-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
vg247-PS4: new kits shipping now, AMD A10 used as base, final version next summer
- Thread starter Globox_82
- Start date
At least one of them could be.
Evolution_1ne
Banned
why is it so hard to accept that the PS4 isn't going to launch worldwide until 2014?
James Sawyer Ford
Member
It's not hard to accept if there were any genuine information pointing in that direction.
I would rather the PS4 come out in 2014 or even 2015 if need be but there isn't much that says that the PS4 isn't coming in 2013 so I don't see why anyone would have to accept that it isn't coming out in 2013 just yet.
Barack Lesnar
Banned
From a hardware standpoint I'd like it to launch in 2014 assuming it would mean newer tech and hardware, but from a sales perspective I'd worry about it being overlooked with 720 taking the Fall/Holidays all by itself. I want it to have a good start and early momentum, I think that's important after PS3 sales disappointed.
-Pyromaniac-
Member
It makes no sense to launch right after christmas. Giving your competition 2 (nintendo) and 1 (MS) holidays advantage over you is bonkers.
Barack Lesnar
Banned
I doubt they really consider Wii U direct competition. While there is some overlap, they mostly cater to different demographics. It doesn't matter if Sony let em have 3 Xmas'... who's #1 or #2 isn't important when console gens seem to trending towards fractured releases (i.e., one releasing early, one in the middle, and so on).
I doubt steamroller was going to be in PS4 anyways. I guess that reinforces the Jaguar rumour.
Barack Lesnar
Banned
I love the code names they give these things...
Bulldozer
Piledriver
Steamroller
Excavator
lol
Bulldozer
Piledriver
Steamroller
Excavator
lol
Evolution_1ne
Banned
becasue Sony saying it multiple times, and final dev kits not release to mid next year isn't enough?
Sony says alot of things.
True, and what Sony has to say about something usually turns out unrealistically optimistic.
That's the plan. The porting nightmares and lead console redundancy is over.
I guess going x86 was a way of easing developers pains and also cutting R&D costs, I might not like it but maybe some customization can turn the x86 to a gaming apt CPU...
Also, if that means shorter development cycles, that will be great, no more TLG or FFXIIvs BS please..
James Sawyer Ford
Member
Near Final dev kits are releasing in January according to the rumor, with the final versions releasing around summer of next year....that suggests to me a 2013 release.
KojiKnight
Member
Final dev kits in summer means late 2013 release at the earliest... not a guarantee 2013 release. Final Wii U kits were launched just a month or two before E3 and it releases in a week... If the final PS4 dev kit hits later than E3, I wouldn't expect the system till after march next year (there is no point in releasing a new game system in the final quarter of the fiscal year since it has a high chance of bringing down your financial results)
Well in that case Sony havent said anything. Just a rumour.
Iacobellis
Junior Member
It's weird how everyone but Nintendo is now announcing their new systems the same year as release. I remember the PS3 got exposure in 2005 and Nintendo even made a small mention that they were working on the next console at E3 2004, two and a half years before the Wii hit shelves.
Nizz
Member
Then this means Sony will be behind again? I really feel they need to release close to the next X-Box with Microsoft apparently getting out by the holidays next year. Another head start for MS.
Yea, a few months at the most. It all depends upon what software will be shown off for PS4 during next E3 to ensure that PS4 gets good initial sales despite being out later and after holidays 2013.
Also aren't there reports of MS having problems with the dies that could push the next xbox to miss Christmas next year?
KojiKnight
Member
I don't think the "headstart" matters nearly as much as people think it does. Obviously Sony closed the gap between the 360, and even if they hadn't they'd have still made profits on it.
I'm also not saying they can't get the system out by 2013, but I just don't see it happening if the final dev kit only hits 3-4 months before launch. I also don't see them if it goes into 2014 releasing before the fiscal year ends. In general, Sony's 2013 business year is already probably not going to be the hottest, but tack in what will certainly be extra losses at the very end of it (that's January to March 2014) simply because the console will sell at a loss and there likely won't be enough software sales to cover it (there rarely are the first few months) and it just makes more sense for them to put it off till after March that way they have more time to accrue sales.
Yeah, final dev kits don't mean all that much for launch games. They'll know what the target specs are for a while beforehand.
Which brings us back to the Battlefield rumor.
EA knows alot more than what they are leading on.
Canis lupus
Member
Just finished MGS4 for a 2nd time, the game and cutscenes look so good for a 2008 game. Looking forward to next gen graphics.
I don't think it matters if PS4 comes out after the next Xbox. As long as it's not something like over 5 months after the next Xbox releases.Sony just has to hype up the PS4 and let people know it's coming not too long after the next Xbox.
The reason PS3 suffered from a bad start wasn't just because it came after, it was
- Too expensive, mainly because of the cell and bluray
- All the games are the same as 360 (many multiplats look inferior), this wouldnt matter if it wasn't as expensive
- Not many big exclusives launched on PS3 in its 1st year and 360 multiplats looked better
- 360 had it's 2nd wave of big games when PS3 launched, such as Gears of War
I think this time Sony could get away with PS4 coming after the next Xbox. I am sure price won't be that much more than the enxt Xbox as this time Sony are not introducing new tech like Cell or Bluray. Also platform support from 3rd partys won't suffer in the early years like PS3 did due to Sony likely going for something more easier to develop for. Finally Sony being more prepared when it comes to games, Sony probably expected tons of 3rd party support like PS1 and PS2, but that wasn't the case. Now Sony is expecting to just get multiplatform games and has invested a lot in their own studios. I'm sure Sucker Punch, Guerilla Games, Santa Monica, Polyphonic, Naughty Dogg and Evolution are all busy developing PS4 games.
The reason PS3 suffered from a bad start wasn't just because it came after, it was
- Too expensive, mainly because of the cell and bluray
- All the games are the same as 360 (many multiplats look inferior), this wouldnt matter if it wasn't as expensive
- Not many big exclusives launched on PS3 in its 1st year and 360 multiplats looked better
- 360 had it's 2nd wave of big games when PS3 launched, such as Gears of War
I think this time Sony could get away with PS4 coming after the next Xbox. I am sure price won't be that much more than the enxt Xbox as this time Sony are not introducing new tech like Cell or Bluray. Also platform support from 3rd partys won't suffer in the early years like PS3 did due to Sony likely going for something more easier to develop for. Finally Sony being more prepared when it comes to games, Sony probably expected tons of 3rd party support like PS1 and PS2, but that wasn't the case. Now Sony is expecting to just get multiplatform games and has invested a lot in their own studios. I'm sure Sucker Punch, Guerilla Games, Santa Monica, Polyphonic, Naughty Dogg and Evolution are all busy developing PS4 games.
The BF4 rumor doesn't mean all that much either. If it's cross gen game it could still launch on PS3/360/PC in October and then just have a "Next Gen Edition" come out day and date with the PS4 and next Xbox.
Planning BF4 for next gen just seems to imply that they'll be coming by spring. Any longer than that and it wouldn't be nearly as profitable to release it on them.
-Pyromaniac-
Member
I think we can all agree the 720 is launching by next christmas, right? Sony is just asking for disaster if they launch in January/feb/march/april/may/august. Talk about instant momentum killing. And there's no way they're going to wait till christmas 2014.
Could sony launch in january/feb/march? Yes. But is there a high probability of instantly killing momentum? Yes. Momentum means a lot. Just look at the vita. Never got any sort of traction and it's just dwindling off. That's what a holiday launch is for. High sales, builds momentum, and hopefully gets enough traction to get you through the slow summer months. Launching in those months is dangerous.
I see them doing everything in their power to launch by next holiday so I can't understand when people make it sound obvious that it will come out later.
Could sony launch in january/feb/march? Yes. But is there a high probability of instantly killing momentum? Yes. Momentum means a lot. Just look at the vita. Never got any sort of traction and it's just dwindling off. That's what a holiday launch is for. High sales, builds momentum, and hopefully gets enough traction to get you through the slow summer months. Launching in those months is dangerous.
I see them doing everything in their power to launch by next holiday so I can't understand when people make it sound obvious that it will come out later.
You could risk a launch with limited titles. If there isn't much difference between January devkits and the final ones, then the more conservative third party titles might be fine anyway (maybe with some rough edges), and sony's first party teams will presumably have access to kits earlier anyway.
jeff_rigby
Banned
Continuation of posting information on SOI Vs Bulk and High performance Vs. Low performance designs.
Summary to this point: Global Foundries 20nm is low power as leakage is an issue with bulk silicon. SOI supports both low power and High performance with less leakage but publicly AMD and TSMC will be bulk process at 20nm and catering to low power handheld designs.
What about high performance are there to be no Desktop high clock chips? Kaveri (desktop high performance 28nm bulk process) is delayed with the roadmap stressing Kabini 28nm low power. That tends to support 28nm bulk can't support high performance either.
Found this:
Linkedin Katherina Babich SMTS at GLOBALFOUNDRIES Greater New York City Area
In the news is only Bulk for GloFlo except for ONE article for a producer of SOI wafers saying they will be used by GloFlo, IBM and others for 20nm and 14nm contrary to the AMD CEO saying they are going Bulk for 28nm and smaller. The article also mentions Finfet and trigate devices just as the above Linkedin post. This latest article has links confirming the use of SOI and why as well as a production ramp up to supply demand.
High clock speeds generate heat and the higher the leakage in devices the higher the heat so generally SOI is needed if you want high clock speeds. This is the reason the Cell (3.2 ghz) was on SOI. Kabini on bulk process at 28nm has clock speeds for the Jaguar CPUs of 1.6Ghz. Partly to extend battery life, partly because at higher clock speeds the chip would overheat.
Will next generation need high performance SOI? My uninformed guess is yes. If so then next generation has to be 32nm SOI or 22nm SOI as that's all that I can find research work on SOI at IBM and GloFlo (correct if wrong but cite please).
ST Microelectronics transferred SOI technology to GloFlo so that they can make Mobile SoCs for them.
Highlights from the IEEE 2011 SOI Conference include presentations by ST, ARM, IBM, Intel, Leti, Peregrine, GlobalFoundries and more.
There are cites that contradict each other so it's confusing. I cited earlier that IBM was skipping SOI @ 28nm and below because the pressure to cause straining caused transistor junction damage reducing yield. IBM is going gate last instead of gate first below 28nm also because of yield issues; GloFlo following IBM's lead. Apparently at 22nm and below different processes will be used and SOI is again viable with IBM going SOI.
Summary to this point: Global Foundries 20nm is low power as leakage is an issue with bulk silicon. SOI supports both low power and High performance with less leakage but publicly AMD and TSMC will be bulk process at 20nm and catering to low power handheld designs.
What about high performance are there to be no Desktop high clock chips? Kaveri (desktop high performance 28nm bulk process) is delayed with the roadmap stressing Kabini 28nm low power. That tends to support 28nm bulk can't support high performance either.
Found this:
Linkedin Katherina Babich SMTS at GLOBALFOUNDRIES Greater New York City Area
22nm SOI Process Integration starting at GloFlo the same time Microsoft and Sony started next generation (2010) with Katherina Babich moving from IBM to GloFlo. In this time period Sony announced they will be skipping the 32nm node and another Linkedin IBM post has research on 22nm cell register arrays.
In the news is only Bulk for GloFlo except for ONE article for a producer of SOI wafers saying they will be used by GloFlo, IBM and others for 20nm and 14nm contrary to the AMD CEO saying they are going Bulk for 28nm and smaller. The article also mentions Finfet and trigate devices just as the above Linkedin post. This latest article has links confirming the use of SOI and why as well as a production ramp up to supply demand.
High clock speeds generate heat and the higher the leakage in devices the higher the heat so generally SOI is needed if you want high clock speeds. This is the reason the Cell (3.2 ghz) was on SOI. Kabini on bulk process at 28nm has clock speeds for the Jaguar CPUs of 1.6Ghz. Partly to extend battery life, partly because at higher clock speeds the chip would overheat.
Will next generation need high performance SOI? My uninformed guess is yes. If so then next generation has to be 32nm SOI or 22nm SOI as that's all that I can find research work on SOI at IBM and GloFlo (correct if wrong but cite please).
ST Microelectronics transferred SOI technology to GloFlo so that they can make Mobile SoCs for them.
Highlights from the IEEE 2011 SOI Conference include presentations by ST, ARM, IBM, Intel, Leti, Peregrine, GlobalFoundries and more.
There are cites that contradict each other so it's confusing. I cited earlier that IBM was skipping SOI @ 28nm and below because the pressure to cause straining caused transistor junction damage reducing yield. IBM is going gate last instead of gate first below 28nm also because of yield issues; GloFlo following IBM's lead. Apparently at 22nm and below different processes will be used and SOI is again viable with IBM going SOI.
if you launch in November, the first holiday isn't really a holiday. You'd be expecting to sell through your launch allocation regardless of when you launch. You need to ramp up your production (and software) first.
So if MS launch in November 2013, their first 'proper' holiday would be Fall 2014. If Sony release in March they would be OK I think.
But I don't see how a delay of 3-4 months would bring a big jump in performance. They should obviously try to avoid being too late to market though.
jeff_rigby
Banned
"But ARM’s showing that if you move to the next node and move to FD-SOI, you get really phenomenal results, especially at the lower supply voltages." "An existing bulk design can be directly ported to FD-SOI".
I don't think there is much of a problem releasing two consoles a few months apart. The first real test will be during holiday 2014. If Sony were to release their console for holiday 2014 with the 720 having a year headstart, then we would have issues, but I can't see that happening. Nov 2013 for 720 and Mar/Apr 2014 for PS4 is what I think will happen.
I don't expect a 2013 release for PS4, based on the fact that Sony still have a few big hitter games to release towards the end of next year on PS3, I don't see why they would want to pull the rug from under them, knowing that the PS4 will likely sell well for the first few months regardless of whether it is a holiday or spring launch.
jeff_rigby
Banned
The rumor has production to start after June 2013 for a release holiday 2013, there is not enough ramp-up time to get an inventory to meet demand for a 2013 holiday launch which is what mrklaw is pointing out. I think we are seeing an effort to get a presence in the market to counter the WiiU launch November 2012.mrklaw said:
So the choice I think is 28nm bulk or later 22nm SOI. The Kavini delay might indicate we are not getting a high performance SoC at 28nm (possible to have a SoC + GPU). One definition of low power/performance is less than 57 watts and high performance less than 90 watts in one SoC package. With 22nm SOI we could get high performance (high clock) and relatively low power in one package which if yields are good at 22nm (higher clock means fewer CUs which should increase yield) is perfect.
jeff_rigby
Banned
Yes you can for most designs. From the roadmaps though, 28nm is late and 20nm Low Power bulk is on track for 2013. We have no information for 22nm SOI but I expect it's the same roadmap as 20nm LP.
To have the cheapest design you want one SoC with one memory pool. With Kavini being delayed it appears you can't get that with 28nm Bulk and there is no 28nm SOI as far as I can see; that leaves 22nm SOI. OR a 28nm SoC + 28nm descrete GPU with it's own memory which is more expensive and moving to a 22nm SOI SoC which could be made 6 months later which doesn't make sense economically. At the same performance SOI is cheaper than bulk.
McHuj
Member
I don't think any of the consoles will launch on a smaller process than 28nm. Next year 28nm will be very mature and will have good yields for mass production. 32nm wouldn't surprise me either.
Even if a smaller node is ready, it won't be ready for such a volume. A 22/20 node has to be ready for mass production around summer for a fall/winter launch, not at the end of the year.
I think both Sony and MS will want to get as many consoles out as they can when they launch. I don't think either wants a repeat of what happened with the 360 shortages when it launched.
Even if a smaller node is ready, it won't be ready for such a volume. A 22/20 node has to be ready for mass production around summer for a fall/winter launch, not at the end of the year.
I think both Sony and MS will want to get as many consoles out as they can when they launch. I don't think either wants a repeat of what happened with the 360 shortages when it launched.
systemfehler
Member
I thought that STMicroelectronics and Globalfoundries had some FD-SOI with 28nm deal going on but I am not sure how mature the process is so far. Roadmaps, universities, experimental projects, ... you can never be sure when that "tech" really hits the street.
No, it really wouldn't at all make a difference. 3-4 months wouldn't give them enough time for any changes worth mentioning.
You seriously just don't get it. They are not going to jump on new die sizes for a system where cost is likely already an issue. Your giving people false hope by even hinting they may go 20 or 22nm.
jeff_rigby
Banned
STMicroelectronics Secures Additional Sourcing for its Leading-Edge 28nm and 20nm FD-SOI Technology with GLOBALFOUNDRIES
AMD would not be designing a PS4 SoC starting in 2010 counting on ST Microelectronics to give GloFlo the technology to use FD-SOI just in time to produce the SoC.
Collaboration at 28nm, 20nm and 14nm: IBM, Cadence, ARM, GLOBALFOUNDRIES, Samsung
I think 28nm is not considered more than a stepping stone to 20nm where most of the effort seems to be targeted.
They may jump on 22nm SOI because cost is likely an issue, you just don't get it....the Microsoft and Sony game console design is a HIGH volume multi-year design and is unique with the Apple iPad second but the SoC for those change each model revision with a 20nm SoC coming in a year.KageMaru said:
It can be done for Apple @ 20 nm but not Microsoft or Sony? We are talking only a few months difference from guesses and leaks.http://www.techpowerup.com/173726/TSMC-Seen-As-Only-20-nm-ASIC-Supplier-to-Apple-in-the-Next-2-Years.html said:
It could be 32nm SOI or 22nm SOI but not likely 28nm SOI. It could be 28nm bulk in which case it might be SoC + Discrete GPU but it then could not be produced by TSMC. In fact 32nm and 22nm SOI would probably be only IBM and GloFlo too.
Point is that 22nm SOI appears to be the best choice and waiting till it's ready, worth it. This is supported by Sony stating they are skipping the 32nm node for cell and work on 22nm Cell registers already done (plans to sell the PS3 in China pretty much require a cheaper lower power design). It's supported by an IBM employee with 22nm experience moving to Global Foundries in 2010 to work on 22nm.
jeff_rigby
Banned
Yes I can be wrong but at least I'm boldly wrong with style. <grin> You might as well tack on my 1PPU4SPU theories again as I'm thinking I was wrong @ 32nm but might be correct at 22nm. I assumed the voluntary EPA Goldstar Power specs as well as a PS3 end of life about 2015 meant the refresh would occur June 2012 not 2014 (too close to end of life to be practical). With Sony planning to sell the PS3 in China it extends the PS3 life and makes waiting for 22nm as stated and shown in the Linkedin cite understandable, I should not have dismissed those two points.
NemesisPrime
Banned
It is going to be funny and ugly at the same time.
I cant help but smile in this thread.
I have yet to decide if Jeff is some kind of misunderstood genius or completely insane. I don't even wanne know what he does in real life. It scares me.
jeff_rigby
Banned
From BY3D
On Durango everything requires 1080P
Makes sense. Don't want to confuse the TV with having to resync the HDCP and frame buffers in the TV and Console with a resolution change. Bringing up a browser or any application with text is going to be difficult to read at 1080P much less some other resolution so everything should target 1080P. The TV display resolution can probably be scaled from 1080P to whatever the user chooses, 1080P being the best choice.
On Durango everything requires 1080P
Makes sense. Don't want to confuse the TV with having to resync the HDCP and frame buffers in the TV and Console with a resolution change. Bringing up a browser or any application with text is going to be difficult to read at 1080P much less some other resolution so everything should target 1080P. The TV display resolution can probably be scaled from 1080P to whatever the user chooses, 1080P being the best choice.
Judging by the responses he gets from other, more knowledgable, members here, I'm sure you can figure out which of those two he is. =p
It doesn't make sense and it's not like the resolution in these consoles would be changing on the fly. It'll likely be similar to how the 360 is now. Set it to a desired resolution and everything is either run natively or scaled to said resolution. I can browse the internet just fine at 720p on both the ps3 and 360, so again I'm not sure where you are getting any of this from.