Carmack again on PS3-360

[SPM]

Looks to me like Carmack is simply loathe to rewrite his game engine/game code to take advantage of multi-threading, because he has something that works and doesn't need to. Who can blame him, it is a lot of work. He is simply expressing his view from the position that he is in now: it is not worth him/his company spending time re-coding to take advantage of multi-threading given resources/finances/cost-benefit. For companies writing game engines/game code from scratch or recoding their game engines/game code anyway, the same factors don't necessarily apply. Why are people trying to turn Carmack's remarks about personal views of his conpany's internal issues into a panacea?

 

[Blazkowicz]

Well Carmack basically was the first to get his hand dirty and try some multi-threading in a game, it brought something like 10-20% performance to the somewhat limited market of people playing quake 3 on dual pentium pro workstations with 3D cards (at the price of hard work and less stability). similar work was done on the doom3 engine (the quake 4 on dual core thing)

 

[Ike Turner]

Well Carmack basically was the first to get his hand dirty and try some multi-threading in a game, it brought something like 10-20% performance to the somewhat limited market of people playing quake 3 on dual pentium pro workstations with 3D cards (at the price of hard work and less stability). similar work was done on the doom3 engine (the quake 4 on dual core thing)

And Doom on the Atari Jaguar , no ?

 

[assen]

Looks to me like Carmack is simply loathe to rewrite his game engine/game code to take advantage of multi-threading, because he has something that works and doesn't need to. Who can blame him, it is a lot of work. He is simply expressing his view from the position that he is in now: it is not worth him/his company spending time re-coding to take advantage of multi-threading given resources/finances/cost-benefit. For companies writing game engines/game code from scratch or recoding their game engines/game code anyway, the same factors don't necessarily apply. Why are people trying to turn Carmack's remarks about personal views of his conpany's internal issues into a panacea?

The problem is not that he needs to rewrite. You're making it sound as if it would be easy only if he was starting from scratch. The problem is writing solid multithreaded code, even on a clean slate, and than keeping it solid as you (and the 20 programmers) constantly rummage around in it.

 

[JardeL]

carmack is 1 hell of a programmer, its amazing how some of u can sit here and trash him

Yeah he is , but why cant we criticise him ?.. Whenever i read his comments he's always complaining ... And he contunie doing that...

I dont know if any other developer/programmer complain this much ...

When you see " Carmack on PS3 , multi threading etc. " on subject , you can guess whats inside ...

I dont know whats he doing [ except complaining ] now but if he continue this , i think he will be an old-gen 1 hell of a programmer ...

 

[dobwal]

Focusing on adding performance via increasing complexity/speed was not working. The decision to go multi core was forced on them.

Core 2 isn't a bunch of Netbursts cobbled together. Intel boosted the performance by increasing IPC on its cpu. Core 2 isn't as fast Ghz wise, but it has always shown the ability to OC very well and seems destined for official GHz specs beyond the capability of any retail level Netburst cpu without being power hungry.

Going multi-core was forced on them but that doesn't mean going forward, MS or Sony is forced to focus more on core additions than on core complexity/speed.

Remember, there the presentation of the 6Ghz Cell so speed in not a limitation.

 

[Reverend]

Yeah he is , but why cant we criticise him ?.. Whenever i read his comments he's always complaining ... And he contunie doing that...

I see nothing wrong with complaints.

Other than it displeases fans of a particular console, especially if such a fan don't know the scales involved in creating a multi-platform game.

I dont know if any other developer/programmer complain this much ...

Maybe it's because there aren't that many developers being interviewed...?

When you see " Carmack on PS3 , multi threading etc. " on subject , you can guess whats inside ...

Can you tell me if he is absolutely wrong with the following :

<talking about PS3> and the extra theoretical CPU power of the cells isn't going to translate into improved games.

The bolds are mine because I think that's the main focus, right?

I'm not "defending" John but I happen to think all of his complaints about the PS3 are valid ones.

How can you improve discussions beyond simply complaining about John complaining a lot about the PS3?

 

[dobwal]

Where did I write that? I really tried to keep it balanced by adding stuff like "You'll find more playstation centric developers praising or complaining about different aspects of the latest generation of consoles and their comments should be viewed in a similar way". I am not sure I can comment on the view of any person without it being "ad hominem" attacks by your standards.

I would say that one obvious reason for Cormacks praise of Microsofts tools and downplay of Sonys hardware efforts is that he don´t want to bite the hand that has fed him through out the years. Remember he was on the advisory board for the original Xbox as well.
Microsoft has supplied the vehicle(mainly windows) for his games and he wants it to continue to be that way. Microsft will supply similar APIs for the 360 and PC, very convenient for him. If the 360 becomes the winner in the "console war", that means good business for him.

You seem to be implying that Carmacks comments are due to his relationship with Microsoft. You seem to be downplaying his arguments by infering bias on his part thats classic ad hominem. The rest of the statement is general and doesn't apply specifically to you. Its just that recently anytime a developer makes a statement that doesn't vibe with everyone, out comes the character attacks.

This does not really make sense to me. Once you've gone beyond 3-4 cores the level of added complexity does not change that much with each new core, but it may let you add one more fully animated close-up character with a complex AI to your scene..

I don't see how devoting a whole core to animating and implementing one character is efficently leveraging the potential peformance provided by additional cores. To take advantage of additional cores you have to assign them tasks that previously were handled within the original number of cores. The more cores you add the more the more parallelism you have to add to your code. If you plan trying to leverage a large percentage of performance from those additional cores, you going to do more than assign them menial tasks.

 

[blakjedi]

Id and Carmack have been staunch OpenGL/Nvidia supporters for a long time. I dont see how how his general enthusiasm for MS' symmetric multi-core DX API system is based on not biting the hand that feeds him.

 

[3dilettante]

Other than it displeases fans of a particular console, especially if such a fan don't know the scales involved in creating a multi-platform game.

Is that the context of Carmack's statements with regard to the (next/current? what are they called after being released?)-gen consoles?

Is it his worry that the costs of developing for a theoreticaly more powerful, but also more difficult architecture, make excessive demands on either limiting the market (going exclusive) or limiting the product (sub-optimal code/features)?

The bolds are mine because I think that's the main focus, right?

Can you paraphrase a little more of the surrounding context? I don't doubt that you have the main point, but you've quoted about half a sentence. I figure there may be some nifty points either immediately before or after that snippet.

 

[nelg]

I understood his comments to be more towards the difficulties with multi-core in relationship to development teams. I don't doubt for a minute that he, and many others, are more than capable of utilizing multiple cores, it is that debugging becomes much more of a bitch. IMHO debugging such finely threaded aps will not only require everyone to be one the same page but on the same word.

 

[3dilettante]

I understood his comments to be more towards the difficulties with multi-core in relationship to development teams. I don't doubt for a minute that he, and many others, are more than capable of utilizing multiple cores, it is that debugging becomes much more of a bitch. IMHO debugging such finely threaded aps will not only require everyone to be one the same page but on the same word.

Then his comments are observations on inevitability, then.

 

[almighty]

carmack is 1 hell of a programmer, its amazing how some of u can sit here and trash him

I think some of the guy's at Naughty Dog would kick his ass IMHO

 

[Crossbar]

You seem to be implying that Carmacks comments are due to his relationship with Microsoft. You seem to be downplaying his arguments by infering bias on his part thats classic ad hominem.

If that is ad hominem, so be it. I did not say that that it was the only reason though. And just to set the record straight, I am a big fan of his work and used to enjoy reading his "blog" during the Doom3 development. Hell, the man is a bloody renaissance man, building space rockets in his spare time and everything. However, I reserve myself the right to question his opinion about multi-core development as I most likely have quite a few years more of experience from computer systems distributed over several CPUs than what he has.

I don´t think that interview gives the complete story. I don´t think that John is that naive that he really believed that sticking to single core CPUs was really a viable option for the latest generation high end consoles, where raw performance/die space is crucial. Within 1-2 years it will be standard in the PC space as well, within 3-5 years you will likely have asymmetric CPUs in the PC as well, either Cell style or through integration of GPUs. Sooner or later he would need to bite the bullit. The current situation was inevitable. From that point of view I can see that a lot of people think his comments are a bit strange, but as many also have pointed out they are probably coloured by his current business situation, with the risks that come from introducing new technology.

I don't see how devoting a whole core to animating and implementing one character is efficently leveraging the potential peformance provided by additional cores. To take advantage of additional cores you have to assign them tasks that previously were handled within the original number of cores. The more cores you add the more the more parallelism you have to add to your code. If you plan trying to leverage a large percentage of performance from those additional cores, you going to do more than assign them menial tasks.

Maybe you missed the end of my post where I refered to the job manager of NT as described here. Do you think one more core would be anything but a benefit to them? It may as well add the extra raw power for an extra character as suggested.