• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

John Carmack Keynote

Carmack's main complaint is just that you can't port your existing engine to the consoles because it would require some re-design. Boo-hoo. He is an incredibly smart guy and far more experienced than I am, but I don't see what the big deal is. Console programmers are used to multi-threaded programming and multi-processor machines as well as molding the engine around the hardware.

At work, we've created our game engine around the concept of multiple CPU cores while reducing stalls/synchronization to a minimum and have had no problems running it on XB360. All it takes is right design.
 
I know you think Carmack is just a PC coder and no one in the console market should care, but show me a singler developer who has actually written code on a Super NES, 32x, Sega Saturn, N64, Atari Jaguar, Dreamcast, PlayStation, PlayStation 2, Xbox, and now Xbox 360(and I'd be willing to be PS3). Let's not forget all different Computers he has coded on, and his latest hobby cellphone game coding. Sits on the advisory board of nVidia, ATi, Microsoft, Apple, et al.

Sure id and Carmack didn't do all the ports themselves, but if you know anyone who has worked with id you will know they keep their noses stuck deep in the porters business.
 
most of the ports to consoles were done by another company. Doom3 for Xbox iirc was the first time id Software actually developed a console game.
 
DopeyFish said:
most of the ports to consoles were done by another company. Doom3 for Xbox iirc was the first time id Software actually developed a console game.

I thought that was ported by Vicarious Visions. I think id did Doom 64 themselves though.
 
Doom 64 was Midway, I thought. Virtually every version of Doom/Quake/etc. has had their supervision...although I think the god-awful-shitty Saturn version somehow slipped through the cracks...
 
Zeenbor said:
Carmack's main complaint is just that you can't port your existing engine to the consoles because it would require some re-design. Boo-hoo.
Another person misses the point entirely.
If anything, I'm more interested in the opinions of the Japanese devs out there, guys who are already used to learning fundamentally different architectures from one console generation to the next and aren't quite as heavily invested in keeping things relatively static.
You'll probably never get that though. Japanese developers are too polite to critique publically, and too beholden to 1st parties. There will probably only be one or two bigmouths like Mikami or Itagaki that say anything....and Itagaki's already obviously got the X360 cock up his ass if you read the interviews so I'm not expecting anything other than glowing praise from him.
 
Single vs. Multithread Processing

Carmack points out that there could be diminishing returns with the next generation of consoles, thanks to the architecture of the hardware. Although the PS3's Cell processor is powerful, it's still a single-thread processor (unlike high-end PC processors). There's not much more technology can do to crank more speed out of a single-threaded processor. To make up for this deficiency, multiple processors are used: multithreading and parallelism is the way to go. But this isn't easy to do in games! "The returns will initially be disappointing," he explains.

It will take a long time before game developers will figure out how to get the most out of parallel processors, and in the meantime, it's going to make high-end game development more difficult -- "Not a good thing," Carmack says. There's no easy solution to this problem, no special compiler that'll take away the grunt work. "There's no silver bullet for parallel programming."

Sony's position seems to be similar to the company's stance with the PS2: Sure, it'll be hard, but the really good developers will suck it up and figure it out. But Carmack wonders aloud: wouldn't it have been better to use multi-threaded processors to begin with?

Can't be sure, but I think they mangled this bit. I don't think he ever said Cell was single-threaded - in fact he went on later to discuss multi-threads on Cell. I also thought he was wondering why they didn't use OoOE processors ;)
 
breakdown by GameMaster from TXB:http://forum.teamxbox.com/showpost.php?p=5768839&postcount=130

1) Carmack seemed to *REALLY* like the ease of development for the XBox360 as well as the enviroment for console game development being very developer centric.

2) Carmack indicated that the XBox360 has a very thin API layer which can allow for the developer to "Talk directly to the hardware" and thus allowing the developer to make the hardware do exactly what that developer wants.

3) Carmack praised the XBox360 as being BOTH powerful AND easy to develop for, which is in contrast to Nintendo, Sony, and Sega (according to his comments).

4) Carmack stated that the Cell CPU is a SINGLE THREADED processor and that there is not much more technology can do to crank out much more speed out of a single threaded processor. To make up for this deficiency the Cell CPU uses parallelism, but ends up making things more complicated.

5) Carmack stated that initial returns from games using multithreading/parallelism will be limited and disappointing at the beginning and that it will make things more difficult for developers, but eventually developers will figure out how to get the most out of these types of systems.

6) Carmack seems to believe Sony's position with the PS3 is similar to their stance with the PS2 in regards to game development.

7) Carmack went on to refute that graphics have reached their peak and how the extra processing power can be dedicated to things like physics and AI. He relates back to a conversion with an IBM engineer at IBM who told him how graphcis was basically done and how Carmack disagreed with that engineer.

8) Carmack was skeptical about AI being offloaded to another processor and the need for overly complex AI routines.

9) Carmack also tounched on physics and some of the problems with physics in games and their scalability.
... and other comments on various subjects related to game development.

The comments about how the Cell CPU is actually only single threaded (as it appears to the developers anyway) is interesting as I originally believed the Cell CPU was dual threaded (though one of the threads is used for scheduling), but is also at the same time parallel. I suppose my original comments was indeed accurate in this context. He said that the API is very thin in the XBox360 and allows for directly working with the machine... in the PS3 the hardware is divided into 3 layers and then you have the software layers. The developer CANNOT access the hardware directly and must go through these layers in the PS3. Level 0 is the bottom hardware layer and is not disclosed and is kept secure. Level 1 handles the operations close to the kernel such as scheduling (for the SPEs), the real time kernel, and device drivers. Level 2 is the guest operating system layer and this is the only layer that developers can access. This may impede the developer's abilities to get the most out of the PS3 hardware. This arrangement is part of Sony's security implementation with the Cell processor and system's based on it.

Now many of you will say Why did Tim Sweeney say the opposite: Well if you look at all the epic games there are 3 logos that show at the beginning of every Epic game

1) Epic Logo
2) The publisher logo
3) Nvidia Logo (the chip inside Playstation 3)

And is there another example of People preferring Xbox 360 hardware? try the chief development architect of Bethesda:

Finger: The Xbox 360 is quite a unique piece of hardware when compared to modern PCs, since it basically has three CPUs working at all times inside a single chip. What challenges does this multiple-core architecture create for Bethesda and what is it like working with the Xbox 360 versus the PC?

Todd Howard: We see it more of an opportunity than a challenge. Multi-threaded code is not a new concept, you can do it on the PC easily too. But having a console now, that takes advantage of that, and is built for speed around that, it really opens up the possibilities of what you can do. 360 is a great system with great dev tools. We love it.


The real focus here is that the Playstation 3 is a Cell processor geared towards A.I and Physics. Now we know A.I isnt much dependent on CPU power but Physics is. So Physics wise PS3 will perform better (around 5-15% better) than Xbox consoles. But Graphically PS3 has according to most developers left the graphics part to nvidia and we all know the RSX is a deviant from the G70 just like the Xbox GPU was a deviant of the Geforce 3 GTS which is just under the power of the Geforce 4 TI. The Xenos is a varient of Directx10 technology while at the same time not using DX10 it will use some of its features plus the 95% efficiency makes it graphically according to alot of people slightly better than RSX (from the RSX info on paper). In terms of graphics the visual quality of both console in GAMES (not realtime tech demos or CGI) will remain the same but by the 3rd-4th year of PS3 the draw distance will increase for PS3 compared to Xbox 360 by 5-10%
 
hasanahmad said:

Please don't post his tripe here. Carmack did not say Cell was single-threaded for starters. His comments about low-level access on Cell are also presumptuous and erroneous IMO. "TheGamemaster" is a running joke anywhere outside TXB.

hasanahmad said:
The real focus here is that the Playstation 3 is a Cell processor geared towards A.I and Physics. Now we know A.I isnt much dependent on CPU power but Physics is. So Physics wise PS3 will perform better (around 5-15% better) than Xbox consoles.

5-15%? Care to explain how you reached that figure?
 
gofreak said:
Please don't post his tripe here. Carmack did not say Cell was single-threaded for starters. "TheGamemaster" is a running joke anywhere outside TXB.



5-15%? Care to explain how you reached that figure?


well watch the 94 minute video. why dont you download it at filerush.com . its 275 MB
care to prove any of GameMaster points wrong? 5-15% is an estimate. what do you expect, 50%?
 
hasanahmad said:
well watch the 94 minute video. why dont you download it at filerush.com . its 275 MB
care to prove any of GameMaster points wrong? 5-15% is an estimate. what do you expect, 50%?

I did watch the video. Carmack did not say Cell was single-threaded. In fact he held Cell out as being a tougher cookie to crack from a multi-processing/multi-threading perspective. He discussed how you'd have two threads on the PPE and worker threads on the SPEs. His whole point was that these new console chips WEREN'T single-threaded, single core designs, and he was wondering if they might have been better to have been so.

The stuff about low-level access to Cell hardware - there's no indication of how heavy or light the interface on Cell is. Carmack was not talking about Cell when he was talking about interfaces.

Re. your physics estimate, 500% seems as valuable as 5-15% if you're just pulling figures out of the air.
 
The TXB summary has a pretty pro-Microsoft slant to it. Not that that should be coming as any surprise, given the source.
 
border said:
Do you guys actually know what an order of magnitude is or is it just some number that gets thrown around to mean "Uhhh, something really large"?

Carmack notes a 40% speed increase when working with OOE multiprocessing. Hint: 40% != Order of magnitude

Do you actually know how many cores he is working with? :P

If its a PC environment, then multicore means 2, so a 40% increase is pretty good.

I personally used order of magnitude in relation to simple horsepower - 4GFLops vs 100 potential in CELL.
 
Another view, this time from Game Informer's interview with Todd Hollenshead, CEO of id Software.

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

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.
 
m0dus said:
It's funny. Reading the gamespy breakdown, it sounds as if there were two different keynotes entirely:



How can people from at the same conference come up with wholly different slants? He obviously has some reservations about nextgen, but those reservations are not focused only on MS , or seemingly as critical of them as the first post paints him to be . . . :)

All sorts of crazy interpretations, (judiciously) cut-up quotes and so forth have come out of this speech. It's crazy.

I think the most important part that needs mentioning again is the bit where he says that his comments on the console CPUs are "essentially just quibbles", and that everyone's making really great hardware, and he's actually quite content with it all.
 
gofreak said:
Can't be sure, but I think they mangled this bit. I don't think he ever said Cell was single-threaded - in fact he went on later to discuss multi-threads on Cell. I also thought he was wondering why they didn't use OoOE processors ;)

Yes, that summary is non-sense, just like the others. Honestly, I don't think this statement could be more wrong:

though the PS3's Cell processor is powerful, it's still a single-thread processor (unlike high-end PC processors). There's not much more technology can do to crank more speed out of a single-threaded processor. To make up for this deficiency, multiple processors are used: multithreading and parallelism is the way to go.

The TXB summary has a pretty pro-Microsoft slant to it. Not that that should be coming as any surprise, given the source.

So does the speech, but that TXB summary is terrible and inaccurate. As gofreak said, The Gamemaster is a joke to everyone outside of TXB. I'd really like to know where he got some of those comments from, because they sure as hell weren't in the speech.

I really advise everyone to either watch the video or wait for a full transcript. None of the butchered interpretations do it justice.
 
I don't see how anyone (*cough* gofreak/titanio *cough*) can spin JC's keynote any other way...he doesn't like Cell and some of Sony's current dev environment for PS3...and likes MS' X360 CPU design and dev environment more. He still disagrees with both consoles' CPU choice, though. In the end, I'd expect that more devs will like X360's CPU design more, in comparison to PS3 simply by virtue of its symmetric core design -- one that will come across much closer to what the personal computer space will be more commonly like (SMP for Mac/Win/*nix)...making cross platform work easier in some ways. Development environment is a landslide in MS' favor, AFAICT.
 
Shompola said:
Gamemaster is a nice guy.

He could be a saint for all I know, but he still regularly makes gigantic half-technical posts filled with baseless (often inaccurate) speculation that he passes off as fact.
 
MightyHedgehog said:
I don't see how anyone (*cough* gofreak/titanio *cough*) can spin JC's keynote any other way...he doesn't like Cell and some of Sony's current dev environment for PS3...and likes MS' X360 CPU design and dev environment more.

Err, did I take issue with that? I've been discussing his points on physics/ai in games, I can fully appreciate the appeal X360's CPU has for him relative to Cell (a symmetric model, as you point out), and it is easier to work with, sure.

MightyHedgehog said:
Development environment is a landslide in MS' favor, AFAICT.

I'd call Xbox/PS2 a landslide, I wouldn't call X360/PS3 a landslide in the same way ;) There's more to the dev environment than hardware or the CPU.
 
gofreak said:
Err, did I take issue with that? I've been discussing his points on physics/ai in games ...

My bad. I woke up and skimmed the thread with a full-on raging XBOT hard on when I awoke this morning and felt the urge to put someone's name in my post. ;p


I'd call Xbox/PS2 a landslide, I wouldn't call X360/PS3 a landslide in the same way ;) There's more to the dev environment than hardware or the CPU.

Again, my bad...as a poorly educated non-techie. As usual, I'm overly redundant and poorly worded in my posts. I meant that JC, AFAICT from the keynote, seems to feel that way...in a general way.
 
Zeenbor said:
Carmack's main complaint is just that you can't port your existing engine to the consoles because it would require some re-design. Boo-hoo. He is an incredibly smart guy and far more experienced than I am, but I don't see what the big deal is. Console programmers are used to multi-threaded programming and multi-processor machines as well as molding the engine around the hardware.

At work, we've created our game engine around the concept of multiple CPU cores while reducing stalls/synchronization to a minimum and have had no problems running it on XB360. All it takes is right design.



Intresting....any plans for PS3 games??
 
Kleegamefan said:
Intresting....any plans for PS3 games??

From the GI interview

With the next project that we’re working on at id, like Carmack said yesterday, is that we’re already focused on 360, and we’ve got PS3 dev kits, and this will be the first game for id that all three console versions will be worked on internally at the company.
 
Well after reading this thread I noticed a trend. Two types of folks like to post into these types of threads. First I'll describe you so you know who you are. Then I have a couple of questions for you .
1)
I like it when someone like John Carmack, who has proven himself to be one of the few top tier developers, makes comments with completely logical reasons behind them and people on message boards come in and ridicule him. You guys that step in with the ridicule are type 1.

2)
I like it even more when misc developer types chime in and say they have no idea what he's talking about because at their company they got things working and obviously Mr. Carmack has lost it. You people who are working on stuff that blows away everything Carmack has ever done and will ever do are type 2.

Question for #1 types: Where did you learn so much about game engine development? You have enough experience to question someone like Carmack? Wow! I hope that one day you are gracious enough to share some of your skill with the gaming community and make a game that blows us all away!

Question for #2 types: If you guys put together such rocking engines why are there so few AAA engines out there? Why are so many game lacking technically? Why is it that everyone anxiously awaits to see what Team X (Kojima, Ninja, id, Epic, etc) do with Y technology? It's always the same skilled players people look to when there's new hardware...Why aren't you guys in that list?

I'm only half serious here, but I'm really curious what's with all the JC hate and the folks who's statements lead me to believe there are tons of people out there writing game code that runs circles around JC's previous, current, (and since he's lost a step) future stuff.
 
It is curious that he's been in contact with both the Xbox360 and PS3 devkits and he stated that powerwise, the difference will be smaller than current gen. This could be better for Microsoft than for Sony, since they (MS) have the cheaper hardware (apparently) and more straightforward development tools, AND the 6months/1year+ headstart.

Even so, it seems as if multiplatform developing companies will stick more to PS3 as primary development console since it'll be easier to port to the Xbox360 than the opposite.
 
Kleegamefan said:
Intresting....any plans for PS3 games??
Of course, it is in the plans. There are just more XB360 kits available at the moment. I don't know if I can say too much more due to the big ol' NDA.


border said:
Another person misses the point entirely
Would you please elaborate as to how I missed the point of what Carmack was saying? I'm trying to be respectful here.


ddkawaii said:
I like it even more when misc developer types chime in and say they have no idea what he's talking about because at their company they got things working and obviously Mr. Carmack has lost it. You people who are working on stuff that blows away everything Carmack has ever done and will ever do are type 2.
...
Question for #2 types: If you guys put together such rocking engines why are there so few AAA engines out there? Why are so many game lacking technically? Why is it that everyone anxiously awaits to see what Team X (Kojima, Ninja, id, Epic, etc) do with Y technology? It's always the same skilled players people look to when there's new hardware...Why aren't you guys in that list?
I assume you're referring to me, partly or wholly. First, let me correct your gross generalizations/misinterpretations. I never said that Carmack lost it or that I am working on stuff that blows away everything Carmack has ever done. To answer your question partially, not every developer gets 2 1/2 - 4 years of game development time with amount of resources and talent (and not to mention experience) that those teams have.
 
ddkawaii said:
1)
I like it when someone like John Carmack, who has proven himself to be one of the few top tier developers, makes comments with completely logical reasons behind them and people on message boards come in and ridicule him. You guys that step in with the ridicule are type 1.

And I like it when someone comes along and equates any disagreement with John Carmack's opinions to 'ridicule'. Saying 'he's coded some impressive single-threaded 3D engines for PC's, but I don't think this necessarily makes him an expert on the new platforms appearing now' is ridicule?

ddkawaii said:
Question for #1 types: Where did you learn so much about game engine development? You have enough experience to question someone like Carmack? Wow! I hope that one day you are gracious enough to share some of your skill with the gaming community and make a game that blows us all away!

Here we go with this bullshit again.

Yes, I am questioning Carmack. I'm putting that in no uncertain terms. I don't agree with his take on the new hardware, plain and simple. You don't need to be a master of game engine development to know that coders tend to become attached to their platforms of choice, and that this usually becomes more pronounced as they get older, no matter how brilliant they may be. Unlike Japanese console developers, who are forced to learn new, nonstandard hardware architecture every few years, Western PC devs have been living in a sort of 'comfort zone' for decades. The x86 PC architecture, while not static, is something that's been incrementally extended over the years, and the guys coding for it seem to have grown accustomed to that steady progression. Now, with the PC gaming audience continuing to dwindle, formerly PC-centric devs are finding themselves increasingly forced to look to console development in order to stay afloat. And with this shift, they're running up against the same thing their Japanese counterparts have been resigned to dealing with for years now - a mandatory shift to new hardware that's rendering their legacy code and the techniques they've honed over decades useless. And they don't like it. Gabe Newell doesn't like it. John Carmack doesn't like it. While the Japanese seem to be heaving a collective sigh and putting their nose to the grindstone like they've done with new hardware in the past, these big-name Western PC devs are rattling off a litany of reasons why it's going to be so hard for coders to adjust to developing for these new platforms, to the point where it almost seems like a plea for console hardware manufacturers to abandon their plans for this gen, or for coders to shift their focus back to the more 'developer-friendly' PC platform. So no, I don't see Carmack (or Gabe Newell, for that matter) as some paragon of detached perspective, and I'm inclined to take his remarks about the potential drawbacks of the new platforms with a large grain of salt. IMO, he's shown a predisposition to dwell on the negative, for the reasons I've explained above. If you still can't see why I feel the way I do, or want to keep insisting that I'm not qualified to have an opinion until I've written a top-selling FPS myself, I don't know what else I can tell you.

I'm only half serious here, but I'm really curious what's with all the JC hate and the folks who's statements lead me to believe there are tons of people out there writing game code that runs circles around JC's previous, current, (and since he's lost a step) future stuff.

Again, if you're referring to my posts, I don't consider disagreement to be hate. I'm not going around shouting, 'Carmack is a talentless hack!' - I'll be the first to admit that he's gifted when it comes to coding fast, visually stunning engines. I just don't believe that's a good enough reason to worship the ground the man walks on, or to assume that everything he says about coding should be taken as unassailable fact. :p
 
Vince said:
Uh, the question was rhetorical. The answer is an emphatic No, as we've already seen Cell outpreforms a generic x86 OOOE core at computationally expensive tasks by a very wide margin, these tasks are a specific subset that are mostly mappable to those functions used in games, such as it's fast fourier transform rate, MPEG decoding ability, raycasting|tracing, etc. The problem is that unlike the more generalized x86 cores we're used to, the current generation console microarchitectures are specifically geared towards running only a subset of the tasks the former encounter and, thus, you need to alter your development strategies. Preformance wise, by any meaningful metric (eg. flop/watt, flop/gates, flops/mm2, etc) the console CPUs utterly outclass x86.

Cell was benchmarked at over 50 times the normalized preformance of a G5 running IBM's raycasting based Terrain Rendering Engine.

In addition, Cell is hardly a "Very, very cheap" ASIC. In fact, much to the contrary, it's an extremely large IC (250mm2 vs a Pentium4's 112mm2) that's fabricated on a more expensive PD-SOI substrate. Your comments are entirely fallicious and unfounded, in fact, they are plain dumb as fuck.

G5's dont run x86 code....
 
Alright so everyone here except for Tellaerin is focusing on what John Carmack *has actually done* in terms of writing multi-threaded game engines, and Tellaerin is basing all of his arguments on him pretending that he is John Carmack's brain.

Bravo.
 
Azih said:
Alright so everyone here except for Tellaerin is focusing on what John Carmack *has actually done* in terms of writing multi-threaded game engines, and Tellaerin is basing all of his arguments on him pretending that he is John Carmack's brain.

Bravo.

As opposed to pretending that John Carmack is utterly devoid of any personal biases that might be coloring his outlook, despite being a human being the last time I checked.

'Bravo', indeed.
 
Tellaerin said:
As opposed to pretending that John Carmack is utterly devoid of any personal biases that might be coloring his outlook, despite being a human being the last time I checked.

'Bravo', indeed.

Strawman, no one has said John Carmack isn't human or utterly devoid of biases of every sort.

The thing here is that you are ASSUMING *why* he is biased towards single core as opposed to multi-core cpus and using that huge logical leap as a foundation for all of your arguments. You are not his brain.
 
Azih said:
Strawman, no one has said John Carmack isn't human or utterly devoid of biases of every sort.

The thing here is that you are ASSUMING that he is biased in the case of multi vs single core cpus and using that huge logical leap as a foundation for all of your arguments.


yea lets just assume someone with more knowledge and experiance than myself is biased cause i disagree with him.
 
tellaerian, you do realize your argument is nothing more than a thinly veiled ad hominem attack against Carmack, right? Just pointing it out to you.
 
Tallearin said:
Yes, I am questioning Carmack. I'm putting that in no uncertain terms. I don't agree with his take on the new hardware, plain and simple. You don't need to be a master of game engine development to know that coders tend to become attached to their platforms of choice, and that this usually becomes more pronounced as they get older, no matter how brilliant they may be. Unlike Japanese console developers, who are forced to learn new, nonstandard hardware architecture every few years, Western PC devs have been living in a sort of 'comfort zone' for decades. The x86 PC architecture, while not static, is something that's been incrementally extended over the years, and the guys coding for it seem to have grown accustomed to that steady progression. Now, with the PC gaming audience continuing to dwindle, formerly PC-centric devs are finding themselves increasingly forced to look to console development in order to stay afloat. And with this shift, they're running up against the same thing their Japanese counterparts have been resigned to dealing with for years now - a mandatory shift to new hardware that's rendering their legacy code and the techniques they've honed over decades useless. And they don't like it. Gabe Newell doesn't like it. John Carmack doesn't like it. While the Japanese seem to be heaving a collective sigh and putting their nose to the grindstone like they've done with new hardware in the past, these big-name Western PC devs are rattling off a litany of reasons why it's going to be so hard for coders to adjust to developing for these new platforms, to the point where it almost seems like a plea for console hardware manufacturers to abandon their plans for this gen, or for coders to shift their focus back to the more 'developer-friendly' PC platform. So no, I don't see Carmack (or Gabe Newell, for that matter) as some paragon of detached perspective, and I'm inclined to take his remarks about the potential drawbacks of the new platforms with a large grain of salt. IMO, he's shown a predisposition to dwell on the negative, for the reasons I've explained above. If you still can't see why I feel the way I do, or want to keep insisting that I'm not qualified to have an opinion until I've written a top-selling FPS myself, I don't know what else I can tell you.

Do remember, that Western Developers are usually far better at creating engines than Japanese developers. Japanese developers main strengths lie in art direction and execution, whereas American/Western Developers strengths lie in creating engines (in all aspects, graphics, sound, networking, etc.).

So I would take these Western Developers' opinion far more seriously than any Japanese developer.

Who's word do you take over John Carmack's? Who would you trust otherwise? Who do you claim is not attached to any hardware?

Your lack of answers to these questions are very telling about how much you know about the issue at hand.
 
Azih said:
Strawman, no one has said John Carmack isn't human or utterly devoid of biases of every sort.

The thing here is that you are ASSUMING that he is unreasonably biased in the case of multi vs single core cpus and using that huge logical leap as a foundation for all of your arguments.

As I mentioned above, I see a pattern developing here. So far, it's been two big-name Western developers (Carmack and Newell) that are heavily invested in PC development doing the all the griping about the new multicore machines. I'm not hearing this sort of doom-and-gloom rumbling from the Japanese dev community. Hell, it's not even a universal response on the part of Western developers. Mark Rein and the guys at Epic don't seem to be angsting over the new hardware--if anything, they seem enthusiastic. And they've been getting results--impressive results, even at this early stage of the game. (UE3 was designed with multithreading in mind, something that doesn't seem to have proven insurmountable to the Epic crew.) That alone is enough to lead me to believe that this isn't just a case of Carmack and Newell 'seeing farther' than their contemporaries, of the two of them perceiving pitfalls that no one else can. So I'm forced to wonder: Why are Carmack and Newell playing up the potential difficulty of coding for the new consoles, when other developers--ones that aren't venerated as coder gods--don't seem to be running into anywhere near the kind of problems you'd expect after hearing those two talk? And the one constant I see there is that the two of them are partial to the PC as a development platform - are still invested in it, both personally and professionally - to a degree that a company like Epic isn't.

You're entitled to draw whatever conclusions you like from Carmack's keynote. I feel there's an element of bias there that's coloring his perception of the new platforms, and that he's (unintentionally) overstating the potential difficulties of developing for multicore processors. Considering that most of the counterarguments I've heard here so far have amounted to little more than, 'He's a genius, so we must all defer to his opinion', I'm inclined to stand by my position.

Fight for Freeform said:
Do remember, that Western Developers are usually far better at creating engines than Japanese developers. Japanese developers main strengths lie in art direction and execution, whereas American/Western Developers strengths lie in creating engines (in all aspects, graphics, sound, networking, etc.).

So I would take these Western Developers' opinion far more seriously than any Japanese developer.

Just to name an example off the cuff, I was impressed with Team Ninja's Xbox output this generation, myself. AFAIK, their engines were developed in-house, and were pretty damn impressive technically, not just from an artistic standpoint. Japanese developers also tend to eschew middleware solutions in favor of coding engines for their games from scratch as a general rule, inefficient as that approach may be. So while you may feel the word of any given Western developer supercedes that of a Japanese dev, I don't share your opinion. I think I've just explained my reasons for that in satisfactory detail.

Fight for Freeform said:
Who's word do you take over John Carmack's? Who would you trust otherwise? Who do you claim is not attached to any hardware?

Well, I just mentioned Epic above - they've positioned themselves as a major middleware provider for PC and next-generation console hardware, and I daresay they'd have a little more perspective on things than Messrs. Carmack and Newell. (At the very least, I'd take the word of the guys at Epic over the aforementioned gentlemen.)

Fight for Freeform said:
Your lack of answers to these questions are very telling about how much you know about the issue at hand.

I obviously know more than you've credited me for. About the only thing that's telling about that last sentence is that you're being unnecessarily patronizing and arrogant.
 
Tellaerin said:
Just to name an example off the cuff, I was impressed with Team Ninja's Xbox output this generation, myself. AFAIK, their engines were developed in-house, and were pretty damn impressive technically, not just from an artistic standpoint. Japanese developers also tend to eschew middleware solutions in favor of coding engines for their games from scratch as a general rule, inefficient as that approach may be. So while you may feel the word of any given Western developer supercedes that of a Japanese dev, I don't share your opinion. I think I've just explained my reasons for that in satisfactory detail.

From a technical standpoint, TN's engines didn't even compare to Starbreeze's games. So even the best example you could come up with doesn't even compare.

Well, I just mentioned Epic above - they've positioned themselves as a major middleware provider for PC and next-generation console hardware, and I daresay they'd have a little more perspective on things than Messrs. Carmack and Newell. (At the very least, I'd take the word of the guys at Epic over the aforementioned gentlemen.)

Have they said anything that contradicts Carmack's opinion on multicore CPU's? I don't remember anything of that nature...

I obviously know more than you've credited me for. About the only thing that's telling about that last sentence is that you're being unnecessarily patronizing and arrogant.

I don't consider it unnecessary... :D
 
Tellaerin said:
Why are Carmack and Newell playing up the potential difficulty of coding for the new consoles, when other developers--ones that aren't venerated as coder gods--don't seem to be running into anywhere near the kind of problems you'd expect after hearing those two talk?

Because Carmack has always been pushing limits, whereas most other developers don't (at least not on the same scale). John Carmack is looking at hardware like 'I should have enough processing power to do this specific lighting model nobody's ever tried in realtime which makes my scenes look more dynamic and blabla' whereas most other developers look to improve on the current features like 'great my engine runs at a higher resolution and I can even do more blurring'. Now this is very stereotypical I know, but John Carmack is like a scientist in the game industry. He invents new algorithms, he pitches ideas to gpu manufacturers, he is not your average game engine developer.

Another thing is (concerning multiprograms): Carmack is being totally honest. He's in a position do be like that. (And really, Epic was a terrible example because their engine is with the PS3 devkit, of course you won't see them say 'oh man the PS3 really is a bitch to program, but that aside) Sony basicly provides developers with hardware with the potential performance of a supercomputer and tells the developers to get as much power out of it as they can. It's been known for decennia that parallel programs are very difficult to design/write. It's pretty hard to make use of multiple threads because of the current game engine design. It requires quite some efford to write a multithreaded game engine, and even then some parts of the engine aren't suited for paralizing. So there the pc/2 statement comes in, which doesn't mean xenos (in this case) is about half as fast as a normal computer. It means that games written for pc's (+- 3 ghz) will run at appox. half the speed on the x360. I think with about 2 or 3 threads the performance meets the pc performance. That many threads is doable, since game logic and rendering aren't that much tied to eachother. But on PS3 you have 7 units all of a sudden and that's quite a different story. So to meet the performance of a modern pc you need to do a lot more efford. But the worst thing is, you can't debug multithreaded code.
 
Apenheul said:
Because Carmack has always been pushing limits, whereas most other developers don't (at least not on the same scale). John Carmack is looking at hardware like 'I should have enough processing power to do this specific lighting model nobody's ever tried in realtime which makes my scenes look more dynamic and blabla' whereas most other developers look to improve on the current features like 'great my engine runs at a higher resolution and I can even do more blurring'. Now this is very stereotypical I know, but John Carmack is like a scientist in the game industry. He invents new algorithms, he pitches ideas to gpu manufacturers, he is not your average game engine developer.

At the same time, he's a rubbish game designer.

Lets not kid ourselves; he has a large influence on the games that Id churns out...

if he says, I want to use the graphical technology that I'm making to churn out an edgy, shit your pants thriller, then thats what Id invariably ends up doing.

If he's decrying the flexibility that the CELL provides at the expense of say... more dedicated graphics processing power, then he can be safely ignored in that sense, as, as good a developer as carmack is... he's probably not all that great at knowing what games need to make them great...

or rather he doesn't understand the meaning of balance between graphics, AI, physics, etc, in producing an exceptional gaming experience.
 
Fight for Freeform said:
From a technical standpoint, TN's engines didn't even compare to Starbreeze's games. So even the best example you could come up with doesn't even compare.

The lighting effects and normal-mapping in Riddick were very impressive, yes, but on the other hand, it's not like the engine had to handle huge draw distances, either. I'm just as impressed by TN's ability to render large, complex environments at rock-solid framerates (and without the occasional lapses into lo-res grainy-vision you'd get from time to time in Riddick, either). Now, do you really want me to sit here and get into which other Japanese devs I feel have managed to get impressive performance out of consoles this gen (particularly on the PS2), or can we set aside the 'Western engine coders > Japanese engine coders' thing as the gross and inaccurate overgeneralization it is?

Fight for Freeform said:
Have they said anything that contradicts Carmack's opinion on multicore CPU's? I don't remember anything of that nature...

When Coder A's saying, 'Hey, we're getting all these good results from this new hardware and we're really psyched about what we can do with it,' and showing impressive results, while Coder B's saying, 'There are all these inherent pitfalls to writing code for these machines that academics haven't been able to solve for years now and I think we're moving to this new architecture too soon', I consider that a contradiction. You might be able to reconcile the two to your satisfaction after some mental gymnastics, but my personal opinion is that it's a big stretch.

Fight for Freeform said:
I don't consider it unnecessary... :D

I'm sure you don't, which is pretty telling.
 
Apenheul said:
I would agree he's not much of a game designer, but that's not really the issue right now.

It kinda is; if he believes the focus of games should be on prettier prettier graphics, without balance...

and his comments on the new consoles reflect that...

Then isn't that what we've kinda been arguing over?
 
Tellaerin sorry if I offended you with my comments -- I didn't mean to. I was more talking about an orverall tone whenever I read a Carmack quote anywhere (GAF, B3D, /.). BTW- you seem to know alot of Japanese devs are you currently working in Japan now or for a big Japanese company working on next-gen stuff?
 
Top Bottom