Print 84 comment(s) - last by gonchuki.. on Feb 28 at 12:54 PM

The notorious Cell processor inside the PlayStation 3
The engineers behind the Cell processor lend a hand to the people behind the PS3 games

It’s no secret that the Cell processor makes the PlayStation 3 a very powerful machine, but its complex architecture also makes it difficult for developers to program their games. While Sony is likely doing its best to improve software tools, some developers are longing for lower-level detail on the Cell.

IBM, one of the creators of the Cell architecture, is stepping in and giving developers tips on how to better harness the PS3 processor. According to Next Generation, IBM engineers are participating at a game developer’s workshop at High Moon Studios in Carlsbad, Calif. with Vivendi Games studios' Radical Entertainment and Swordfish to learn the ins and outs of the multi-core Cell.

High Moon chief technical officer Clinton Keith says that that IBM's expertise goes beyond what Sony's support can offer in regard to the Cell. “We’ve been talking to Sony for almost two years now, but they didn’t create the Cell,” he said.

“They created the architecture for the PS3 and they’ve created a lot of the developer libraries. We’ve had access to those [Sony] engineers… but they’re not the hardware engineers,” continues Keith. “We want to hit [the Cell] on all fronts. We’re talking to the guys [IBM] who designed this chip and have been working on it for five years now.”

As part of the workshop, programming teams will use software development kits from IBM’s Global Engineering Solutions labs to create the best Cell-based game development algorithm. The teams will share their knowledge and findings from their work on the processor.

Sony’s Phil Harrison said that the current PS3 launch window titles use “less than half” of the system’s computational power, and that “nobody will ever use 100 percent of [the PS3’s] capability.”

Developers haven’t kept quiet on the challenges for making games on the Cell. John Carmack, lead programmer of the Quake and DOOM 3D engines, has expressed his thoughts, saying, “I think the decision to use an asymmetric CPU by Sony was a wrong one. There are aspects that could make it a winning decision, but they’re not helpful to the developers … It’s not like the PlayStation 3 is a piece of junk or anything. I was not a fan of the PlayStation 2 and the way its architecture was set up. With the PlayStation 3, it’s not even that it’s ugly--they just took a design decision that wasn’t the best from a development standpoint.”

Japanese developer Hideo Kojima shares Carmack’s sentiment, though his team was able to achieve stunning results on the PS2 with Metal Gear Solid 3. “Overall I think it is difficult, of course. But if it's easy and simple, then how can you differentiate from other games?” posed Kojima. “Other people will do the same thing as you. Therefore, we're trying to use [the Cell's] difficulty to create something different.

“If I say too much, Mr. Kutaragi might get really upset. I think that the PS3 is difficult to work with, but so was the PS2.”

Comments     Threshold

This article is over a month old, voting and posting comments is disabled

By Shoal07 on 2/22/2007 9:10:57 AM , Rating: 4
Developers aren’t stupid. Developing for a difficult architecture (and we thought the emotion engine in the PS2 was a PITA!) requires more people and more time. This means to makes the same game for the 360 or the PS3, the PS3 will have significantly higher development costs. What’s the 360 really require, .net, C3 or J#? How many thousands of developers know these backwards and forwards. Now throw in the PS3 where they’re working with the actual hardware developer to design a new language to handle programming for the cell... What a mess. Why even develop and invest the resources into this mess unless Sony is giving HUGE incentives. The risk ratio between 360 and PS3 development is going to throw alot of developers onto 360. GTA 4 on 360 in October anyone?

RE: Why?
By Shoal07 on 2/22/2007 9:12:22 AM , Rating: 2
"What’s the 360 really require, .net, C3 or J#?"
What does the 360 really require, .net, C# or J#?

Oh where oh where has my edit button gone?

RE: Why?
By thebrown13 on 2/22/2007 11:26:44 AM , Rating: 2

Which is already the best in the business, soon to be complemented by XNA.

RE: Why?
By Trisped on 2/23/2007 1:05:19 AM , Rating: 2
Really, with .Net you should be able to use any MS IDE language since they all use the same libraries and such.

RE: Why?
By maximal on 2/23/2007 12:59:04 PM , Rating: 2
Are you serious? Who in the right mind would write a game in managed code? It (Xbox 360) can support .NET all it wants, but it won't be used for game development. You need direct hardware access in order to squeeze every last ounce of performance out of it, and that means some sort of efficient C compiler front-end, and with Power architecture there is no shortage of those.

RE: Why?
By robber98 on 2/23/2007 3:54:23 PM , Rating: 2
So are you suggesting that Microsoft is an idiot? Yes, C/C++ offer better performance, but does it justify the cost (development time, debugging..etc)? For pure performance, why don't you suggest Assembly or, to be extreme, machine code? ;)

RE: Why?
By gonchuki on 2/28/2007 12:54:15 PM , Rating: 2
hardware may not be limited right now, but using inefficient programming languages only leads to reach the console limit earlier.
If your game engine already struggles in today games to run at 60fps, better optimize it or tomorrows games with more bling will be crawling.

The PS3 has the big advantage of raw power, it just needs more time to create a good engine, and it will live longer than its competitors from the same generation (just like what happened with the PS2 and its exotic hardware)

RE: Why?
By jskirwin on 2/22/07, Rating: 0
RE: Why?
By therealnickdanger on 2/22/2007 10:54:40 AM , Rating: 3
Sony needs to really, really start kissing some tails in order to keep developers in the fold

That's about the only part of your comment I agree with. While it's ignorance on our part to claim that it is "easy" or "easier" to program for the Xbox360 or the PS3, what is widely known is that the 360 has more power that's more readily accessible along with incredible hardware and software support from Microsoft.

If both the Xbox360 and PS3 were 500hp track cars, the PS3 would be the car having air/fuel problems.

RE: Why?
By robertgu on 2/22/2007 1:57:23 PM , Rating: 5
That's a good analogy.

I think people seem to get too caught up on the performance of the Xbox vs. the PS3 on one aspect of the system...the CPU Subsystem. On the whole, the Xbox and PS3 are equal in power; the Xbox has a more powerful Graphics Subsystem, the PS3 has a more theoretically powerful CPU Subsystem (if developers ever figure out how to properly use it).

Additionally, even with the relatively easy to use developement tools for the Xbox and the use of symmetrical CPU processors, developers are still not able to fully tax the Xbox. What I believe will happen is the Xbox will get closer to its theoretical potential much quicker than the PS3. And its graphics, A.I., and physics support will be superior to the PS3 for many years to come due to its more standard structure and better developer support.

I can’t see how the PS3 will ever get close to its theoretical potential. Even if developers squeeze every ounce of the Cells impressive theoretical power, what’s the point, as the graphics subsystem will be the bottleneck way before the Cell even gets taxed? You look at gaming systems everywhere, what’s to first thing that needs to be upgraded because of obsolesce for each generation of PC games? It’s the graphical subsystem. You can usually get away with using the existing CPU for a couple incremental graphic subsystem upgrades before it really starts to limit the system.

So for me, I would prefer a more standard, balanced, and easy to develop structure (al la Xbox), vs. the entirely unbalanced and difficult to develop structure of the PS3 of which has an avg. graphical subsystem and an amazing CPU subsystem. To me, the Cell is more a marketing aspect than actual usable substance for a gaming console. But that's just my useless opinion :)

RE: Why?
By ADDAvenger on 2/22/07, Rating: -1
RE: Why?
By abakshi on 2/22/2007 3:15:38 PM , Rating: 2
What? The PS3 has a G70 derived GPU and the 360 has an ATI R500.

RE: Why?
By 4745454b on 2/22/2007 3:37:05 PM , Rating: 2
The 360 uses a G70 derived GPU, the PS3 uses a G80 derived GPU, now which one is more advanced?

Not true. The 360GPU was made by ATI, and was code named Xenos. 10MBs of ram, USA, etc. The PS3s GPU is based off an Nvidia chipset. After doing some digging, I can't find the exact one. Most people say its a modded G70(1?) I couldn't find any links that showed its based off of the G80. If you care to post one, be my guest. 360 info

RE: Why?
By FITCamaro on 2/22/2007 3:45:16 PM , Rating: 3
Seriously dude. Did you even read anything about either console?

The PS3 GPU is a castrated 7800GTX with half the ROPs(8 vs. 16) and half the memory bandwidth(128-bit vs. 256-bit).

The 360s GPU is sort of a hybrid R580/600. It supports DX10 level features with it's unified shaders, has a full 256-bit memory bus, and offers free 4x AA processing on every rendered frame. It just has the same number of unified shaders as the X1900XT/X has pixel shaders.

So which do you think is more powerful?

RE: Why?
By afkrotch on 2/22/07, Rating: -1
RE: Why?
By obeseotron on 2/22/2007 8:48:05 PM , Rating: 2
At the very least it has generalized shader units assigned on the fly to the type of shader required. This is a DX10-level feature. The GPU is significantly more programmable than a standard DX9c, though absolutely not DX10 level.

RE: Why?
By willow01 on 2/23/2007 1:06:58 AM , Rating: 2
He didn't say that the X1900XT/X had unified shaders he just said that the Xenos had as many unified shaders as the X1900XT/X has pixel shaders .

RE: Why?
By baddog121390 on 2/22/2007 10:08:14 PM , Rating: 3
"offers free 4x AA processing on every rendered frame."

Not at 720p it doesnt. 10mb of edram isnt enough space to fit a 1280*720 frame with 4x anti-aliasing.

RE: Why?
By otispunkmeyer on 2/23/2007 4:31:26 AM , Rating: 2

it cant even fit a 720p image with 2xAA in it.

to overcome this they use a tile based renderer. they just make the tiles small enough to fit in with 4xAA applied.

however, i reckon that these tiles are abit too small, and that the machine will then be sampling the same geometery points over n over (ie if lots of geometry spans multiple tiles) so thats a lot of wasted work and i think the 360 isnt powerful enough to mince through it all and it just gets bogged down.

with 2XAA they can make the tiles larger and thus have less wasteful processing.

everyone should read that, its a very indepth articel detailing the inner workings of Xenos. its a very good read.

RE: Why?
By vgermax on 2/23/2007 3:17:16 PM , Rating: 2

Sorry, what math are we using to say 10MB isn't sufficient for 2XFSAA at 720p?

1280x720=921,600 pixels
3 bytes/pixel (24 bits/pixel * 8 bits/byte)
2,764,800 bytes
2.6 MB
10.5 MB

Given that both Ars, and Beyond3D (page 4, first paragraph) indicate that the design was intentionally targeting 4X FSAA for free, it seems a bit silly to say otherwise. If the quantity of RAM is assumed to be precisely 10 MB, then the buffer is more than sufficient for 2X FSAA, and just a bit shy for 4X. One would also hope then, that the 10 MB value referenced is an approximation, with the real value being sufficiently large to fit the entire data set.

RE: Why?
By otispunkmeyer on 2/23/2007 4:19:51 AM , Rating: 3
i want some of your leaf.

360 uses Xenos... or R500 as some like to call it. its a Unified architecture which gives it immense flexibility, and it also features 10mb of eDRAM on board supposedly for applying free AA to games. the 10mb isnt big enough to fit a 720p image with AA applied inside, but they get round this by using rendering in tiles.

this chip, in terms of speed, isnt even top of the DX9 pile (thats Radeon X1900-X1950) but it does have some additional flexibility that DX9 doesnt. add to that its got a full 512mb shared pool of memory to work with too.

in the PS3, the GPU is designed by Nvidia and derived from the successful G70 core, but that uses the traditional fixed number of vertex and pixel shader approach so is a little less flexible. also its probably got G70's nasty filtering quality too and will take a larger hit from applying AA.

im not entirely sure how they expect a G70 GPU to push 1080p for the next five years either. it also has 256mb of ram right on the GPU package (not on the die like Xenos), its got 4 chips surrounding the gpu. the other 256mb is located around the Cell so i would hazard a guess that if the PS3 needs more than 256mb of VRAM its going to take a penalty by leeching memory from around the cell.

RE: Why?
By ViperROhb34 on 2/23/2007 11:00:19 PM , Rating: 2
The PS3 bottlneck is its Processor memory, not the GPU itself. THEORETICAL GPU power doesn't take into account the memory.

This statistic includes the eDram logic to memory bandwidth in the Xbox 360 GPU and not internal CPU bandwidths. The eDram internal logic to its internal memory bandwidth is 256 GB/s
Sony's CPU is ideal for an environment where 12.5% of the work is general-purpose computing and 87.5% of the work is DSP calculations.

That sort of mix makes sense for video playback or networked waveform analysis, but not for games.

In fact, when analyzing real games one finds almost the opposite distribution of general purpose computing and DSP calculation requirements.

A relatively small percentage of instructions are actually floating point.

Of those instructions which are floating-point, very few involve processing continuous streams of numbers.

Instead they are used in tasks like AI and path-finding, which require random access to memory and frequent branches, which the DSPs are ill-suited to.

Based on measurements of running next generation games, only ~10-30% of the instructions executed are floating point.

The remainders of the instructions are load, store, integer, branch, etc.

Even fewer of the instructions executed are streaming floating point?probably ~5-10%.

Cell is optimized for streaming floating-point, with 87.5% of its cores good for streaming floating-point and nothing else.

Game programmers do not want to spread their code over eight processors, especially when seven of the processors are poorly suited for general purpose programming.

Evenly distributing game code across eight processors is extremely difficult.

Game programmers do not want to spread their code over eight processors, especially when seven of the processors are poorly suited for general purpose programming.

Evenly distributing game code across eight processors is extremely difficult.

GPU Even ignoring the bandwidth limitations the PS3's GPU is not as powerful as the Xbox 360's GPU.

Below are the specs from Sony's press release regarding the PS3's GPU.

RSX GPU 550 MHz Independent vertex/pixel shaders 51 billion dot products per second (total system performance) 300M transistors 136 "shader operations" per clock The interesting ALU performance numbers are 51 billion dot products per second (total system performance), 300M transistors, and more than twice as powerful as the 6800 Ultra.

The 51 billions dot products per cycle were listed on a summary slide of total graphics system performance and are assumed to include the Cell processor.

Sony's calculations seem to assume that the Cell can do a dot product per cycle per DSP, despite not having a dot product instruction.

However, using Sony's claim, 7 dot products per cycle * 3.2 GHz = 22.4 billion dot products per second for the CPU.

That leaves 51 - 22.4 = 28.6 billion dot products per second that are left over for the GPU.

That leaves 28.6 billion dot products per second / 550 MHz = 52 GPU ALU ops per clock.

It is important to note that if the RSX ALUs are similar to the GeForce 6800 ALUs then they work on vector4s, while the Xbox 360 GPU ALUs work on vector5s.

The total programmable GPU floating point performance for the PS3 would be 52 ALU ops * 4 floats per op *2 (madd) * 550 MHz = 228.8 GFLOPS which is less than the Xbox 360's 48 ALU ops * 5 floats per op * 2 (madd) * 500 MHz= 240 GFLOPS.

With the number of transistors being slightly larger on the Xbox 360 GPU (330M) it's not surprising that the total programmable GFLOPs number is very close.

The PS3 does have the additional 7 DSPs on the Cell to add more floating point ops for graphics rendering, but the Xbox 360's three general purpose cores with custom D3D and dot product instructions are more customized for true graphics related calculations.

The 6800 Ultra has 16 pixel pipes, 6 vertex pipes, and runs at 400 MHz.

Given the RSX's 2x better than a 6800 Ultra number and the higher frequency of the RSX, one can roughly estimate that it will have 24 pixel shading pipes and 4 vertex shading pipes (fewer vertex shading pipes since the Cell DSPs will do some vertex shading).

If the PS3 GPU keeps the 6800 pixel shader pipe co-issue architecture which is hinted at in Sony's press release, this again gives it 24 pixel pipes* 2 issued per pipe + 4 vertex pipes = 52 dot products per clock in the GPU.

If the RSX follows the 6800 Ultra route, it will have 24 texture samplers, but when in use they take up an ALU slot, making the PS3 GPU in practice even less impressive.

Even if it does manage to decouple texture fetching from ALU co-issue, it won't have enough bandwidth to fetch the textures anyways.

For shader operations per clock, Sony is most likely counting each pixel pipe as four ALU operations (co-issued vector+scalar) and a texture operation per pixel pipe and 4 scalar operations for each vector pipe, for a total of 24 * (4 + 1) + (4*4) = 136 operations per cycle or 136 * 550 = 74.8 GOps per second.

Given the Xbox 360 GPU's multithreading and balanced design, you really can't compare the two systems in terms of shading operations per clock.

However, the Xbox 360's GPU can do 48 ALU operations (each can do a vector4 and scalar op per clock), 16 texture fetches, 32 control flow operations, and 16 programmable vertex fetch operations with tessellation per clock for a total of 48*2 + 16 + 32 + 16 = 160 operations per cycle or 160 * 500 = 80 GOps per second.

Overall, the automatic shader load balancing, memory export features, programmable vertex fetching, programmable triangle tesselator, full rate texture fetching in the vertex shader, and other "well beyond shader model 3.0" features of the Xbox 360 GPU should also contribute to overall rendering performance.

Can u put the pics how do they look Hmmm..Can we compare XBox and Sony PSP ?

WHat the heck,I am content with my PC
Bandwidth The PS3 has 22.4 GB/s of GDDR3 bandwidth and 25.6 GB/s of RDRAM bandwidth for a total system bandwidth of 48 GB/s.

The Xbox 360 has 22.4 GB/s of GDDR3 bandwidth and a 256 GB/s of EDRAM bandwidth for a total of 278.4 GB/s total system bandwidth.

Why does the Xbox 360 have such an extreme amount of bandwidth?

Even the simplest calculations show that a large amount of bandwidth is consumed by the frame buffer.

For example, with simple color rendering and Z testing at 550 MHz the frame buffer alone requires 52.8 GB/s at 8 pixels per clock.

The PS3's memory bandwidth is insufficient to maintain its GPU's peak rendering speed, even without texture and vertex fetches.

The PS3 uses Z and color compression to try to compensate for the lack of memory bandwidth.

The problem with Z and color compression is that the compression breaks down quickly when rendering complex next-generation 3D scenes.

HDR, alpha-blending, and anti-aliasing require even more memory bandwidth.

This is why Xbox 360 has 256 GB/s bandwidth reserved just for the frame buffer.

This allows the Xbox 360 GPU to do Z testing, HDR, and alpha blended color rendering with 4X MSAA at full rate and still have the entire main bus bandwidth of 22.4 GB/s left over for textures and vertices.

CONCLUSION - When you break down the numbers, Xbox 360 has provably more performance than PS3.

Keep in mind that Sony has a track record of over promising and under delivering on technical performance.

The truth is that both systems pack a lot of power for high definition games and entertainment.

However, hardware performance, while important, is only a third of the puzzle.

Xbox 360 is a fusion of hardware, software and services.

Without the software and services to power it, even the most powerful hardware becomes inconsequential.

Xbox 360 games?by leveraging cutting-edge hardware, software, and services?will outperform the PlayStation 3.

RE: Why?
By AnotherGuy on 2/22/2007 3:55:47 PM , Rating: 3
Not a good analogy.... Id say more like the PS3 is having mechanics problems or drivers problems... noone knows how to drive this beast or how to use all its performance

RE: Why?
By jtyson on 2/22/2007 7:04:25 PM , Rating: 3
Actually, that's a piss-poor analogy. How a car runs has nothing to do with how difficult it is to engineer. With your logic, PlayStation 3 software will need to be constantly paused or rebooted. That's just... not... well, suffice to say, it's just a bad analogy.
A better analogy would be how well a car is tuned. For example, a Mustang is pure muscle. On the other hand, a car like a Skyline or NSX (whose engines were made by hand) can easily hang with a Mustang and keep pace. But in order to edge the Mustang out, the car has to be finely tuned. Is it harder to do? Sure. Does it cost more money? Absolutely. But the end result is a thing of beauty. Pure quality. Have you seen Devil May Cry 4? Metal Gear Solid 4? Final Fantasy XIII? Difficulty or no, these developers are doing some pretty sweet things with the PlayStation 3 hardware. These games look amazing.

To me, it's not developing that's the problem; the Cell processor is going to be used in a lot more things than just the PlayStation 3. Developers are going to have to learn this stuff anyway. To me, the only thing holding the PS3 back (besides the price) is the lack of vibration in the controllers. I can't get over it. I tried. I pretended like I didn't care. But I do. It bothers me. I want it back (and so does Kojima). If they did that, then I think you would see developers a little more eager to make some PS3 software.

RE: Why?
By Locutus465 on 2/22/2007 7:17:25 PM , Rating: 2
That's about the only part of your comment I agree with. While it's ignorance on our part to claim that it is "easy" or "easier" to program for the Xbox360 or the PS3, what is widely known is that the 360 has more power that's more readily accessible along with incredible hardware and software support from Microsoft.

Ever use visual studio 2005? One of Microsoft's greatest strengths are their development tools and documentation. These things are big part of the reason why windoz pwns all in market share. I've already heard from several publications that game developers (such as John Carmac) feel that developing for X-Box is a bunch easier. I would tend to beleive it. Sony doesn't have Microsoft's experience with developing tools that are friendly to end developers that Microsoft has gained through nessesity.

RE: Why?
By thebrown13 on 2/22/07, Rating: -1
RE: Why?
By saratoga on 2/22/2007 1:17:40 PM , Rating: 3
Quit yer whining!
The Cell is new technology that requires learning new coding methodologies. But the payoff is the ability to tap into a truly powerful processor that will result in some seriously cool games.

Any machine is only as good as the people using it. If you build a machine that people cannot use efficiently, you've built a poor machine. Sony's best bet now is to build up the tools and middleware to compensate for the hardware trouble. But its a tough fight. MS already took over the world in part by writing good dev tools, now they're trying it on consoles too, and Sony played into their hand.

RE: Why?
By Panurge on 2/22/2007 2:07:36 PM , Rating: 3
If we're utilizing less than half of the processor now and making games that look as good as the more established and lower tech Xbox 360, just imagine what the games are going to look like in a year or two when developers learn how to use that power.

You're confusing the abilities of the CPU and GPU in this statement. First, you say less than half of the CPU is being utilized in current gen games, but then go ahead and ask how great they'll look in the future.

This is one of the problems with the "potential" of the Cell. Current games tend to focus on two areas for improvement, graphics and AI (gameplay aside, since it is a different topic, and I can't think of many situations offhand where a high end CPU would change gameplay without focusing on one of the previous two areas).

For graphics, the solution is a graphics processor. Obviously, most new consoles have focused on this to some degree, with the end result being an approximate equality in what each can output. The GPU is, for the most part, separate from the CPU.

For AI, the multi-core chips being used are limited in what can be done. Having 6 SPUs all capable of large amounts of real-time floating point calculations doesn't particularly help in AI. In fact, the architecture (if I have read correctly) doesn't provide much in the way of branch handling, which makes AI even more difficult to handle (at high performance levels).

To sum all this up, I don't see the point of adding a complicated, high performance, asymmetric chip of this type with the hopes that game development will catch up and make use of it. Unless some big change comes to the world of gaming, two of the biggest "improvements" make regularly won't get huge strides from increaded utilization of the Cell chip (or the Xbox360 chip for that matter).

RE: Why?
By Holytrinity on 2/22/2007 10:53:20 PM , Rating: 4
The Cell is new technology that requires learning new coding methodologies. But the payoff is the ability to tap into a truly powerful processor that will result in some seriously cool games.

I disagree. The last "seriously cool game" I played was made cool by the gameplay of the game, not the processor it was running on.

RE: Why?
By ElJefe69 on 2/23/2007 9:05:51 PM , Rating: 2
The best last game I played was Fallout 2

2nd best graphically advanced I played was vampire bloodlines, masquerade. Good games, stories, interesting, not bs, not kiddie shit.

find a game like that now. stories suck ass. graphics are meaningless efforts of a 500 dollar gpu with no artistic merit.... could go on for a long time.

RE: Why?
By msva124 on 2/22/2007 2:54:01 PM , Rating: 3
What does the 360 really require, .net, C# or J#?

Yes, the 360 processor directly executes .NET bytecode. :rolls eyes:

I do agree with the sentiment that building a machine that people cannot use efficiently is a poor machine. If the cell architecture is so convoluted that Sony was unable to make decent libraries for it, and a complete rewrite is required to achieve faster performance than the Xbox 360 version, Sony is in big trouble. Microsoft has many years experience in making good programming libraries. And they came out with their system first, so many games have been developed or began development using the Microsoft libraries.

Along comes a new system with more processing power, but that's a pain in the ass to program for. Hey boss, we can either spend additional time and money re-writing all our code to wring extra performance out of this new architecture that noone's even bought yet, or we can just do a port from our Xbox 360 code that looks more or less the same. It's not a hard choice. Sony mucked up big time when they thought they'd have a monopoly at this point.

RE: Why?
By ObscureCaucasian on 2/22/2007 3:39:19 PM , Rating: 2
You make good points. If you look at why the PSone became so popular, it was because Sony was much more developer friendly than Nintendo. Microsoft is a company that made itself by creating developer tools, and as a result they made a system that is much easier to develop for. Just look at what they have with XNA and the opportunities they have made for the homebrew community. Sony needs to really change the way they treat third parties or else they're going to lose some serious market share. Many of the Japanese developers are slow to move to the Xbox, most likely due to its unpopularity in Japan. I also think it could have to do with pride in the Japanese culture. They'd rather support Sony, but the advantages of the Xbox are just too great for some of them.

RE: Why?
By AstroCreep on 2/22/2007 4:57:20 PM , Rating: 2
Why? I think it may be because IBM doesn't want the PS3 to 'tarnish' the Cell's good name. Also this way, by giving a bit of a technical hand to those (who may be) in need, it will look good for IBM and help the aforementioned developers 'hit the ground running'.

PR Move? Maybe, but it sounds functional.

RE: Why?
By afkrotch on 2/22/2007 7:29:55 PM , Rating: 2
128,085,000 ppl in Japan and the PS3 is winning the console war there.

Now you tell me, why isn't it worth developing for it? Even if you have one game that sells to 1 million users for 7000 yen, that's $57,634,461

RE: Why?
By robber98 on 2/22/2007 10:06:07 PM , Rating: 2
the PS3 is winning the console war there.

What are you smoking? Since when PS3 winning in Japan? Lot of PS3 leave on shelves and collecting dust, game store slash price on PS3 (20% off), 2nd-hand store stop taking any PS3 and the most recently sales number from are indicated that PS3 isn't winning (not even close).

RE: Why?
By chrnochime on 2/23/2007 1:46:53 AM , Rating: 2
And what are you smoking?? The PS3 has sold more compared to the 360, so it's winning there. The primary competitor is 360 so of course the comparison of sales numbers should be between the PS3 and 360, NOT between PS3 and Wii. If you're going to say it's losing to Wii, yes but then so is 360, so that isn't making the 360 look any better.

RE: Why?
By SmokeRngs on 2/23/2007 3:27:37 PM , Rating: 2
Now you tell me, why isn't it worth developing for it? Even if you have one game that sells to 1 million users for 7000 yen, that's $57,634,461

You're ignoring some realities here. If there are difficulties in programming for the PS3, it will raise development costs. It takes more time to do things and time is money. You have to pay those people that are spending extra time to work on it. This reduces the profits from the game. That's a plain and simple fact.

If you can do the game for another console that isn't as difficult to program for, you can get basically the same game out for less money since development costs are lower. This means a higher profit.

I think this is one of the main reasons many developers are not making Sony exclusives anymore. The other consoles have the numbers out there to sell games.

Not every game will have higher development costs on the PS3 compared to the Xbox360 and in time it will be easier to develop for the PS3 which should reduce costs.

Remember, everything boils down to profit. This can mean a company will have to take smaller profits on a game to get out the game they want or keep about the same profits and not make the game as good as they had planned.

RE: Why?
By afkrotch on 2/22/2007 8:14:59 PM , Rating: 2
128,085,000 ppl in Japan and the PS3 is winning the console war there.

Now you tell me, why isn't it worth developing for it? Even if you have one game that sells to 1 million users for 7000 yen, that's $57,634,461

RE: Why?
By cheetah2k on 2/22/2007 8:49:03 PM , Rating: 2
While I'm not a programmer, or have anything to do with the software industry, I feel the problem these days is that Developers are so caught up on their end of year revenue report, rather than implimenting new technology. Obviously programming CELL is hard (this is understandable), and the learning curve in making the transition from PS2 to PS3 will take time and cost.

I am sure DX10 on the PC will suffer too, as the GPU architecture has changed alot from the good olde DX9 days, and developers will, again, have to invest time and cost learning (albiet not as much as CELL).

Time will tell - lets see what happens.

RE: Why?
By akugami on 2/22/2007 9:53:53 PM , Rating: 2
You might want to recheck your facts there buddy. The PS3 is currently in second place with the Wii in first, followed by the PS3, then trailing is the Xbox 360.

Check the above link for lifetime sales numbers and you'll see the Wii is in first place. The PS3 is second place and under 700k and assuming they sold enough to reach 700k by now, the best a game can hope for is to sell to every single installed user and that's 700k units while a game can sell to half of the installed Wii user base to reach 700k. Most games do not sell to half of the installed user base and if a game makes 500k total sales in it's first month then it's a smashing success. At this point, with the PS3's installed user base, I don't think 500k is possible unless the game is a revolutionary one. The next Mario 64 or Final Fantasy so to speak.

The PS3 is in trouble at this point and developers are not as sold on it as they once were. By no means is the PS3doomed to failure but Sony needs to get something out to justify the hype. Either a game so revolutionary that it's a must have even if you're a casual gamer or a major price reduction to make it more within reach of the average household.

They are going about it the wrong way...
By exdeath on 2/22/2007 10:22:07 AM , Rating: 3
Many people see Cell as "a main PPC CPU with 6 SPEs as coprocessors or vector units," trying to apply what they know about PS2 to the PS3.

However it should be the reverse... instead you should target the SPE as your main build target and pretend you have a 6 core CPU. All game code should be threaded SPE code, with the PPE pumping the main loop and handling I/O tasks and engine/OS house keeping.

Beginners should ignore the Power core entirely at first and concentrate on mastering the SPEs, then figure out later what menial tasks the Power core can do. This is pretty much how the Cell is designed in the first place.

I haven't seen the SDK, but hopefully the compiler can hide and abstract the fact that the SPEs even have local storage or DMA, by optimizing loops or data accesses with DMA block pre-fetching from main memory as needed, etc, thus making the main stumbling block of SPE coding transparent to the newbie.

RE: They are going about it the wrong way...
By Chaser on 2/22/2007 10:32:08 AM , Rating: 2
I read an article elsewhere that an independent 3rd party was working on a new SDK for the PS3. They hope to make it less complicated and easier to utilize more of the cell processor.

I believe Carmack was also quoted that he didn't favor multi core PC CPUs neither.

Times a changing.....

By Carl B on 2/22/2007 11:50:42 AM , Rating: 2
It's not third-party, but this may be what you're thinking of:

Sony's PSSG Project:

RE: They are going about it the wrong way...
By exdeath on 2/22/2007 3:06:33 PM , Rating: 3
It's not so much that multi core in and of itself is a problem on the PS3. Standard multi core CPUs, on the PC and the XBox360 for example, can all access main memory and can all execute the same code on any core. You have a universal set of homogenous CPU cores. Any and all code you write can run on any and all cores, a single variable in main memory is accessible from all cores, etc. From a tools and development environment standpoint, nothing changes from a single threaded environment. The only concern on a PC or a 360 is the standard dependency resolution and that is inherit to multithreading on any platform.

What makes the Cell such a pain is that the PPE main core and the SPE cores are drastically different architectures with different memory access methods. This means that the programmer is aware of the differences at all times, which means you focus on the systems programming instead of your game programming. You can’t just write one C function and thread it and know that it can run its code and access system variables from any core. You have to worry about what’s in main memory, what’s in SPE local storage, is this SPE code or PPE code, what’s calling what and how to marshal data between code of two architectures, making sure DMA streams are started early before the data is need, etc. Even within your SPE code you have to be aware of what you have and what you don't.

Anything the compiler or SDK can do to help make the Cell appear as a homogenous set of universal cores will help. If you start by ignoring the single PPE core at the start, and focus on simply 6 x SPE as the primary build target, then you can write all your game code as SPE code that can run on any SPE and forget about the PPE. Then have the compiler automatically setup and optimize the local store memory map and DMA channels to make it work as cache behind the scenes. The only thing you want giving away the fact that there are even multiple cores are the ‘runthread’ type calls.

Basically, the problem is porting. Most developers build games across all platforms. The PS3 architecture is different than everything else and will take special hacking to patch up game engines that otherwise work on all other platforms. Nowhere on a Xbox360 or PC or Wii will you ever see code that has one core doing a raw data copy via DMA to upload code to another core, etc. Coming up with a common interface between all systems will be interesting...

I love Cell in a sadomasochistic hobbyist hacking sort of way for my own amusement, but in the real world where you want to make games and not tinker with exotic hardware, an identical universal set of homogenous cores is the preferred method to go multi core.

By exdeath on 2/22/2007 3:22:24 PM , Rating: 2
You might have something like this if you were developing a cross platform engine with the limitations of each machine in mind:


for(int i = 0; i < 100; ++i)


On PS3, Prefetch() would have to be a SPE call setting up DMA to pull main memory data into the local store. Also, keep in mind the SPE code must also fit in SPE local store as its own independant executeable since it has to be copied into SPE local memory to execute. This means that each outer most SPE thread function, along with any subroutine calls, must all be compiled into one object file. Lots of duplication of common code that would be called across different thread functions. Unless you want to start getting into your own virtual memory management to dynamically copy in subroutine code via DMA to any free address in local store..., etc, etc... as you can see this is why the architecture is a pain.

On PC/360 Prefetch() may just do a dummy read to prime the caches, or may do nothing at all...

By saratoga on 2/22/2007 4:14:29 PM , Rating: 2
love Cell in a sadomasochistic hobbyist hacking sort of way for my own amusement, but in the real world where you want to make games and not tinker with exotic hardware, an identical universal set of homogenous cores is the preferred method to go multi core.

Indeed. The cell seems to have been designed to combine the most annoying aspects of DSP programming and traditional multithreading into one machine.

RE: They are going about it the wrong way...
By FITCamaro on 2/22/2007 3:55:25 PM , Rating: 2
You can't just ignore the main PowerPC core. The SPEs don't do anything without it. They don't think for themselves. They're just given a single task, execute it, and return a result. You can get more than one to work on a task, but that is done through the PowerPC core. And even still, you don't have 7 SPEs at your disposal. Only 5. The other two are tied up with the OS and I forget the other thing.

But they are by no means separate cores that can make decisions on their own. Yes the PowerPC core is basically the task scheduler, but its not as simple as running a single main loop that tells the SPEs to each run their own individual programs.

RE: They are going about it the wrong way...
By exdeath on 2/22/2007 5:41:18 PM , Rating: 2
You can have the SPEs self multitask. Basically as soon as the current task finishes, control returns to a small microkernel scheduler stub that grabs the next task off a task list, indefinitely. Tasks can spawn other tasks, so even the main loop can be a SPE task that just spawns itself over and over again.

If you wanted to, you could write everything on the SPEs and let the PPC core idle (other than only the PPC can handle system interrupts, etc) But my point was, the SPEs are the *heart* of the Cell, not just co-processors. Writing multi-threaded apps on the Cell should start first and foremost with the SPEs. The SPEs should be looked at as a 6 core CPU the way the 3 PPCs in the 360 are looked at as a 3 core CPU. The key to multithreading is having any code executing on any available free core. Since there are 6 SPEs and 1 PPC and you can’t run SPE code on the PPC, you might as well forget about it and focus on the SPEs. Any code that is to be multithreaded is automatically SPE only code. Any code written on the PPC defeats the purpose of Cell in the first place.

PPC is suited for single threaded non repeating tasks, things done ones per game loop, yes of course, and it would be a waste to ignore it completely. I was just making a point, that for the beginner, forget the PPC for now and focus on the SPEs only so you have a homogenous universal multi core CPU where any of your 6 cores can run any of our 6 threads equally. Don't think of the SPEs like you would think of the PS2s VUs, instead think of them as your main cores where 99% of you work should be going on. All object processing, AI, physics, path finding, geometry processing, collision detection, etc, should be re-entrant and thread safe in SPE code. Use the PPC for things done once per frame and for feeding network, input, and disc data to the SPEs.

The Cell chip has 8 SPEs. The PS3 has 6 useable SPEs. One is reserved for the PS3 BIOS/OS and another is reserved for yields so that at least one SPE can have flaws and be disabled and still have a useable Cell chip.

By exdeath on 2/22/2007 5:43:42 PM , Rating: 2
"you have a homogenous universal multi core CPU where any of your 6 cores can run any of our 6 threads equally."

you have a homogenous universal multi core CPU where any of your 6 cores can run any of your many threads equally.

By wetwareinterface on 2/23/2007 5:10:50 AM , Rating: 1
you forget one problem with coding for the spe directly that affects your theory of how to improve code for ps3

it is a specialized floating point processor with no scheduling of it's own. it has to be handed it's tasks by the ppe hence you have to code for the ppe first and worry about what fp instructions you hand off to the spe's and let the ppe do it's thing first and foremost. the spe's are there simply to handle fp tasks. not do scheduling or integer compares etc...

By exdeath on 2/23/2007 9:47:12 AM , Rating: 2
The SPE is just a in order CPU like the ARM7 in the Gameboy Advance, the Vector Units in the PS2, etc. You can easily set up a microkernel on the SPE that will grab tasks off a queue as they become available. In fact this is one of the three modes of operation documented explicitly by IBM, the three being 1) explicit PPE scheduling of the SPEs, 2) self multitasking of the SPEs without PPE intervention, and 3) pipelined streaming of the SPEs.

But I think everybody is missing my point completely... I didn't mean never ever use the PPE, I was simply making the case that all game logic code should be SPE code so that it can be universally multitasked on any available SPE as one becomes free...

Example, the PPE might run the main game loop and do basic sphere-sphere collision detection for example, in the collision detection phase of the UpdatePhysics() call in the main loop. But any sphere-sphere checks that pass will then be added into a job queue for the SPEs. Say you have 1000 objects and 150 of them pass the sphere-sphere test, you have 150 potential collision pairs which result in 150 jobs in the SPE work queue. Assuming standard linear game loop running on the PPE exploiting SPE parallelism at each discrete step, 6 of those 150 collision pairs will be examining further mesh-mesh collision resolution simultaneously, and the first SPE of the 6 available SPEs that finishes will grab the next one, and so on, until all jobs are finished. The SPEs go idle for a split second while the PPE synchronizes and moves on to the next part of the main loop and fills the job queue again with AI tasks or whatever is next.

Also game logic such as AI decision making and path finding, should be broken down into SPE modules such that multiple objects can process their AI simultaneously across all available SPEs. In other words, all game code where actual game state changes should be SPE code to exploit parallelism.

On the PC and 360 it's not an issue, you just write all code the same as you normally would, and any code can run on any core and access any memory without any special considerations other than the usual multithreading pitfalls such as data consistency (which can be handled by double buffering all thread data to guarantee the data is constant throughout the frame and swap buffers on a frame sync. No locking or thread synchronization needed in the game code).

By exdeath on 2/23/2007 10:02:05 AM , Rating: 2
You don't send SPEs individual instructions, you send them entire programs, and one of the first programs could be a scheduler if you wanted. I don't recall the specifics but I believe the SPEs have access to the entire system memory map via DMA, such that they could take control of the graphics chip and submit their own display lists, etc, without PPE intervention.

My point was, that for a beginner who barely understands multithreading, much less with non-homogenous cores with incompatible instruction sets and memory models, it might be best to think of the Cell as just a standard 6 core CPU and think in terms of SPEs and not be too concerned with the PPE at first.

I'm not saying intentionally bypass and never use the PPE at all, I'm just strongly making the point that rather than write a game for the PPE core then later figure out 1 or 2 ways you can use the SPEs as coprocessors , your design from ground up should be SPE-centric from the start. Anyone compiling 90% of their program for the PowerPC core and trying to use the SPEs as coprocessors will never ever realize the full potential of Cell.

The SPEs is what the Cell is all about. The PPE is basically... just there. =D Personally I think they could have added a few basic things to the SPEs to make them slightly more general purpose, ditched the PPC, and just have like 12 SPEs… would have made things easier from the start.

internal conflicts?
By shimman on 2/22/2007 9:15:58 AM , Rating: 4
We’ve been talking to Sony for almost two years now, but they didn’t create the Cell

so who did create the cell processor? has sony been misleading us to believe that the cell processor was a fruit of sony's engineering?

Sony’s Phil Harrison said that the current PS3 launch window titles use “less than half” of the system’s computational power, and that “nobody will ever use 100 percent of [the PS3’s] capability.”

then, what's hell is the point of having such processing power?

But if it's easy and simple, then how can you differentiate from other games?” posed Kojima. “Other people will do the same thing as you. Therefore, we're trying to use [the Cell's] difficulty to create something different.

choosing a difficult route when there is an easy way wouldn't make much sense unless there is a big pay off later. not sure why kojima said that because creating games isn't about overcoming technical difficulties. design of game itself counts more & by having an easier to use architecture, they could concentrate on creative parts of game making; perhaps, that's why ps3 games either suck or no different from existing games so far?

RE: internal conflicts?
By SquidianLoveGod on 2/22/2007 10:25:29 AM , Rating: 5
I saw an AD which showed that Final Fantasy 8 Used approximately 90% of the PSones power.
Perfect Dark Probably used Most of the Nintendo 64's Available Power, even with the 4Mb of Rambus Memory Pack thingy.
With the Play Station 2, So to say that no game will ever utilize all of the PS3's power is probably true, as Game developers generally keep their games just under the max performance of the system, to keep games running smoothly, So there isn't any frame rate loss during intensive situations which was exhibited in Perfect Dark on the Nintendo 64, But the way they are saying it, Is its impossible no developer could tap all the Power of the PS3 And Yadda Yadda Yadda which is un-true, Just developers choose not to.

I am also confused on who "made" the Cell processor, To many mis-truths have been told but here is line from wikipedia
"Cell is a microprocessor architecture jointly developed by a Sony, Toshiba, and IBM alliance known as STI"
So I guess Sony, Toshiba and IBM all helped design the cell.

The Power of a system really doesn't influence the different types of games that a console can play, Since when has the Graphics of a game made a game great?
For instance the Wii may not be up to standards per-say when it comes to Graphical Power, But it indeed does have alot of new and unique ideas, The DS is the same it may not look as pretty as the PSP when it comes to games, But god damn! It makes you wonder how you ever gamed on a hand held without a touch pad and Dual screens! And most games use those ideas differently, Which makes the Cell's hard to tap power to make unique games kind of a stupid reason to think its going to influence the variety of games, if anything at all There would be less quality games, because the amount of money that has to be put into making the game run on the Cell efficiently.

RE: internal conflicts?
By ObscureCaucasian on 2/22/2007 3:48:47 PM , Rating: 2
It was probably financed by Sony, Toshiba, and IBM, but IBM almost definitely did most of the dirty work engineering the chip. It would be foolish to partner w/ IBM to make a CPU, then not let them do what they do best.

I always thought the Sony/Toshiba partnership was kinda odd, because think about it, Sony/Toshiba are partners on the Cell, but direct competitors in the BluRay/HD DVD war.

RE: internal conflicts?
By afkrotch on 2/22/2007 8:33:50 PM , Rating: 2
The point of having such processing power is so that developers can eventually get near it's full capability, before the next generations of consoles will come out.

What's the point of coming out with a console that only survives a couple years?

RE: internal conflicts?
By akugami on 2/22/2007 10:03:49 PM , Rating: 2
I found the Kojima comment on the ease of accessing a system's power extremely odd. The easier it is to access the power of a system the better you should be able to concentrate on game design which should result in a better game. For instance, you can have a great game design concept but if you're struggling with the system because it is difficult to program for then your game will suffer...or you spend a huge amount of time (wasted $) developing your game. And we've seen plenty of games with great sounding designs get crushed by bugs and lack of polish.

Or you can have a good game design and have a relatively easy time of programming and implementing what you wanted giving you more time to polish up your game. The final product, while nothing revolutionary may very well be an extremely polished game that is well received.

RE: internal conflicts?
By Darthefe on 2/23/2007 4:52:21 AM , Rating: 2
why r u putting down the PS3 so badly i dont get it. programming a ps3 game yeah i get it its difficult but no developer really learnt how to programm with a Cell processor and why would it cost more to programm with a Cell can anyone explain me that? it takes more time but dose not necceirly cost more. and ps3 games have as many bugs as the X-box 360 get it.

RE: internal conflicts?
By akugami on 2/23/2007 2:29:18 PM , Rating: 2
First, please read through what I wrote. Second, I did not put down the PS3. My thoughts would apply to any and all game console platforms. If you at the very least thought through what I wrote it would be very obvious why the PS3 costs more to develop for.

My thoughts were on Kojima's comment that a difficult to develop for system separates the good developers from the bad. My argument is that with an easier to develop for system (which is not the same as a weak system) the overall level of quality raises because one can concentrate on making a game good and not wrestling with it to do what you want.

Moving from a single core to dual core processor already requires more programming work to effectively harness the power of a dual core processor. This is a necessary step as almost all newer systems that are higher powered are moving towards multi core designs. However, the Cell is different in that the SPE's are highly specialized cores that are not suited for the typical code that one would run on a single core CPU or even multi core CPU's that use the same cores. That is not to say the Cell is a bad CPU, just that it might not be the best suited for a video games console and that while powerful, it is difficult to effectively use all of it's power.

Someone used a car analogy and I think it works well in this case. The Cell is like a powerful car engine that requires a lot of fine tuning and coaxing to get it to perform optimally. This takes a lot of time and effort in order to do so. The CPU in something like the Xbox 360 (or Intel Core 2 Duo or AMD Athlon X2) is like a weaker but still powerful engine that is easy to utilize and get close to maximum power out of.

With the Cell, it is more costly to get it performing at a high level due to the difficulty of programming for it. This is because you can either use the same man power to code for the Cell as you would on the Xbox 360 (or Wii) but it would take you longer to get good results or you can throw more man power at it which results in shorter development times but again, much like a longer development time, it means more money. More people required for a project or an equal number of people in the project but with a longer development time means it costs more to finish the project. This is not a hard concept to grasp. It would be the same if the roles were reversed and the PS3 was the easier to program for, but not as powerful, system and the Xbox 360 was the powerhouse but difficult to program for.

Another aspect that reduces cost is that the easier to develop for system likely will have less bugs because you're not struggling just to get a game running. You also will have more time to test for bugs and an easier time at it since you can get your project finished more quickly. Think of trying to find out where you went wrong in a series of equations where you're adding up a series of math equations. It's easier to find out where you added wrong if all 10 math problems are simple addition, subtraction, multiplication and division. But when the series of equations include trigonometry, calculus and statistics problems, then it becomes increasingly difficult to check where you went wrong.

As I stated before, a decent game concept and design that is well polished is better than a great game concept and design with poor polish. Imagine a game like Grand Theft Auto but with poor collision detection, poor AI, hard to use control, enemies that can run or see through walls, etc. These are some of the symptoms of a game that lacks polish. And as great of a gaming concept as the GTA series has been, no one would play it.

Bottom line, a game that has a shorter development time costs less because the longer it takes the more money you have to pay to your programmers and artists and other folk working on the game. It's an extremely simple concept to grasp. The easier a console is to program for the less likely you will run into bugs and even if you do, it is easier to fix the bugs.

RE: internal conflicts?
By Darthefe on 2/25/2007 11:42:45 AM , Rating: 2
I did not mean that u put down the PS3 this message was supoosed to go to those pessemstic folks which think it is impossible for the PS3 to win the console or at least het second and thats actually the most possible ranking at the end of this generation. What i wrote about the price thing was'nt thinking. But just a point to make it is not that it is difficult to program with it is just that the programmers have to get used to programing with it and also get the concept so give those folks time and after a year they get the concept the reason is that all programers r used to core proccesors technology not they new cell technnology.

RE: internal conflicts?
By togenshi on 2/26/2007 11:17:34 AM , Rating: 2
Somewhere along the lines, open source was totally forgotten. Lets take 2 examples --> PowerPC linux and the PSP.

Start with PSP because it is easier. Most people think that the PSP is just your ordinary hand-held game console with a few added features such as music, pictures, movies and the internet. Well, that is just the tip of the ice berg. The PSP is now a shoutCAST streamer, GPS device (with waypoint and now supported by Sony), language translater, emulator for pretty much anything (windows 95, mac os, n64 to gamecube), playstation 1 console, etc, etc. This is all because Sony entered an Open Source agreement. Though it did kinda backfire (UMD dumping), it still played a major role in the PSP's development stage and sales.

Now the PowerPC linux is also open source. Therefore the capabilities of the PS3 extends greater than just being an ordinary gaming console. Instead of having a closed source like microsoft, sony has allowed people outside the games development group to input their efforts into the PS3. PS3 can be capable of having mod packs that extend the gaming experience of ordinary games. In most cases this would be free. Hopefully the community will see PS3 an opportunity to expand upon like the PSP did.

But i was dissappointed about the PS3 XDR ram reduced from 512mb to 256mb (ages ago), i still think it is a really powerful machine. Fair enough each processor type is going to have its pros and cons (not much you can do about that if damage has been already made) though the japanese aren't stupid to invest so much money to receive so little.

Lets take the Emotion engine for example. NO-ONE had a clue what that thing did, yet look at the outcomes. Fair enough at lack of digital output, but the PS2 vs 360 (analog for analog signals), i would reckon the PS2 would put up a decent fight from games such as FFXII, Gran Turismo, Devil May Cry etc. Im talking about prehistoric (in tech terms) console vs next gen console. Hopefully it doesnt take 6-7 years for the PS3 to mature like the PS2 did. All that developers need is experience compiling these games. As soon as they develop and get used to the PS3 environment, possibilities are quiet endless.

The linux install was a good move as all distros come with a C complier, therefore people with some C programming knowledge would be able to explore and expand the functions of the PS3. I think this would be its greatest asset. Unlike the 360 where everything is pretty much fixed, PS3 can be free to modify. Fair enough Hardware acceleration will not be supported in linux (to make it impossible to mount and play ps3 games), there are still avenues to explore.

The ps3 just needs time to mature to a point where development teams can exploit the full potential. Let time tell whether Sony did do the right decision or not.

I'll answer this one...
By tfranzese on 2/22/2007 1:32:47 PM , Rating: 2
“Overall I think it is difficult, of course. But if it's easy and simple, then how can you differentiate from other games?” posed Kojima

Gameplay, depth and imagination. Graphics shouldn't be what separates a good game from a bad one, but rather how fun the game is.

RE: I'll answer this one...
By msva124 on 2/22/2007 2:24:46 PM , Rating: 2
Which is why the Wii is (right now) winning the console wars. Graphics are a nice touch, but they are like DVD vs VHS. It looks nicer, but if the movie sucks I'll take VHS any day of the week.

RE: I'll answer this one...
By FITCamaro on 2/22/2007 3:58:52 PM , Rating: 2
I wouldn't exactly compare console wars to VHS vs. DVD. Anything that can go on VHS can go on DVD and look way better. There's no difference in the content between the two other than the quality. And DVDs do it in a smaller, cleaner, more reliable package.

RE: I'll answer this one...
By jtyson on 2/22/2007 7:24:42 PM , Rating: 2
I didn't see where Kojima mentioned graphics at all in that quote. Maybe I should read it again...

... nope. Still nothing...

I think what Kojima was referencing was the ability to make games more in-depth and imaginative, ultimately adding to the gameplay. But then again, it sounds like we're reading different quotes...

RE: I'll answer this one...
By StevoLincolnite on 2/23/2007 7:47:40 PM , Rating: 1
Actually he was referring on how hard it is to code for a game on the cell, And how that will lower games quality as games get rushed out the door. It makes sense really Game developers are yet to start making multi threaded games efficiently and it seems the Cell is pushing that to the extreme not only do allot of developers suddenly have to start multi threading their game but they also have to change their code even more to make it run on the SPE's This means, Physics, Sound, Geometry Calculations, A.I, all of those things have to be split up with and be executed on each specialized SPE' What makes it worst is that each SPE' cannot directly talk to each other which creates another hurdle and will increase latency, Simply put, your having a 3 way phone call, Instead of talking directly to both of the other people, You have to hang up, call one, talk to them, hang up, then the person you talked to has to call the 3rd person tell them what you said etc. See where I'm going?
All of this takes time, Developers have to do allot of trial and error, And because they usually are timed constrained, Games will get rushed out the door and will be less quality than what it could have been.

What they need on the PS3 is a Software layer, So that developers will be making a game with 7-8 threads or more etc, And the software layer will automatically direct the required thread to a spe or something. (Thats my imagination going to work, not sure if it would work in the real world).
Or somehow only make a game single threaded and the software layer breaks it up into chunks and sends it all to the SPE's perhaps?

The high moon chief!
By yacoub on 2/22/2007 8:50:14 AM , Rating: 5
"High Moon chief" sounds like a Native American legend.

Interview w/High Moon Cell Research Senior Architect
By Carl B on 2/22/2007 11:56:16 AM , Rating: 2
Interviewed Mike Acton a while ago, who was the senior research guy at High Moon in charge of Cell. You can read some of the ways High Moon is really tackling this, and Acton himself is a top-notch coder.

He went over to Insomniac this year to lead their engine development team, but it's an awesome interview lending a lot of insights into Cell, how to program for it, and what High Moon is doing to build tech around it.

Tapping the Power of Cell:

By msva124 on 2/22/2007 2:35:38 PM , Rating: 2
More like a demonstration of how much someone with a big ego can ramble when given the chance. I'm not sure who that article is targeted towards. For beginning game programmers it is too complex, and for experienced game programmers it is unnecessary. Sony already has a nice big optimization manual. The problem is that they can't be bothered to rewrite all of their code for a completely new architecture when a crappy port from their Xbox 360 version will suffice.

By jodhas on 2/22/2007 6:14:24 PM , Rating: 2
The success of Wii wasn't really dependent on the spec pissing contest that PS3 and Xbox360 fans are engaged in, but rather the lost art of "gameplay" and "innovation".

The reason why the titles such as the Zeldas, Marios, GTAs and Halos were popular and fun were because of their gameplay! There were surely better looking games out during those times, but those games won the hearts and mind of the gamers because of the gameplay.

Now... having said that, if the developers are not enthusiastic about a particular platform and begin to withdraw development, then you will have problems.

Again, good gameplay can CARRY a console through generations while today's best graphics will never be tomorrow's best.

RE: Gameplay
By bigbrent88 on 2/22/2007 7:36:19 PM , Rating: 2
I'm going to take a stab at this and compare the Cell to any DX10 GPU. Based on DX10 being designed around unified shaders, where they can reinstructed to do any work on the GPU isnt that what an SPU is doing? Couldnt you relate next gen GPU programming to Cell programming?

Game Engines?
By Mitch101 on 2/22/2007 9:11:33 AM , Rating: 2
Isnt that what game engines help resolve? A good quality game engine would allow the developer to work on other aspects as a quality game engine would solve the difficulties a company might have in creating the core of the game?

Maybe IBM needs to get together with Carmack and Sony to create a high end 3D engine? Not that Carmack wont produce something high quality but right now Sony should be kissing someone's booty for a good sound engine to build upon.

RE: Game Engines?
By EclipsedAurora on 2/22/2007 10:27:45 AM , Rating: 1
Saddly, so far Carmack is good at PC game engine in terms of visual quality only, but not efficiency. From Quake to Doom3, which Carmack's product can produce good graphic even with older hardward. Console engine is a different game. Developers always need to work with the same hardware and get best possible quality out of it. Therefore historiclly best console engine did never come from Mr. Carmack's hands!

By encryptkeeper on 2/22/2007 9:51:36 AM , Rating: 2
its complex architecture also makes it difficult for developers to program their games

Um...Sony you might have made a goof here. Difficulty in programming for the hardware is a large part of what caused Nintendo to lose a lot of third party support in the N64 era. And yes, I know there were others like the resistance to the ROM cartridge format.

By UsernameX on 2/23/2007 11:35:52 AM , Rating: 2
I've read quite a few posts here basically saying that the Cell processor is good in some ways and bad in others. This is true of ANY processor. I mean common guys the real proof is in the games and what we play. What we see on paper (in most cases) isn't what is truly being displayed before our eyes. Find a game that's sold on both consoles, that has such a noticeable difference in quality, (weather it be graphics, AI, etc,) that it would prevent someone from buying the console... you probably wont find that many.

If you go back to the PS2 - XBOX days... even then you saw some games that looked better on the XBOX (Sports games) and some games that looked better on the PS2 (Burnout). But it really wasn't enough to change anyones mind about dropping down an extra couple hundred to buy a console. (I'm not saying all people, but the majority)

Sony PS3 and Cell are finished
By Lexington on 2/24/2007 12:26:34 AM , Rating: 2
The joke is on sony, stringer and IBM. Talk about egg on face. Sony got suckered by IBM into buying into this cell. It is too hard to program and so complext and started Sony on this silly over expensive over complext gamebox. I was wadering around best buy today, PS3s for sale, WIIs sold out and xbox360 being bought by a couple customers. In 3 years WII will be 40% of market, xbox 40% and sony PS3 < 20%. The games will be far and few, developers will be focusing on xb360 for the high end and fun games for WII. PS3 will be where gamecube was for the 2nd generation. The disaster is that they are losing their shirts on the hardware. Stringer will be fired before 2008 arrives for the fiasco.

Another sucker fooled again by IBM

Marcus you're an idiot.
By somegeek on 2/22/07, Rating: -1
RE: Marcus you're an idiot.
By tuanming on 2/22/07, Rating: -1
Whose being quoted in the last paragraph?
By SunAngel on 2/22/07, Rating: -1
RE: Whose being quoted in the last paragraph?
By Marcus Yam on 2/22/2007 7:21:09 PM , Rating: 3
If you do not see a closing quotation mark at the end of a quote, then the following quote is from the same speaker.

In case it is still unclear to you, the person speaking in the last quote is Hideo Kojima.

By jtyson on 2/22/2007 7:36:23 PM , Rating: 3
That's what you call getting chumped with class.

The sad part is that if he's too simple to grasp fifth grade literary concepts, he won't be able to understand that he's been chumped, for the same reason.

By Sungpooz on 2/22/07, Rating: -1
RE: what?
By Roy2001 on 2/22/2007 12:46:50 PM , Rating: 1

"Vista runs on Atom ... It's just no one uses it". -- Intel CEO Paul Otellini
Related Articles

Most Popular ArticlesSmartphone Screen Protectors – What To Look For
September 21, 2016, 9:33 AM
UN Meeting to Tackle Antimicrobial Resistance
September 21, 2016, 9:52 AM
Walmart may get "Robot Shopping Carts?"
September 17, 2016, 6:01 AM
5 Cases for iPhone 7 and 7 iPhone Plus
September 18, 2016, 10:08 AM
Update: Problem-Free Galaxy Note7s CPSC Approved
September 22, 2016, 5:30 AM

Copyright 2016 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki