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

RUMOR: Sony chooses Imagination Technologies for the PS4

H_Prestige said:
But Cell is just another EE. Why should they spend time putting band aids on a CPU that is for all practical purposes "gimped" when they could just go with a symmetrical multi core CPU that can run everything no problems?

God of war and uncharted 2 both prove nothing about the cell is gimped.
 
H_Prestige said:
But Cell is just another EE. Why should they spend time putting band aids on a CPU that is for all practical purposes "gimped" when they could just go with a symmetrical multi core CPU that can run everything no problems?

Something that is a games console doesn't need everything in the same ratio. Some characteristics are more needed in games than others
 
The cell cost should be quite low when next consoles are launched, so it may be useful to have 2 or 3 cells, one for main purposes, another to help GPU with AA, postprocessing or to do raytracing, and another one for physics. I think that would be a better solution than a standard multicore CPU as cell does some specific things very well and that does really fit games.

And those games would suddenly look like crap on an Intel CPU and any modern GPU?
the thing is they look great on a 2007 cpu and gpu. Consoles usually have a long life, 5-6 years. And Cell, is really pushing PS3 despite RSX being a bit worse than 360'sGPU and double pool memory.
 
NinjaFusion said:
God of war and uncharted 2 both prove nothing about the cell is gimped.

There's more than one way to be gimped. The Cell, in practice, is too difficult to extract the best possible results from unless you target it in a very specific way, which means that for the purposes of third-party developers it is "gimped" -- it takes more work to produce the same level of performance.

Unless there's some reason to believe that no less-baroque architecture could've been powerful enough to produce the equivalent of some of the (truly impressive) first-party PS3 titles, "our own dedicated teams did good work with it" isn't really the most important factor for Sony to consider in determining the architecture of their next system.
 
People think that they should abandon Cell now? :lol

Considering developers are getting the hang of it now and that it gives PS3 the edge when it comes to physics and graphics...

They should improve Cell and use it in PS4 IMO.

Just as with home Cell will be much better next generation...
 
tzare said:
The cell cost should be quite low when next consoles are launched, so it may be useful to have 2 or 3 cells, one for main purposes, another to help GPU with AA, postprocessing or to do raytracing, and another one for physics. I think that would be a better solution than a standard multicore CPU as cell does some specific things very well and that does really fit games.

Wouldn't it be even harder to code for something like that than it already is with just one single Cell?
 
charlequin said:
There's more than one way to be gimped. The Cell, in practice, is too difficult to extract the best possible results from unless you target it in a very specific way, which means that for the purposes of third-party developers it is "gimped" -- it takes more work to produce the same level of performance.

Unless there's some reason to believe that no less-baroque architecture could've been powerful enough to produce the equivalent of some of the (truly impressive) first-party PS3 titles, "our own dedicated teams did good work with it" isn't really the most important factor for Sony to consider in determining the architecture of their next system.

It takes "more work" simply because they chose to develop for the 360 first. If lead was on PS3 first, it would take them "more work" to get identical perfs on the 360. This has nothing to do with the hardware itself.

antiloop said:
People think that they should abandon Cell now? :lol

Considering developers are getting the hang of it now and that it gives PS3 the edge when it comes to physics and graphics...

They should improve Cell and use it in PS4 IMO.

Just as with home Cell will be much better next generation...

Well the problem is that unlike the EE or any other console GPU out there, it wasn't simply develop for the PS3 alone and kind of died in other areas. If IBM and Toshiba give up, I'm not sure Sony is willing to keep it on their shoulders alone. Which would be a shame, because it's apprently a very potent CPU, unlike early reports were implying.
 
antiloop said:
People think that they should abandon Cell now? :lol

Considering developers are getting the hang of it now and that it gives PS3 the edge when it comes to physics and graphics...

They should improve Cell and use it in PS4 IMO.

Just as with home Cell will be much better next generation...

Is STI even working on Cell anymore?
 
H_Prestige said:
But Cell is just another EE. Why should they spend time putting band aids on a CPU that is for all practical purposes "gimped" when they could just go with a symmetrical multi core CPU that can run everything no problems?

If Cell were so useful, wouldn't it be more widely adopted by now? Why isn't IBM enthusiastic about anymore?

I think u have remember Sony put lot of money in developing Cell and i believe they dont have to pay any royalty on it.. so they do get it cheap.. Secondly wot u r missing is that IBM is more of a consultant company, they r selling what ever customer wants.. if customer wants Linux they are not going to push for Windows.. So for avg user all the content is on i86.. so why they should push Cell when their is no app/software/tool for it.? n why customer develop anything for it when they dont know how long sony is going to produce it. :)

Lastly and which is most imp, games run on 1920 n 1024 or something like that.. so in the end u dont need a hugely powerful machine for it.. even if they go 3D they will need 120 frames per second instead of 60fps.. so a fine tuned 2x powerful Cell will be able to easily handle it.. until or unless they dont produce 3d holographic displays with 12000 x 8000 resolution console manufactures dont need powerful machines..

The important thing is that the whole infrastructure of Sony Entertainment is getting build around Cell and if they dump it and goes for a new architecture than i think its a lot of pain and maynot be worth it.. Sony maynot be able to rebound second time from PS3 like launch.. They need to have awesome graphics, they need to have a solid library from get go and they need to put more developers on developing content for consumer than building the backend on new architecture.
 
I'd say they should go with PC-like architecture and keep the cell for backwards compatibility. The PS3 was pretty expensive at launch mainly because of the Blu-ray drive. Now that Blu-ray has already won the format wars, I think it will be the standard. And the standard formats usually go down in price dramatically in terms of manufacturing the discs and the drives. Also, I'm sure Sony has learned their lesson from this generation.

Having PC-like architecture would make it much easier for developers and it would attract western PC devs and having the cell for backwards compatibility ensures the continuous support for the PS3. They should launch at $399 and bundle the Eye 3 and Move in addition to the dual shock 4 and I'm sure it'll be successful.
 
hxa155 said:
I'd say they should go with PC-like architecture and keep the cell for backwards compatibility. The PS3 was pretty expensive at launch mainly because of the Blu-ray drive. Now that Blu-ray has already won the format wars, I think it will be the standard. And the standard formats usually go down in price dramatically in terms of manufacturing the discs and the drives. Also, I'm sure Sony has learned their lesson from this generation.

Having PC-like architecture would make it much easier for developers and it would attract western PC devs and having the cell for backwards compatibility ensures the continuous support for the PS3. They should launch at $399 and bundle the Eye 3 and Move in addition to the dual shock 4 and I'm sure it'll be successful.

This is pretty much what I think they should do. Except for move/eyetoy. Keep that as a peripheral and dualshock the standard. They're not going to out Nintendo Nintendo. If MS wants to try their luck with their Natal box, let them.

Also, I'm really struggling to keep up with the logic of some posters here. People think it's better to go with harder to work with, non standard CPU's rather than standard Intel CPU's which have better support, are simpler to work with, smaller, and more power efficient. I mean, who the hell complains about Intel not having enough power for games?
 
Why would Sony continue to use Cell for anything in PS4 other than BC, when IBM has dropped Cell development?

Raist said:
It takes "more work" simply because they chose to develop for the 360 first. If lead was on PS3 first, it would take them "more work" to get identical perfs on the 360. This has nothing to do with the hardware itself.

But didn't they do 360 first since it was a more familiar platform/architecture with more robust and mature tools already at hand? Had they done PS3 first, 360 would still have been easier.
 
The only reason to hold on to Cell is backwards compatibility. It was and still is a horrible decision as far as CPUs go for a game system. It may be good for scientific processing or video decoding, but it just seems ill suited for something a branchy as video games.

Here's a question for those more knowledgeable than myself. Would a PPE with a local cache be capable of emulating an SPU, or more concisely, does the PPE support the same SIMD instructions that the SPU is built for? Would a custom solution that included 8 PPE(2?)s be able to emulate a Cell? If the PPE supports those same SIMD instructions, then it seems like a somewhat elegant solution to the backwards compatibility problem. Past that, as long as the GPU is capable of the gymnastics it would take to emulate the PS2's GPU, then the entire history of the Playstation brand is there. Pop in a disc, or download it from PSN.

Alternately, PS3 backwards compatibility could be completely forgone, but for the fact that it's an amazing hook for prospective buyers. You can still play your old games; they'll load faster and may even look better. How many truly iconic games does the PS3 have that aren't 1st party? Could updated executables be provided for certain selected games? If you could run MGS4, FFXIII, Uncharted 1/2, the GoW games, LBP, and Trico, would that be enough?
 
Lets not forget the incredible amount of value that was created by developers who mastered cell. Developers now know how to program for the system. And that is worth a LOT.

If you have the possibility you should opt for the same architecture you should. This is also true for MS. Try to scale it up.

The only thing that doesn't make it automatic is that there are new technologies that you might not be able to integrate in the old architecture.
 
bmf said:
The only reason to hold on to Cell is backwards compatibility. It was and still is a horrible decision as far as CPUs go for a game system. It may be good for scientific processing or video decoding, but it just seems ill suited for something a branchy as video games.

Here's a question for those more knowledgeable than myself. Would a PPE with a local cache be capable of emulating an SPU, or more concisely, does the PPE support the same SIMD instructions that the SPU is built for? Would a custom solution that included 8 PPE(2?)s be able to emulate a Cell? If the PPE supports those same SIMD instructions, then it seems like a somewhat elegant solution to the backwards compatibility problem. Past that, as long as the GPU is capable of the gymnastics it would take to emulate the PS2's GPU, then the entire history of the Playstation brand is there. Pop in a disc, or download it from PSN.

Alternately, PS3 backwards compatibility could be completely forgone, but for the fact that it's an amazing hook for prospective buyers. You can still play your old games; they'll load faster and may even look better. How many truly iconic games does the PS3 have that aren't 1st party? Could updated executables be provided for certain selected games? If you could run MGS4, FFXIII, Uncharted 1/2, the GoW games, LBP, and Trico, would that be enough?

Speaking of which, how does the pc emulator handle the GS problem? It seems to work pretty well and apparently you don't need a screaming GPU to run it. Should we expect ps2 emulation to be a problem on the ps4?
 
bmf said:
Alternately, PS3 backwards compatibility could be completely forgone, but for the fact that it's an amazing hook for prospective buyers. You can still play your old games; they'll load faster and may even look better. How many truly iconic games does the PS3 have that aren't 1st party? Could updated executables be provided for certain selected games? If you could run MGS4, FFXIII, Uncharted 1/2, the GoW games, LBP, and Trico, would that be enough?


3rd party games are also iconic, GTA4, COD4, MW2, etc. And don't forget PSN games. They need backwards compatibility to have a successful launch.
 
Drastic said:
But didn't they do 360 first since it was a more familiar platform/architecture with more robust and mature tools already at hand? Had they done PS3 first, 360 would still have been easier.

The fact that the 360 launched earlier is probably a more likely reason.
 
lowrider007 said:
eh? since when?

Since November 2009. Probably because Sony dropped Cell. Like someone pointed out, IBM is a work-for-hire consultant company these days, so they do pretty much what people order from them. The PS4 speculation said that in 2008-2009 Sony was exploring different architectures, including Cell continuation. In November 2009 news came out that IBM won't do much anything with Cell. So probably Sony told them that they chose another path.
 
Raist said:
The fact that the 360 launched earlier is probably a more likely reason.

So, is your argument that had PS3 come out at the same time as 360 (or earlier), then PS3 would have been easier for the devs, even considering the different architecture?
 
hxa155 said:
3rd party games are also iconic, GTA4, COD4, MW2, etc. And don't forget PSN games. They need backwards compatibility to have a successful launch.
I totally agree that my list is really incomplete. I'm also very conscience that may 3rd parties would not want to do this at all. They want to sell the next version of their yearly franchise. As far as PSN games go, do many really dig into the SPUs? Would being able to recompile for something that isn't running to the metal on SPUs be enough? Given that the PS3 includes a farily robust OS that the games run on top of, I'm going to assume that they will want to carry most of that over, especially the APIs.

EDIT: As far as 3rd party game lists go, would scraping the top 20 or 30 off of metacritic be sufficient?
 
Backwards compatibility is a must next gen, given the nature of PSN and XBL. Next gen, your account will still have to show trophies for past gen games, still allow you to redownload any games/movies you bought through the store, etc. You are tied to your account, and they will have to support your account fully with future iterations of PlayStation/Xbox. Removing that would be like Steam suddenly dropping support for games more than five years old. The only exception might be handheld games vs console games (ie, PSP games would work on PSP2 but not PS3, though they may allow some cross platform play for older titles, like playing PS1 games on either PSP or PS3).

Actually, thinking about this gives me hope that PS2 BC will return one day in software form, if not for the PS3 than for the PS4. The PS4 will presumably be able to play PS1/PS3/PS4/PSN titles, and it'd be stupid to let you play all but PS2 games.
 
Chittagong said:
Since November 2009. Probably because Sony dropped Cell. Like someone pointed out, IBM is a work-for-hire consultant company these days, so they do pretty much what people order from them. The PS4 speculation said that in 2008-2009 Sony was exploring different architectures, including Cell continuation. In November 2009 news came out that IBM won't do much anything with Cell. So probably Sony told them that they chose another path.

Makes total sense when you think about it, sense Sony was the number one Cell buyer.
 
H_Prestige said:
Speaking of which, how does the pc emulator handle the GS problem? It seems to work pretty well and apparently you don't need a screaming GPU to run it. Should we expect ps2 emulation to be a problem on the ps4?
They way I understand it is that it involves doing a lot of setup on the CPU, then passing off to the GPU to do actual rendering and display. I may be misinformed though.
 
makingmusic476 said:
Actually, thinking about this gives me hope that PS2 BC will return one day in software form, if not for the PS3 than for the PS4. The PS4 will presumably be able to play PS1/PS3/PS4/PSN titles, and it'd be stupid to let you play all but PS2 games.

PS2 emulation on PC is already pretty good, so if they went with PC-like architecture for the PS4, that provides PSOne and PS2 emulation. I honestly think that's the best solution.
 
jett said:
Ken Kutaragi am cry. I wonder what is the guy doing these days.

Probably sipping a cocktail on a beach somewhere while managing his phantom company.

Drastic said:
So, is your argument that had PS3 come out at the same time as 360 (or earlier), then PS3 would have been easier for the devs, even considering the different architecture?

That's likely. I mean the PS2 had a fucked up architecture and devs didn't jump ship when the xbox and gcn came out. Especially the xbox which was basically a PC.
 
Raist said:
That's likely. I mean the PS2 had a fucked up architecture and devs didn't jump ship when the xbox and gcn came out. Especially the xbox which was basically a PC.

I think that's just because the PS2 was selling amazingly well.
 
I feel like a lot of people are assuming the only options are a new version of the CELL or going x86 PC architecture.

Why wouldn't Sony simply use the latest non-CELL IBM CPU, along with a Nvidia or ATi GPU?
 
Raist said:
That's likely. I mean the PS2 had a fucked up architecture and devs didn't jump ship when the xbox and gcn came out. Especially the xbox which was basically a PC.

But I thought the argument was why the PS architecture was more difficult to program for, not just which was first/ported? Were devs saying the OG Xbox was more difficult to program than PS2?
 
panda21 said:
also i think they would be insane not to drop cell, since IBM are slowly dropping support for it completely, which i doubt they would be doing if it was actually a good architecture.

If they drop Cell, then they also drop PS3 backward compatibility. That would be insane, because they spent a good chunk of the last 3½ years building a library of downloadable games on the PlayStation Store. Most of those games (except for PS1 classics and maybe Minis) would be unplayable on a system that didn't contain (or emulate) Cell.

I'd prefer they did what Nintendo did this generation with Wii (being a smaller upgrade from GameCube, built on the same architecture), and go for a smaller upgrade in power but retain Cell (and therefore retain backward compatibility), rather than go for a huge upgrade in power with a new architecture and throw out compatibility with PS3 and PSN games.
 
Drastic said:
But I thought the argument was why the PS architecture was more difficult to program for, not just which was first/ported? Were devs saying the OG Xbox was more difficult to program than PS2?

Come to think of it, wouldn't the OG Xbox be easier to code for than the PS2 due to the Xbox being more similar to a PC than the PS2 was?
 
yurinka said:
I think it's false because it make sense, the only think I don't like is the release date. I would launch it in late 2012 / early 2013.

$399 / 399€
-Same CELL CPU, just faster or with more cores
-4GB or more of PC standard RAM DDR3-2.133
-Some good Nvidia GPU
-1TB HDD
-WiFi 802.11n, USB 3.0, Bluetooth 2.0 (or 3 if available), HDMI 1.4
-EyeToy 3, Move 2, DualShock 4 included with console
-Full PS1,PS2,PS3,PSP,PSP2/Playstation Phone/PlaystationPad,Neogeo, PCEngine & Sega consoles BC
-All current PSN services + new ones. You can use your current account and redownload your PSN stuff (including trophies) purchased in PS3.
-Cheaper transition for developers since they would use the same PS3 devkits, toolkits etc
-Storage in both HDD and Playstation Cloud.
-Obviously 3D ready

So Sony would copy Nintendo, since PS4 would be 2 (4 in this case) ducktaped PS3.

Best guess of what PS4 will be like IMO. Nothing too wild but a logical upgrade of the basis of what is currently in PS3. It will be sold at a slight profit, following Wii model, Sony can't and won't be as silly as they were with PS3 launch.
 
Agent X said:
If they drop Cell, then they also drop PS3 backward compatibility. That would be insane, because they spent a good chunk of the last 3½ years building a library of downloadable games on the PlayStation Store. Most of those games (except for PS1 classics and maybe Minis) would be unplayable on a system that didn't contain (or emulate) Cell.
Remember Sony is the company that axed PS2 BC and decided that the PSP Go could be released without a UMD Drive. They really don't give a shit about BC.

That doesn't mean that you won't be able to play your current gen Uncharted games, Killzones and Little Big Planets, they could just re-release those titles. People who think that they're able to play all their current generation content on next generation systems will end up being dissapointed. The library and backlog of most gamers would keep them busy for months if not years so they're not going to buy that many new games. It's the difference between 20 games or 50 games. Not to mention that most current gen games will be useless considering they're mostly based on online multiplayer and by 2012 most servers of current gen games are probably shut down or empty.
 
Agent X said:
If they drop Cell, then they also drop PS3 backward compatibility. That would be insane, because they spent a good chunk of the last 3½ years building a library of downloadable games on the PlayStation Store. Most of those games (except for PS1 classics and maybe Minis) would be unplayable on a system that didn't contain (or emulate) Cell.

I'd prefer they did what Nintendo did this generation with Wii (being a smaller upgrade from GameCube, built on the same architecture), and go for a smaller upgrade in power but retain Cell (and therefore retain backward compatibility), rather than go for a huge upgrade in power with a new architecture and throw out compatibility with PS3 and PSN games.

The problem is that if they just keep upgrading cell for each new playstation, what are they going to do when they eventually have no choice but to move to a new architecture? They will have all this legacy software that they somehow have to emulate on completely different architecture. And Cell is a dead end, Sony can't keep using it forever. It's better to make the break now and just have to deal with ps3 b/c. If they can't emulate it, then they can put the cell in the ps4 strictly for ps3 b/c.

Remember Sony is the company that axed PS2 BC and decided that the PSP Go could be released without a UMD Drive. They really don't give a shit about BC.

I don't think they don't care about BC. I'm sure they understand the importance. If the PR people say otherwise and say it's better to focus on ps3 games, blah blah blah, it's just to justify the removal of ps2 BC. That sounds better than saying "shit, we're bleeding money and we had no other hardware to throw out of the ps3". Removing ps2 BC was just an unfortunate result of boneheaded decisions made earlier.

And psp go means nothing. It's not the standard psp and never will be. It's just another psp sku. A dumb one, but just another sku.
 
Combichristoffersen said:
Come to think of it, wouldn't the OG Xbox be easier to code for than the PS2 due to the Xbox being more similar to a PC than the PS2 was?
That's my point.

For Raist's argument to hold water, the lead development console would always be easier to develop on since it would be first.

I think the reality is, the architecture that devs are already familiar with, have mature tools for, or at least similar tools, etc. are naturally going to be easier to work with.
 
DarkoMaledictus said:
Yeah I hope they dump cell... its really not helping them at all going against the norm. They would gain a lot more support just by going pc... Ok backward comp will suffer but guess thats why we have emulators ;p!

But, who's not supporting the PS3? Honestly, everything comes to PS3.

So at this point, everyone understands cell, and by the time PS4 rolls out, there will be a LOT of devs who are experts. They are already through the biggest hurdles. I personally hope they stick with cell at this point.
 
IBM have not stopped development of the Cell, just the advanced 32 core family.

You do not invest billions of R&D dollars into an architecture design to totally dump it a number of years later. The Cell will be driving the PS4, but the surrounding memory addressing architecture will be different to simplify things for devs. They will move to a pseudo-UMA. In fact, the idea of discrete memory blocks for each SPU/PPU I reckon will likely be the way to go. I think 16 SPU core, each with 512Mb standard addressable memory block. Update the GPU with a strong scaleable solution with discrete memory allocation and you will have everything you need, including enough bandwidth to run PS1, PS2 emulation as well as having a simpler way to provide PS3 BC.

The only area Sony got caught out this gen is Online and price. They need to provide standardised infrastructure within PSN so that devs are able to include these features across every game and not rely on them developing their own from scratch. In regard to price, don't use the system as a trojan horse for an advanced architecture and they sould be fine.
 
[Nintex] said:
Remember Sony is the company that axed PS2 BC and decided that the PSP Go could be released without a UMD Drive. They really don't give a shit about BC.

They need to be concerned, though. There's a good chance both Nintendo and Microsoft will be backward compatible next generation. If Sony throws that away, then they risk handing both Nintendo and Microsoft a huge competitive advantage, particularly when it comes to selling downloadable games.

H_Prestige said:
The problem is that if they just keep upgrading cell for each new playstation, what are they going to do when they eventually have no choice but to move to a new architecture? They will have all this legacy software that they somehow have to emulate on completely different architecture. And Cell is a dead end, Sony can't keep using it forever. It's better to make the break now and just have to deal with ps3 b/c. If they can't emulate it, then they can put the cell in the ps4 strictly for ps3 b/c.

They don't have to "upgrade Cell". They'll probably have to emulate it at some point anyway. By the time PlayStation 5 rolls around, it should be capable of emulating PS3 in software.

Meanwhile, they need to do something for the PlayStation 4, whether it includes an actual Cell processor (as early PS3 models did with PS2 chips this generation), or some sort of emulation. Cell doesn't need to be the PS4's main processor. But it needs to be in there somewhere.

BTW, if Sony hadn't removed the PS2 chipset from newer PS3 models, they'd be in a good position to sell downloadable PS2 classics now.
 
Agent X said:
They need to be concerned, though. There's a good chance both Nintendo and Microsoft will be backward compatible next generation. If Sony throws that away, then they risk handing both Nintendo and Microsoft a huge competitive advantage, particularly when it comes to selling downloadable games.



They don't have to "upgrade Cell". They'll probably have to emulate it at some point anyway. By the time PlayStation 5 rolls around, it should be capable of emulating PS3 in software.

Meanwhile, they need to do something for the PlayStation 4, whether it includes an actual Cell processor (as early PS3 models did with PS2 chips this generation), or some sort of emulation. Cell doesn't need to be the PS4's main processor. But it needs to be in there somewhere.

BTW, if Sony hadn't removed the PS2 chipset from newer PS3 models, they'd be in a good position to sell downloadable PS2 classics now.

Yeah, that's basically what I'm saying. Include the ps3 cell in the ps4, but not as the main cpu. It should just be there for BC. If Sony is smart with how they engineer the ps4, and not try to make it do more than it should, there wouldn't be a need to toss out the cell later on like what happened with the ps3.
 
Drastic said:
That's my point.

For Raist's argument to hold water, the lead development console would always be easier to develop on since it would be first.

I'm just saying that the whole "too hard to program for" and "easier because it's like PC" basically means nothing. The primary factor is the first console of a new gen being released.
When everyone was releasing games left and right for the PS2, the 360 was completely different and devs had to adapt. Once they had their tools and engines running and the PS3 came out, leading development on the 360 was what made most sense for them. It has nothing to do with the actual difficulty of developing games for a system.
Had the PS3 been released a year before the 360, developers would have started getting experience and tools for it, and the PS3 would likely be the lead platform in most cases, and engines not really adapted to the 360's architecture. I really doubt developers would have been like "hey, this new system looks like a PC, let's forget a year of work and develop for this one first"
 
H_Prestige said:
Yeah, that's basically what I'm saying. Include the ps3 cell in the ps4, but not as the main cpu. It should just be there for BC. If Sony is smart with how they engineer the ps4, and not try to make it do more than it should, there wouldn't be a need to toss out the cell later on like what happened with the ps3.
This already exists in enterprise business. There are compute/render nodes which use both an AMD Opteron and a Cell, with the Opteron acting as the main CPU. The main program is an X64 binary while the smaller job/task programs are Cell SPU binaries. It's an amazingly powerful and flexible combo.
 
I'm sure developers don't like it, but it is nice having a bit of an exotic architecture to the PS3. CELL seems to be a pretty good CPU, as when it is in the best hands the PS3 is doing things the 360 dreams of accomplishing.

Besides, most devs are trained to work on PS3 nowadays, might as well extend it. Just get a better GPU in PS4.
 
Raist said:
I'm just saying that the whole "too hard to program for" and "easier because it's like PC" basically means nothing. The primary factor is the first console of a new gen being released.
When everyone was releasing games left and right for the PS2, the 360 was completely different and devs had to adapt. Once they had their tools and engines running and the PS3 came out, leading development on the 360 was what made most sense for them. It has nothing to do with the actual difficulty of developing games for a system.
Had the PS3 been released a year before the 360, developers would have started getting experience and tools for it, and the PS3 would likely be the lead platform in most cases, and engines not really adapted to the 360's architecture. I really doubt developers would have been like "hey, this new system looks like a PC, let's forget a year of work and develop for this one first"

All else considered equal, I think it's safe to say the PS3 is more difficult to develop on than the 360. At least it sure would seem so from developers' comments. Sure, whichever platform comes first gets a head start, but that doesn't mean that platform is easier to work with overall. I think the PS3 had two things working against it - released later and more difficult to work with.

A more familiar/less complicated architecture is easier to work with no matter if it releases shortly before, after or the same time as a competing but less familiar/more complicated architecture. Which was released first can affect familiarity, but it doesn't affect complexity at all.

I think the bottom line is this: If you think the PS3 would be just as easy to work with as the 360 had they released at the same time, then we'll just have to agree to disagree.
 
Top Bottom