ps3 easier to program for

Status
Not open for further replies.
N

ne0-reloaded

[H]ard|Gawd
Joined
Jul 1, 2003
Messages
1,216
Not sure if this has been posted, but some developers are saying that programming for the ps3 is just like programming for the pc, thanks to Open GL ES. hopefully this means better games initially

"Some new information regarding the process of programming for the Playstation 3, suggests that we may have been misled about the difficulty of writing for Sony's new console.

PS3 developers are beginning to surface and are claiming that developing for the Cell powered console is, in fact, very similar to creating content for the PC. It appears that PS3 utilises a cut-down version of Open GL called Open GL ES as its graphics API, making programming for the console a fairly straight-forward process. Two major PS3 developers have recently commented on the process of content creation for Sony's console and both claim that comments released, around XBox 360 launch time, suggesting that PS3 was hard to program for were completely unfounded.

Guerilla Games, developers of the exciting Killzone franchise and Volatile, creators of the Zombie-fest Possession for PS3, commenting to Playstation.com and the Guardian Unlimited: Gamesblog respectively, both agree that the use of Open GL ES makes programming for PS3 much easier than programming for PS2 ever was.

Volatile's lead PS3 programmer, Lyndon Homewood goes even further and suggests that the use of Cg, a version of the C programming language modified to make programming for graphics chips easier, makes working on PS3 similar to creating for PC. At the end of the day it's just a multi-processor architecture. If you can get something running on eight threads of a PC CPU, you can get it running on eight processors on a PS3 - it's not massively different. says Lyndon.

Mr. Homewood does go on to make some direct comparisons between X360 and PS3 and does claim that each console will have its unique benefits and drawbacks. One example the Volatile programmer uses has to do with the way the PS3 accesses video memory, ...the main processor can access all the machine's video memory, but each of the seven SPE chips has access only to its own 256k of onboard memory - so if you have, say, a big mesh to process, it'll be necessary to stream it through a small amount of memory - you'd have to DMA it up to your cell chip and then process a little chunk, then DMA the next chunk, so you won't be able to jump around the memory as easily, which I guess you will be able to do on the Xbox 360.

Lyndon does go on to say however that PS3, mainly thanks to the enhanced capacity of its Blu-Ray drives, will be able to offer much more HD content in games, something which may make PS3 a more enticing option in the long run.

So there you have it, as launch time approaches for PS3, more and more developers are willing to step up and defend Sony's console and the programming process involved when creating content for it. Let's hope more developers give us honest accounts of their experiences working with next-generation consoles."

Link
 
cool, thanks for the article, it would be really nice if the PS3 was easyer to code fo than the 360.......... ;)
 
I dont get how what was said in that article makes it much easier to program for? The GPU is basically a 7800 so it makes sense that you would be able to use a version of OpenGL to program for the GPU. That has no effect on the Cell chip however, which is the hard part of programming.

When the developer finally talks about the important bits, the pseudo C code for the Cell, he basically says "If you can do 8 threads on a PC you can do it for Cell"... Well DUH? Of course you CAN utilize the entire cell :rolleyes:. The hard part is actually doing the 8 different threads. I am glad to see that the development tools are good instead of being clunky, but I dont see them mentioning a compiler that optimzes use of the Cell chip. That is what would really make a difference for the PS3.

I am not doing this to try to flame the Cell or the PS3. I think it has great potential...I am merely referring to the article and the fact that it doesn't seem that the PS3 is any easier to code for than we thought. All it is basically saying is that you code for it in a familiar language...not that it handles the multithreading for you.
 
Good news. Though I would like to see Sony's market share slip this generation, the more developers can focus on executing concepts, the better. If PS3 came out on top, but at the expense of developer resources, I can't say any of us benefitted for it.
 
Nomikal said:
lol...I like you Nas quote...thought I was the only fan...
Click to expand...

ya ima big nas fan, i been listenin to ny state of mind all week non stop. u heard bout the def jam signing right?

i never thought about it, but the last poster made a good comment. sony losin market share wouldnt be such a bad thing. when a company stays on top too long, the consumers are the ones who suffer usually. hopefully ms and nintendo could give em a big run for their money this time, cuz last gen it was a SLAUGHTER!!!
 
ne0-reloaded said:
Volatile's lead PS3 programmer, Lyndon Homewood goes even further and suggests that the use of Cg, a version of the C programming language modified to make programming for graphics chips easier, makes working on PS3 similar to creating for PC.
Click to expand...

I think nobody ever said that the graphics chip was going to be the problem coding for the PS3. Even John Carmack said that thank god they used good graphic chips in these new consoles.

The problem is the Cell chip.

Mr. Homewood does go on to make some direct comparisons between X360 and PS3 and does claim that each console will have its unique benefits and drawbacks. One example the Volatile programmer uses has to do with the way the PS3 accesses video memory, ...the main processor can access all the machine's video memory, but each of the seven SPE chips has access only to its own 256k of onboard memory - so if you have, say, a big mesh to process, it'll be necessary to stream it through a small amount of memory - you'd have to DMA it up to your cell chip and then process a little chunk, then DMA the next chunk, so you won't be able to jump around the memory as easily, which I guess you will be able to do on the Xbox 360.
Click to expand...

And this is what Anandtech's article said just before it was removed from the face of the internet. That's because the SPE's only have access to their own memory (which is too small) that they'll need to do a DMA request through the PowerPC core in order for it to do anything.

So imagine 7 mighty SPE's that have crap for memory that'll need to stream data through the PowerPC. Now the PowerPC is handling data for 7 other cores while also doing it's own work. Anandtech said that the response time would be so terrible that it'll make the SPE's nearly useless in real world applications.

Which might explain why Sony said that developers were not to use all 7 SPE's. I forget how many they are allowed to use but by limiting the developers how many SPE's they can use it also reduces the response time and actually makes the SPE's usuable. Though you won't be using all 7 SPE's.

At the end of the day it's just a multi-processor architecture. If you can get something running on eight threads of a PC CPU, you can get it running on eight processors on a PS3 - it's not massively different. says Lyndon
Click to expand...

Given that there isn't anything running on the PC with 8 threads. At least not for gamers or home users. We'd be lucky if we see 2 let alone 8 threads.

Creating software on the PS3 won't be nearly as big of a problem as say porting it. Which is the real problem. The SPE's don't act like a GPU nor a CPU. They are specialized and will require extra work to port what ever code is written.

This is what Gabe Newell said.

And I totally see why Sony wants people to write code that runs on seven SPEs and a central processing unit, because that code is never going to run well anywhere else.
Click to expand...

People gotta remember that Sony has done this before. They always have a bunch of developers who are ready to say anything to make their product sound like the next best thing.

If you're wondering what platform developers will be workign on first it'll be the Xbox 360. It's easier to develope games on the X360 when it's design is similar to that of a PC and even the Nintendo Revolution. Which is good because everyone will benefit from having it made on the X360 first.

Instead of having it made on the PS3 first and port it to console systems that aren't similar in design. What Sony thinks is that developers make games for the PS3 first and hope that the difficulty and cost in porting it will make developers think twice before doing it.
 
Considering Guerilla Games is owned by Sony and the only game they made was a piece of shit and they already lied to us about their next game ( BS E3 presentation) not a single thing these guys say is trust worthy.When I start hearing opinions from true third party devs that are not in Sonys or Microsofts pockets then I'll decide who to believe. Right now the only dev's opinion that holds any weight is Carmacks and he said that the PS3 has the potential to be more powerfull based on the prelimanary specks but the 360 is much easier to develop for.
 
If you're wondering what platform developers will be workign on first it'll be the Xbox 360. It's easier to develope games on the X360 when it's design is similar to that of a PC and even the Nintendo Revolution. Which is good because everyone will benefit from having it made on the X360 first.
Click to expand...

They will work on the platform that publishers pay them to work on and that will be the platform with the biggest (or potentially biggest) userbase.
 
ECM said:
They will work on the platform that publishers pay them to work on and that will be the platform with the biggest (or potentially biggest) userbase.
Click to expand...
True but it's not Sony or even gamers who have control in that situation. Especially seeing how big of a jump Microsoft has over the PS3.

Microsoft wasn't stupid when they released the X360 first. They know that when it comes to who'll out sell who if Sony and Microsoft released their systems in the same time frame it'll be Sony hands down.

This gives Microsoft time to build a user base for their Xbox 360. Microsoft is hoping that by the time the PS3 is launched in stores that the Xbox 360 will have just as many sales as Sony will have within a given time frame. Though it will be Sony who'll accelerate faster in sales but that's where the developers will come in.

Developers will have a choice of either making games first on the Xbox 360 or the PS3. Sony doesn't have as big of a bank as Microsoft to "convince" developers to make games first. Though I doubt Microsoft will have to do much convincing since Sony will have a few problems of their own.

Sony is poor. They don't have any money. What ever amount they put on the table Microsoft will just top it. Also considering the cost to develope on the PS3 first and then port it. I've heard reports of it costing in the billions.

Though it's not like Sony is SOL. If they can simply produce enough machines to feed the Sony fans then that'll be more then enough to sway developers and they'll dump the Xbox 360 like a bad dream.

Remember that Sega was a bigger name then Sony in video games and Sony kicked their butt. Simply because the Saturn was too difficult and worked too different from everything else. The Saturn used quad triangle VDP which was considered the future at the time. The Playstation used regular polygons just like the PC and N64 did.

What's even more strange is that the Saturn was in fact more powerful then the playstation. It seems recently people have found out that Sony did in fact lie about the Playstation spec. If you visit Sony's Playstation website you can see that the Playstation puts out less polygons then the Saturn.

Playstation: 180,000 Textured Polygons / 360,000 Flat Polygons
Saturn: 200,000 Textured Polygons / 500,000 Flat Polygons

So who's to say history doesn't repeat with the PS3? After all the Saturn was more powerful and yet all it needed was good programmers to bring out its full power. Sony has lied a number of times about what their hardware can do. Made their system difficult to work on just like Sega did. Had a good reputation like Sega did. Just like Sega they are broke just before their console is launched. Which might explain why the PS3 might cost as much as $500.

At this point I'd say it's up to developers that will make or break the PS3. Much like how it was up to them if Sega would float or sink. Programming for the Saturn was so costly compared to the Playstation that they just got up and said no to Sega and yes to Sony.
 
I agree with dukenukemx and erasmus....the hard part was never programming graphics, it's the cell processor. I think light is simply trying to be brought to the situation as a whole..............nevertheless, I will still be getting a PS3 for MGS4 alone!
 
I seem to remember a John Carmack quote from a couple months ago saying the exact opposite - that the 360 would be far easier to write code for.
 
The X360 and the PS3 have pretty much the same main core (PPE = PowerPC Processing Element = stripped down PPC). It's just a matter of finding a way to write for the in-order cores that are cache crippled... There are 3 of these on the X360, and only 1 of them on the PS3. The PS3, however, adds seven almost-useless SPE's to the mix, which are only even remotely useful for media/graphical content... and that's where the difficulty comes in. No matter which way you look at it, neither system is that easy to program for considering their archaic main architecture. But the X360, in any case, will have far less of a learning curve, since whenever they re-learn how to program in-order again for one PPE, they will already know how to use all 3 of the PPE's. Not so with the PS3, as there's still those SPE's, which don't even have a branch predictor.

Bottom line here is that 95%+ of multiplatform titles will likely just use 1 PPE core on each system, making porting extremely easy. They even run at the same 3.2ghz, since they are so extremely similar, though the PS3's has only half the cache... multicore headaches be damned.
 
Erasmus354 said:
he basically says "If you can do 8 threads on a PC you can do it for Cell"... Well DUH? Of course... :rolleyes:.
Click to expand...
No, you're totally off base here. For months, people have been saying just the opposite, that even if you COULD break an application into eight threads, you still weren't home free on the Cell. That the PPEs were far more limiting than the SPE, and therefore multithreading was only one small step to wards utilizing the Cell.

I've been saying for just as long, however, that these claims are exaggerated, and that the Cell won't be substantially more difficult to code for than any other multicore CPU. Glad to see others are starting to see the light finally.
 
masher said:
No, you're totally off base here. For months, people have been saying just the opposite, that even if you COULD break an application into eight threads, you still weren't home free on the Cell. That the PPEs were far more limiting than the SPE, and therefore multithreading was only one small step to wards utilizing the Cell.

I've been saying for just as long, however, that these claims are exaggerated, and that the Cell won't be substantially more difficult to code for than any other multicore CPU. Glad to see others are starting to see the light finally.
Click to expand...

So how am I off base? Notice the all caps CAN. You CAN use the entire cell...it doesn't make it easy to do that though....which was my entire point. Which is exactly what you just said... which makes me come back to how am I off base?
 
Erasmus354 said:
So how am I off base? Notice the all caps CAN. You CAN use the entire cell...it doesn't make it easy to do that though....which was my entire point. Which is exactly what you just said... which makes me come back to how am I off base?
Click to expand...
I'm not quite sure how to rephrase it any simpler. The developer was countering a persistent anti-Cell rumor. Your "duh" remark misses his point, that the Cell is-- despite its asymmetric nature-- no more difficult to code for than would be a symmetric multicore cpu.

Make sense now?
 
masher said:
I'm not quite sure how to rephrase it any simpler. The developer was countering a persistent anti-Cell rumor. Your "duh" remark misses his point, that the Cell is-- despite its asymmetric nature-- no more difficult to code for than would be a symmetric multicore cpu.

Make sense now?
Click to expand...

I seem to recall saying the hard part was splitting everything up into 8 different threads. Not that it would be more difficult than programming multithreaded for a multicore CPU. Programming a video game for 8 threads is not simple, no matter how you put it. If you dont have an advanced compiler that can figure it out for you it is still going to be hard to code whether it is harder than a symmetric CPU or not.
 
Erasmus354 said:
I seem to recall saying the hard part was splitting everything up into 8 different threads.
Click to expand...
You're still missing the point. For a symmetric cpu, your statement is true-- split your problem into n parts for n cores, and you're over the hurdle. For one with asymmetric cores, its not neccesarily true. For the Cell, the rumor was just this...that the hard part was NOT splitting everything up, but rather the subsequent mapping of gp code into limited, sp cores...not to mention the overhead of dealing with intra-core communications at a much lower level than a symmetric multicore requires.

The article is saying otherwise. That the rumor is wrong, and that your statement applies to the Cell.

Programming a video game for 8 threads is not simple, no matter how you put it.
Click to expand...
This is a side issue, not related to the main point. However, I will point out that its not nearly as difficult as people are claiming. Use a few cores for advanced physics and AI alone and you can easily use more 8 cores. Current videogames don't use a lot of cycles for this...because they don't have the cycles to spare. With multicore architecture, thats about to change.

Games that don't require this can use SPEs to offload function calls. They won't see the full benefit of the architecture this way...but then again, games like that won't need it either.
 
DukenukemX said:
Especially seeing how big of a jump Microsoft has over the PS3.
Click to expand...
Before the 360 was launched, I was thinking the same thing. MS is going to get quite a big jump on the PS3, especially with their couple million sold in the first 90 days and 4-5m by June. Well, so far they have only sold/shipped ~600,000+ That's a pretty big difference considering Sony sold/shipped nearly a million in the first day of the PS2.

Supply will certainly come into play and the million dollar question is, how much supply can Sony create for launch. If they can get their shipments out then MS's early launch was pointless in terms of getting the larger installbase. MS will still have the advantage with games available and perhaps true second gen games available.

It will be interesting to see if we hear anything at the Playstation Conference in Feb.
 
masher said:
This is a side issue, not related to the main point. However, I will point out that its not nearly as difficult as people are claiming. Use a few cores for advanced physics and AI alone and you can easily use more 8 cores. Current videogames don't use a lot of cycles for this...because they don't have the cycles to spare. With multicore architecture, thats about to change.

Games that don't require this can use SPEs to offload function calls. They won't see the full benefit of the architecture this way...but then again, games like that won't need it either.
Click to expand...

The SPE's can't be used for any of that stuff. AI and physics especially. As I said, they are only useful for such functions as graphics or media processing, which does not require any sort of branching code, a lot of memory, or any significant processing load. All of the physics/AI/game code will be running on that single in-order cache-crippled main PPE core. If they wait until spring 2007, you may be right... but not this year.

Supply will certainly come into play and the million dollar question is, how much supply can Sony create for launch.
Click to expand...

If the rumours about Cell fabrication are true (i.e. they haven't even taped anything out completely), expect them to ship a similar number as the X360 has, or less even. The X360 uses 3 PPE's (3 of the same core), an ATI GPU, and a standard 12x DVD-rom. The PS3 has to worry about 1 PPE, 7 SPE's, an NVidia GPU and a 1-2x BD-Rom. If you think Sony will have a million consoles ready by the end of this year, even in all regions, you're probably wrong. At least I wouldn't put any money in it.
 
steviep said:
The SPE's can't be used for any of that stuff. AI and physics especially.
Click to expand...
Where do people get such misinformation? Do you know anything whatsoever about computational physics? The SPEs are vector processor-- their base unit is a 128-bit vector, broken into 32-bit quads. As such, they're ideal for physics calculations. Far better than a general purpose cpu. And the scatter-gather / SIMD programming model used by the Cell is the same model I used as a graduate student in physics, programming hydrodynamic simulations on massively parallel computers.

Seriously, learn a little about your subject.
 
Steviep and Duke have been here posting about the technical aspects of the PS3 like they program for it and know what they're talking about for months. The best thing to do is just take their posts with a grain of salt, or less.
 
Gnu314 said:
Steviep and Duke have been here posting about the technical aspects of the PS3 like they program for it and know what they're talking about for months. The best thing to do is just take their posts with a grain of salt, or less.
Click to expand...


i agree.. i mean how do you guys "know" so much about the architecture.. i mean you can know about it.. but how do yu know if it's godo for media, physics.. etc... serious question.. like how do you know that?

other than reading what someone else said about it...

so i guess i'm asking.. do you guys really understand the cell's architecture?
 
LynxFX said:
Before the 360 was launched, I was thinking the same thing. MS is going to get quite a big jump on the PS3, especially with their couple million sold in the first 90 days and 4-5m by June. Well, so far they have only sold/shipped ~600,000+ That's a pretty big difference considering Sony sold/shipped nearly a million in the first day of the PS2.
Click to expand...

Sony sold 1 million PS2 concentrating on one region. If MS had concentrated solely on the US they would have sold over 1.5 million 360s (thats how many they have sold worldwide).
 
Considering Guerilla Games is owned by Sony and the only game they made was a piece of shit and they already lied to us about their next game ( BS E3 presentation) not a single thing these guys say is trust worthy.When I start hearing opinions from true third party devs that are not in Sonys or Microsofts pockets then I'll decide who to believe. Right now the only dev's opinion that holds any weight is Carmacks and he said that the PS3 has the potential to be more powerfull based on the prelimanary specks but the 360 is much easier to develop for.
Click to expand...

Exactly...i pay more attention to Carmack and the epic people when they say 360 is easier to work with than i do 2 companies who havent even proven anything yet. Would any person in his right mind ever choose to play killzone over gears of war? I mean seriously...
 
RancidWAnnaRIot said:
i agree.. i mean how do you guys "know" so much about the architecture.. i mean you can know about it.. but how do yu know if it's godo for media, physics.. etc... serious question.. like how do you know that?

other than reading what someone else said about it...

so i guess i'm asking.. do you guys really understand the cell's architecture?
Click to expand...

I have no idea of how useful the SPE's are for Pysics or media. I'm just looking at how it has 8 threads at which one of them is a PPC core. Right there you don't need to go any further. I haven't heard of any developers complaining if the SPE's could or couldn't be useful for gaming.

All I know is that John Carmack said it's a pita to program for and Gabe Newell said that anything you code for the Cell's SPE's will be code that wouldn't work well in other platforms. Plus there's Anandtech's missing article that explains how worthless the SPE's are if they need to do a lot of DMA access. Which even Sony's own developers said that the Cell does have a small problem with it's SPE's doing a lot of DMA because of the small memory that the SPE's have access to.

I mean I have sources to prove what I've said.

Lost Anadtech Article
Gabe on Next Gen
John Carmack on Next Gen

Though it's what Gabe Newell said that's important.
"There are incredibly few programmers who can safely write code in the PlayStation 3 environment. And I totally see why Sony wants people to write code that runs on seven SPEs and a central processing unit, because that code is never going to run well anywhere else. They're saying, "Make your code not run on anything but one of our machines, and we're betting that we'll have market share that's so high that everybody will have to write for our platform. We'll just starve the air from the other platforms by absorbing everybody'd R&D budget and making their code less portable.
Click to expand...

Everybody here seems to be skipping steps here. Before you start talking about the benefits of the Cell there are two questions developers always ask themselves before they begin a project.

1. Can I do it?
2. Should I do it?

Because at some point the question is will it be faster then the Xbox 360? Is it worth the extra effort to use the SPE's? Are the benefits higher then the disadvantages?

If step one and step are a yes then developers will do it. Even Gabe Newell said there are few programmers that can work with the PS3 well. I'll go by there word since I don't have the ability to get my hands a developers kit. They obviously know and aren't biased like Sony's own developers. Especially John Carmack.
 
Can I just say, that in 10 years from now that Anand Tech article will have taken on epic proportions. Already it's the "lost" article, but once the whole 360vsPS3 thing is over and done with, people will still be citing the lost Anand Tech article, and I'm sure that by then the article itself will said to have had the cure for aids as well as the first quantum computer.

Anyways, does anybody know why it went MIA? Is AnandTech in Sony's pocket or what? I mean that was a pretty damning article that spread like wildfire, then one day it was gone. What's the word on that?
 
I have no idea of how useful the SPE's are for Pysics or media...
Click to expand...
The SPEs are ideally suited for physics, DSP, and media processing. They're not well suited for AI and certain other types of code. But what people like StevieP need to realize is that "not well suited" doesn't mean "unable". After all, Pentium class cpus are not well suited to physics nor media, even with SSE intidy tacked onto their instruction set. Yet thousands of such programs run on them daily

All I know is that John Carmack said it's a pita to program for...
Click to expand...
Is that what this is about? Some sort of Carmack fan boy fest? Please remember that, although he's well regarded...he's still just a game developer. And a game developer with no direct experience with the Cell.

In any case, I don't recall Carmack calling the Cell "a pita". I think his remarks were that the Cell had more power than the 360's Xenos, but that it would be harder to exploit. A statement that is true...at least without tool and compiler support for the Cell. However, your sophomoric exaggerations of his statements appear to be wholly unfounded.

Cell does have a small problem with it's SPE's doing a lot of DMA because of the small memory that the SPE's have access to
Click to expand...
The entire purpose of the local store-- and the required DMA transfers between it and main storage-- is to explicitly parallelize code, and to reduce memory latency. On a current processor, a cache miss causes a hard stop for several hundred cycles. A DMA setup takes only a few cycles...and once setup, it proceeds asynchronously. In parallel with code. Done properly, memory latency is far _lower_ than an on conventional processor.
 
MYSTic Jedi said:
cool, thanks for the article, it would be really nice if the PS3 was easyer to code fo than the 360.......... ;)
Click to expand...

It'd be nice if both systems were easy to program for and had great games that stand alone exclusively on the console. At least that's my opinion ;)

I don't need to say anything about Nintendo...they'll be just fine. They've managed to stay in business in one form or another since 1890 :D
 
Spaceman_Spiff said:
Exactly...i pay more attention to Carmack and the epic people when they say 360 is easier to work with than i do 2 companies who havent even proven anything yet. Would any person in his right mind ever choose to play killzone over gears of war? I mean seriously...
Click to expand...

Well, from everything I've read, Epic seems to be really comfortable with developing on the PS3. If you take a look at this gamespot Q&A session with Epic's Mark Rein, you should see he's actually quite happy with the PS3. That article has some good Cell info as well ("We're going to be able to do a lot of cool stuff on it. Especially great physics--")
 
Why does this happen. Everytime there is a thread about coding 1. everyone chimes in like they do it for a living, and 2. it always turns into a vs./you don't know/I've got facts see link threads.

In fact, why do these threads even pop up? Do people search? There is about 1 a week. Do you guys know why I don't get into these (for the most part)? Because I'm off enjoying my 360! And I'll enjoy my ps3 too! You guys should follow suit.

*oh* don't want to forget steviep...he'll be enjoying his Rev. :D
 
To everyone who thinks the "Cell" and the PS3 are hard to code for must look at one thing... The E3 2005 UT007 Demo shown off on it was coded and optimized in 2 months...
 
Lol, I wonder how much this person got paid by Sony to say this? I wonder if it competes with the money they paid Americans to graffiti promoting Sony products?
 
I'll make one last comment. Do any of you *dramatic pause* think for a second *another pause* that Sony would put out a console that was hard to code for? It'll be their third console so I think they know a little something about it, and they have engineers out their ass, that they pay out their asses for to make this stuff efficient.

I'm willing to be that almost everything we hear either for or against Sony/M$ is motivated by either money or hate (or some other biased feeling).
 
Firewall said:
I'll make one last comment. Do any of you *dramatic pause* think for a second *another pause* that Sony would put out a console that was hard to code for? It'll be their third console so I think they know a little something about it, and they have engineers out their ass, that they pay out their asses for to make this stuff efficient.

I'm willing to be that almost everything we hear either for or against Sony/M$ is motivated by either money or hate (or some other biased feeling).
Click to expand...

If I'm not mistaken...

PS1: Easy to code for.

PS2: Hard to code for.

So yeah, they've already put out a console that was hard to code for. Even better, they released it after releasing a console that was easy to code for.
 
At any rate, Playstation 3 !!!!!!s shouldn't get all flush over the idea that the Xenon will struggle on non-graphics code. However bad off Xenon will be in that department, the PS3's Cell will probably be worse. The Cell has only one PPE to the Xenon's three, which means that developers will have to cram all their game control, AI, and physics code into at most two threads that are sharing a very narrow execution core with no instruction window. (Don't bother suggesting that the PS3 can use its SPEs for branch-intensive code, because the SPEs lack branch prediction entirely.) Furthermore, the PS3's L2 is only 512K, which is half the size of the Xenon's L2. So the PS3 doesn't get much help with branches in the cache department. In short, the PS3 may fare a bit worse than the Xenon on non-graphics code, but on the upside it will probably fare a bit better on graphics code because of the seven SPEs.
Click to expand...

Source: ARSTechnica

So how much memory do these SPE's have to use for all the things that Sony promises they'll do? Oh, and tell me now... where do these SPE's get their instructions from? Oh yeah, the 1 main in-order cache-crippled PPE core. Call me a fan-boy, but I'll be enjoying my Revolution, and whichever other system (either the PS3 or the X360) provides me with the games I want to play. I go where the good games are, and am not bound by biased opinions that prevent me for enjoying good games, and never have. Now if you'll excuse me, I'll go and play Shadow of the Colossus on my PS2 now, which was also a big pain for development (not quite as much as the PS3, though).
 
Slartibartfast said:
Anyways, does anybody know why it went MIA? Is AnandTech in Sony's pocket or what? I mean that was a pretty damning article that spread like wildfire, then one day it was gone. What's the word on that?
Click to expand...


I'd like to know that, as well!
 
Mind you that both consoles prior to the PS3 were both developed fully by Sony, This time around sony has tons of middleware developers and Nvidia working on the GPU and helping with OpenGL ES.

What made the PS2 so hard to code for was not the architecture but that lack of good middleware, compilers and documentation on the GPU. That is also why alot of the earily titles suffered horribly from jaggies compared to most today, many developers didn't know how to implement good AA without a huge performance hit.
 
Slartibartfast said:
If I'm not mistaken...

PS1: Easy to code for.

PS2: Hard to code for.

.
Click to expand...


Says who, you? I think there a lot more games available for the ps2 than there is for the ps1.

So if that is the case then sony fans won't have to worry about how hard it is to code for. There will be tons of games for it. End of disscusion.
 
Status
Not open for further replies.
Back
Top