooby_dooby]

That's understandable yes, but as I said, It's a learning process & once you've cracked it..

That's his entire point, in his opinion multi-core programming will not be 'cracked' for a long time. Many people have been working on it, for many years, and it's still coming along slowly, he sees no major breakthroughs for quite some time.

Just his opinion, but I'd like to know what experience you have that you can sit there and contradict his statements.

 

[ERP]

That's understandable yes, but as I said, It's a learning process & once you've cracked it..

This is the fundamental misconception, it's certainly a learning process, but there is no such thing as "cracking it" it's a fundamentally hard problem.

Tools and experience can make it easier, but it will always be significantly more difficult.

Multiprocessor architectures are inevitible, using them effectively is a hard problem, and it's one that compounds all of the other hard problems in game development.

 

[Entropy]

Is he basically saying

1) I don't like multicore. It's hard and a hassle.
2) Since it's hard and a hassle and MS has better tools to make it less hard and less of a hassle I prefer the X360 over the PS3.
3) I will back, support, pull for etc. The company that makes my job the easiest.

So it would seem, yes.

Which is rational.
He does seem to have a bit of a problem dealing with the progression to multiprocessing. While he acknowledges that it's the way forward and that old paradigm wasn't holding up anymore, he can't help complaining about the effort involved in learning new skills. I know that he is correct in his cautioning that there are fundamental difficulties in parallell processing, but so what - any change will bring with it pitfalls and possibilities.

If you want to utilize all of that unused performance, it’s going to become more of a risk to you and bring pain and suffering to the programming side.

This quote demonstrate both his points and attitude in one sentence.
Becuase it is also possible to see the new architectures as an exciting new path to explore, a path that not only promises to be the future, but also rewards the curious mind and the learning of some new skills, both intellectually and in terms of achieved results.

 

[pegisys]

@ archangelmorph

I think you are missing his point, you yourself said this

If you wanted to build an xbox game on the xbox 360 then i'm sure it wouldn't be "harder" or run "slower" than on the xbox, even if you decided to only utilise one core, no multi-threading and basic GPu functionality.. It would probably scorch an xbox version of the game in performance purely because the processor speeds (CPU + GPU) have been drastically increased over the xbox)

considering that the xbox 360 is ~equal to the PS3 PPE core, and looking at how the PPE benchmarked compared to high end single core CPUs, and then take the fact that 3 CPUs doesn't equal 3 times the power. It probably was a single core solution that would have somewhat equal in performance and make developing a lot easier.

I think that is a valid argument with the xbox 360. Maybe not so much with the PS3 since all the power is in the SPEs, but I think he validated his statement with that as well when he said it would mostly pay off only if Sony had the same market share and games were tailored for the PS3

Now lets say MS did come out with a single core CPU that offered about the same performance that the xbox Cpu offers now, it bring up 2 possibilities:

one, that is kinda like what we are seeing now, were MS gets a good piece of market share and the multi-platform games are being made for the 360 and ported to the PS3, I think the games would end up being worst then the ports we are seeing now, since it would go from easy single core development to having to write parallel code to just gain similar performance, and if they are just trying to get the game out the door it's not going to be optimal

then the other would be Sony dominates the market again and all the games are tailored to the PS3 and then ported to the xbox 360, which would probably lead to downgrades in any highly parallel CPU intensive task. This scenario may or may not still be true, and I think this is the reason MS went with the CPU they did because it would be easier to port PS3 games to the xbox 360 not because it's the future, though it may be

 

[Arwin]

This is the fundamental misconception, it's certainly a learning process, but there is no such thing as "cracking it" it's a fundamentally hard problem.

Maybe. But didn't we all think the same back when we made our first recursive routines, or when we first started proper modular programming, using classes and headers, and so on?

it will always be significantly more difficult.

I really don't agree.

Multiprocessor architectures are inevitible, using them effectively is a hard problem, and it's one that compounds all of the other hard problems in game development.

Maybe, but the idea of 'cracking it' isn't so far fetched. You just need to change your mindset. As some wise man said some time ago, data-centric thinking is key here.

This is the main thing with technology advancements that move this fast - every once in a while you have to start from scratch, and the more you know and the longer you've worked in a certain way, the harder that gets. And for some developers, that paradigm shift already happened a generation ago.

 

[ector]

Yeah, people have researched ways to program shared-memory multiprocessors since the 60's, and it's still a fundamentally hard problem, especially if you really want to squeeze maximum performance, which of course game programmers want to do.

All he's saying is that ideally the clock speeds should continue to increase exponentially, but since that is not possible in the real world, he has to adapt, and just like anyone else, he doesn't like it. Every programmer in the world would prefer a single infinitely fast CPU vs an infinite number of slow CPU:s.