Valve Says PS3 Complexity Hinders Game Development

[rated pg]

I still don't get how people criticize Valve for being lazy by not programming for the 3rd place console which also has the lowest software sales. Makes sense to me...?

 

[Bat Vader]

"Hahahahahahahahahaha

Thanks for demonstrating your posts are not worth reading!"

seemed to imply that you had a bias against MGS4 and made it seem as though you thought that Theif trumped everything, and while that's not bad per se, if you've never even played the game that you put down and then say that your opponent probably never played it (which he did [http://www.escapistmagazine.com/forums/read/7.118382?page=4#2267484]), that's hypocrisy my friend.

Is english your primary language?

You even added "seemed to imply" which shows you didn't have a clue what the fuck I was saying but opened your mouth anyway.

I would be a hypocrite if I did what I tell others not to - i.e. judging two games which are imo too different to meaningfully compare.

Here's a hint: read my post which you quoted and read it damn hard. When you get to the point when I said "I haven't played MGS4 but if I ever get a PS3 I will be sure to give it a go." reflect hard on just how bad your reading comprehension is.

How is it that something about Valve and the PS3 has turned into this? Why can't people just accept Valve does not want to program their games for the PS3?

 

[yzzlthtz]

I still don't get how people criticize Valve for being lazy by not programming for the 3rd place console which also has the lowest software sales. Makes sense to me...?

If they had come out and said it like that, I'm sure this discussion would be very different.
But people would still be complaining i guess.

 

[Bat Vader]

Crytek is smaller than Valve, and has done a much better job of getting into the multiplatform game. And, plainly put, Valve just doesn't compare to what the guys at Crytek have done.

Morne

Crytek never developed a game like Portal or half-life. Just because some company can multiplatform something does not make that company good. Crysis was never multiplatform it only came out for the PC and yet they are bringing Crysis 2 to the consoles. If 2 then why not 1 as well?

Valve may anger some PS3 owners because of their decision not to program their game's for the PS3 but we should have respect for their decision.

 

[Zer_]

Question:

Is the PS3 hard to develop for, or is it different?

So many devs. fight at each other saying "The PS3 was easy to develop for" while the people at valve can't stand it's architecture.

Sorry if someone already answered this for you (probably).

But it's different.

Yup, different.
I've also heard that pc architecture is arguably more complicated, especially considering you have to account for any wild combination of hardware.

Actually it's not. The only significant incompatibilities come with different video cards or sound cards. Now, when it comes to the video card market the card's drivers does a large portion of the hard work for you.

When graphics hardware supports Unified Shader Model, it can make sense to design its computational units so that any of them can run any type of shader. When computational units are "unified", that is called Unified Shading Architecture. Most often such hardware is composed of an array of computing units and a dynamic scheduling / load balancing unit that distributes shader work to the computing units.

Hardware is not required to have Unified Shading Architecture to support Unified Shader Model, and vice versa. A Direct3D 10 level hardware can still have dedicated vertex, geometry and pixel processors (although because of very similar ISA, they would be quite similar on hardware level). A technologically earlier (e.g. Shader Model 3.0) hardware can also have unified architecture, as is the case with Xenos graphics chip in Xbox 360, for example.

Unified Shader Architecture allows more flexible use of the graphics rendering hardware. For example, in a situation with a heavy geometry workload the system could allocate most computing units to run vertex and geometry shaders. In cases with less vertex workload and heavy pixel load, more computing units could be allocated to run pixel shaders.

The Cell is a pretty unconventional design when it comes to processors. There is the primary core called the Power Processing Element (PPE). Then there are the eight fully-functional co-processors called the Synergistic Processing Elements (SPE). For the SPE to function correctly, it needs to be assigned a task by the PPE. The PPE found on the PS3 runs at 3.2ghz and is basically your average multi-threaded processor.

The PPE, which is capable of running a conventional operating system, has control over the SPEs and can start, stop, interrupt, and schedule processes running on the SPEs. To this end the PPE has additional instructions relating to control of the SPEs. Unlike SPEs, the PPE can read and write the main memory and the local memories of SPEs through the standard load/store instructions. Despite having Turing complete architectures, the SPEs are not fully autonomous and require the PPE to prime them before they can do any useful work. Though most of the "horsepower" of the system comes from the synergistic processing elements, the use of DMA as a method of data transfer and the limited local memory footprint of each SPE pose a major challenge to software developers who wish to make the most of this horsepower, demanding careful hand-tuning of programs to extract maximal performance from this CPU.

Before the Cell was released, an article mentioned one of the potential problems of the Cell. "This may sound like an inflexible system which will be complex to program but it?ll most likely be handled by a compiler with manual control used if you need to optimize." Of course, one of the first complaints by game developers was the lackluster development kit and poor support.

The article later states that certain compilers will simply not work with the Cell architecture, others will end up with poor performance. The PS3's SDK was released with a shoddy compiler. Much of the code had to be pre-isolated since the compiler did a very poor job. Simply put the PS3 was launched way before the SDK was ready. Early versions of the compiler even had trouble compiling the code to run on the six threads available to the developer. The latter problem affects (you guessed it) multi-platform games. This was largely considered a "dick move" by many developers in that they view it as a vain attempt to force developers into exclusively developing for the PS3.

The basic design philosophy for the Cell is to simplify the hardware significantly (thus reducing the amount of power being used) while increasing the complexity of the software that runs on it. The result is alienating any developers who wish to program a multi-platform game for the PS3 and 360 as well as the PC. Current compilers for the Cell are better then the first generation ones, but they still provide a significant challenge to anyone who wishes to port a game over to the PS3. Let us make an analogy to better understand this.

Each SPEs in the Cell can only consume a small cookie at a time. The Cell cannot take a larger cookie and split it into pieces on its own. The compiler's job includes taking a the program (the large cookie sheet) and splitting it into small cookies that can be eaten by the SPEs. For each cookie to be properly ingested by the SPE the chocolate chunks on the cookie must remain intact. If chocolate chunks are cut by the compiler, then the code becomes buggy and usually results in a bug. Current compilers are not all that effective at splitting the cookie sheet in a way that doesn't damage the chocolate chunks. What that means is that the programmer needs to manually move each of the chocolate chunks in a pattern or formation that makes the compiler's job easier. The PPE in the Cell then takes the compiled cookies and distributes them to the appropriate PPE. If any of the cookies are too big, the performance of the game suffers.

An Intel/AMD CPU is able of eating much larger cookies. This allows for more freedom for the compiler when splitting the sheet into the larger cookies and ultimately results in a lot less chocolate chunks being split. The Intel/AMD CPU is able to take a small piece of the larger cookie itself and assign it to a thread for consumption.

Currently, coding a game from the ground up for the PS3 (one good example being MGS4) is a lot less time consuming then porting a game from another platform to the PS3. Valve makes games primarily for the PC. The 360's cookie cutter is able to cut the PC sheets with very little issues. Unfortunately for the sheet to be cut properly to be consumed by the PS3, the whole sheet needs to be reconfigured.

Hope you guys get it now.

 

[TheBadass]

Eh, tell that to Uncharted 2.

Damn, you get your rocks off with Uncharted 2, don't you?

I really, really do. I think it's Chloe or Nate, but pfft, can't be sure. Hopefully the former, eh?

 

[jymbobjones]

dunno how to reply i love the ps3 and i only sorta like valve (mainly cause of there buggy system where u get there games from dont start a flame war)so i honestly cant say not having l4d on the ps3 hurts my views on the console it just makes me feel that valve are to lazy to work on a game for ps3 where they would get more money and ppl but oh well (ps srsly no flame war not bothering with ppl getting uppety with me)

 

[rated pg]

Counter: Why didn't they program for the PS2 back when it was in first place (and it's still going)

Anser: because it was also difficult to program for the PS2 (Also Gabe Newell and Valve are a damn biased company but thats perfectly fine, so are Naughty Dog, Insomniac and Sucker Punch)

Don'[t get me wrong, I got no issues with valve releasing their games on the 360 and PC exclusivly, My hats off to them for having a fanbase so loyal that they would sacrifice their children to Gabes stomach. This is a purely buisness venture people!

Look Valve also said "Oh the Wii is going somewhere" but haven't made any games for it, why? it's simply

Gabe Newell and Valve are on the damn payroll and their isn't any reason they shouldn't be.

A. it's faster for them to do it on the 360/PC due to similar hardware
B. they are getting PAYED to slander a console and they have millions of followers.

But on that note who trusts THEIR opinion on the 'War' thats like asking Naughty Dog for THEIR opinion on the 360. I garun-fucking-tee that they will slander the console as being weaker and behind the PS3 hardware wise. Because they are on Sony's Payroll and are staffed by former Sony employees!

I said earlier that Game Devs and myself have a very simple relationship, they make video games and I play them, beyond that they can spank a naked monkey and lather it in baby oil or whatever.

EDIT: also it's because they aren't saying that the reason they wont dev for it is because it's in last place but because it's too hard. Those are the words of a lazy bastard not someone concerned over buisness.

http://www.metacritic.com/games/platforms/ps2/halflife?q=half-life

You missed my point, which was I don't think they think it's worth it money wise. And they could very well be right. Plus it could delay HL2:EP3 or other projects.

And I am amused at how everyone who dislikes the PS3 is on someone's payroll. Hasn't anyone ever heard of an opinion before? Sure, there are legitimately some companies Microsoft or Sony have paid large amounts of money for exclusivity/timed exclusivity which may cloud their opinion whether by necessity or otherwise, but immediately attacking a respected industry professional's integrity because of their opinion seems very wrong to me, and sets a bad precedent.

 

[rated pg]

Question:

Is the PS3 hard to develop for, or is it different?

So many devs. fight at each other saying "The PS3 was easy to develop for" while the people at valve can't stand it's architecture.

Sorry if someone already answered this for you (probably).

But it's different.

Yup, different.
I've also heard that pc architecture is arguably more complicated, especially considering you have to account for any wild combination of hardware.

Actually it's not. The only significant incompatibilities come with different video cards or sound cards. Now, when it comes to the video card market the card's drivers does a large portion of the hard work for you.

When graphics hardware supports Unified Shader Model, it can make sense to design its computational units so that any of them can run any type of shader. When computational units are "unified", that is called Unified Shading Architecture. Most often such hardware is composed of an array of computing units and a dynamic scheduling / load balancing unit that distributes shader work to the computing units.

Hardware is not required to have Unified Shading Architecture to support Unified Shader Model, and vice versa. A Direct3D 10 level hardware can still have dedicated vertex, geometry and pixel processors (although because of very similar ISA, they would be quite similar on hardware level). A technologically earlier (e.g. Shader Model 3.0) hardware can also have unified architecture, as is the case with Xenos graphics chip in Xbox 360, for example.

Unified Shader Architecture allows more flexible use of the graphics rendering hardware. For example, in a situation with a heavy geometry workload the system could allocate most computing units to run vertex and geometry shaders. In cases with less vertex workload and heavy pixel load, more computing units could be allocated to run pixel shaders.

The Cell is a pretty unconventional design when it comes to processors. There is the primary core called the Power Processing Element (PPE). Then there are the eight fully-functional co-processors called the Synergistic Processing Elements (SPE). For the SPE to function correctly, it needs to be assigned a task by the PPE. The PPE found on the PS3 runs at 3.2ghz and is basically your average multi-threaded processor.

The PPE, which is capable of running a conventional operating system, has control over the SPEs and can start, stop, interrupt, and schedule processes running on the SPEs. To this end the PPE has additional instructions relating to control of the SPEs. Unlike SPEs, the PPE can read and write the main memory and the local memories of SPEs through the standard load/store instructions. Despite having Turing complete architectures, the SPEs are not fully autonomous and require the PPE to prime them before they can do any useful work. Though most of the "horsepower" of the system comes from the synergistic processing elements, the use of DMA as a method of data transfer and the limited local memory footprint of each SPE pose a major challenge to software developers who wish to make the most of this horsepower, demanding careful hand-tuning of programs to extract maximal performance from this CPU.

Before the Cell was released, an article mentioned one of the potential problems of the Cell. "This may sound like an inflexible system which will be complex to program but it?ll most likely be handled by a compiler with manual control used if you need to optimize." Of course, one of the first complaints by game developers was the lackluster development kit and poor support.

The article later states that certain compilers will simply not work with the Cell architecture, others will end up with poor performance. The PS3's SDK was released with a shoddy compiler. Much of the code had to be pre-isolated since the compiler did a very poor job. Simply put the PS3 was launched way before the SDK was ready. Early versions of the compiler even had trouble compiling the code to run on the six threads available to the developer. The latter problem affects (you guessed it) multi-platform games. This was largely considered a "dick move" by many developers in that they view it as a vain attempt to force developers into exclusively developing for the PS3.

The basic design philosophy for the Cell is to simplify the hardware significantly (thus reducing the amount of power being used) while increasing the complexity of the software that runs on it. The result is alienating any developers who wish to program a multi-platform game for the PS3 and 360 as well as the PC. Current compilers for the Cell are better then the first generation ones, but they still provide a significant challenge to anyone who wishes to port a game over to the PS3. Let us make an analogy to better understand this.

Each SPEs in the Cell can only consume a small cookie at a time. The Cell cannot take a larger cookie and split it into pieces on its own. The compiler's job includes taking a the program (the large cookie sheet) and splitting it into small cookies that can be eaten by the SPEs. For each cookie to be properly ingested by the SPE the chocolate chunks on the cookie must remain intact. If chocolate chunks are cut by the compiler, then the code becomes buggy and usually results in a bug. Current compilers are not all that effective at splitting the cookie sheet in a way that doesn't damage the chocolate chunks. What that means is that the programmer needs to manually move each of the chocolate chunks in a pattern or formation that makes the compiler's job easier. The PPE in the Cell then takes the compiled cookies and distributes them to the appropriate PPE. If any of the cookies are too big, the performance of the game suffers.

An Intel/AMD CPU is able of eating much larger cookies. This allows for more freedom for the compiler when splitting the sheet into the larger cookies and ultimately results in a lot less chocolate chunks being split. The Intel/AMD CPU is able to take a small piece of the larger cookie itself and assign it to a thread for consumption.

Currently, coding a game from the ground up for the PS3 (one good example being MGS4) is a lot less time consuming then porting a game from another platform to the PS3. Valve makes games primarily for the PC. The 360's cookie cutter is able to cut the PC sheets with very little issues. Unfortunately for the sheet to be cut properly to be consumed by the PS3, the whole sheet needs to be reconfigured.

Hope you guys get it now.

Thank you for that. You get a cookie now (since I imagine you probably wanted one, after talking about them for so long). +1

 

[teutonicman]

I haven't seen anything this idiotc in awhile. Why the hell would you not want developers to maximize the PS3's potential immediately?! Does Sony have some kind of master plan to some god send game or something?!

 

[Yopaz]

'hai everybody, i have a great idea for our new system

let's make it really hard to use, really esoteric and unintuitive, so it will reveal more of itself over time as people slowly discover what a pain in the ass it is to develop. the industry will really benefit from a tool that is too complex for anyone to figure out! in the meantime, let's also make it too expensive for the average consumer to purchase. we're really smart, i hope one day people can understand what great geniuses we are'

Oh, but don't forget that people like online play, so let's add a crappy online system that too.

 

[MiracleOfSound]

God I hate console wars.

.

Me too... can't we all just agree that we all like games and both consoles are good in thier own ways?

I love my 360 but when God of War 3 comes out I'll be investing in the big black brick for sure.

 

[rated pg]

Ah right maybe my complete lack of care for Half-Life is the reason I never played it.

And yes Valve is on Microsofts payroll, I don't doubt it my mind that Gabe Newell really DOES dislike the PS3 and means every word he says. he is STILL on the Microsoft payroll just like Naughty Dog is for Sony.

And your point was "why would they dev for the minority market" oh by the way the minority? thats US the Wii is the majority market the PS3 and 360 both have miniscule sales compared to it so thats laughable at best.

I know they don't think it's worth it money wise, but that wasn't what they were saying. What they said was "It's too hard." not "It's not worth it from a buisness standpoint" because a 30% market is a LOT.

And I'll set whatever god damn precedent I want.

Well, if you can prove it then I won't say anything else on the subject. Proving that Gabe Newell himself has literally received money from Microsoft as you just claimed.

Yes but The Orange Box and L4D would never have taken off on the Wii; the online support system isn't there, and neither is the hardware or interest. So they put the game on the hardcore console with the largest user base. Based on sales for shooters on the 360 and online statistics, it was a good move.

And part of why it isn't worth it is that it's too hard. The extra development time versus the end result is likely why Valve decided not to bother.

And 30% is probably quite an overestimation, and even if it isn't you still have to factor out the people who just bought it for blu-ray and look at other shooter's performance on the console (Resistance 2 = 500k after 2 months, Killzone 2 = 2 million after a month and a half, etc)

 

[Vlane]

So we have to go through this again? We already did that like 5 months ago.

 

[GonzoGamer]

Ahh, the fanboys are out in full force this post.

Seriously, who cares. Valve is whinging again that the PS3 is too hard to develop for, old news is old.

That's what I was wondering: what PR purpose does it serve to have some dbag from the company come out and say this every couple of months. At this point (when we're seeing some really nice AAA titles on the ps3) their continued whining is starting to make them look stupid. They either need to put up or shut up at this point.

I put up a poll asking ps3 owners if they even care at this point, and it seems that they don't:
http://www.escapistmagazine.com/forums/read/9.118532?page=1