Cell clock is already very high, it consumes a lot of power. How many multi-cores CPUs with 3+ Ghz core clock we have now on the market now? We are probably going to see more cores, not higher frequencies.
PS4 to be based on Cell?
- Thread starter tuna
- Start date
Why is it a con? Are they misrepresenting? People are buying what they're selling - what they're selling is an approachable 'pick-up-and-play' experience.
Yes, but releasing a low spec'd isn't the way to win that sector over. It needs something that they'll buy. Without a Wiimote surprise feature, what will be the selling point of PS3.5? (note this was assuming a 2x the performance console, which clearly isn't going to happen because an updated nVidia CPU would already be several times the capability!)
That's marketing. The Wii's biggest success was based on Nintendo's fun image and a spot-on marketing effort, along with an affordable price-point. The image of the new controller made a difference, the actual controller not so much.
Most of these Sony can replicate; the biggest challenge is trying to be 'fun'. Maybe they won't, and will instead go in a different direction, like, I don't know, 'sophistication'.
I think Sony have to address their SCE* dissonances first, though.
DuckThor Evil
Legend
It goes without saying that PS4 will be much more cost effective than PS3, but I sincerely hope that this 2x total system power talk is false information and I hope they can come up with something significantly better.
Sony is in a tough spot now against MS, because they don't have any chance to go into a price war against X360. I'm sure that's the one mistake they won't repeat.
The disapointing process shrinks of the Cell processor so far give a hint that maybe the Cell used in PS4 won't have much more cores, but I quess time will tell. Hopefully atleast one manufacturer will have a full generational leap in next gen, but MS will surely target the same launch window and they probably also want to have cost effective system, so what can they do?
Sony is in a tough spot now against MS, because they don't have any chance to go into a price war against X360. I'm sure that's the one mistake they won't repeat.
The disapointing process shrinks of the Cell processor so far give a hint that maybe the Cell used in PS4 won't have much more cores, but I quess time will tell. Hopefully atleast one manufacturer will have a full generational leap in next gen, but MS will surely target the same launch window and they probably also want to have cost effective system, so what can they do?
A lot of why the Cell die shrinks have been modest so far are because of the 'dumb shrinks' in conjunction with the poorer scaling of the I/O creating dead space on the chip. Any real re-design, likely to occur at 32nm anyway, should bring the chip forward to where the shrinks should traditionally have brought it.
Flash memory also runs cooler, takes less space, is less failure prone, and we only need a small amount of it standard to aid in game optimization, patches, system updates, etc. Hdd based mass storage is best left optional because it never cost scales down since its always chasing a moving target (bigger and bigger sizes needed over time). I recall a conversation many years ago at Microsoft where they lamented how the standard hdd was killing them cost wise on the original Xbox. Hence why an hdd-less 360 was no surprise when it came out. I'd wager both parties are considering the same next time out.
Its not as high as intially proposed, if Im not mistaken the Cell was originally spec'd at 4+ Ghz with 90nm prototypes running at those speeds. The 3.2 Ghz of the current Cell found in the PS3 is the result on having to lockdown the chip in its infancy at 90 nm. A very mature Cell at 45nm should scale over 3.2 very easily.
If all Sony plans to do is to "wii-tized" the PS4 (my comment was based within the context of the article) then simply increasing frequency of the Cell and pairing it up with a more robust GPU will provide more performance per dollar than adding more cores.
I don't think Sony will wii-tized the PS4 so I honestly expect both more cores and higher speeds.
I think by that time though having onboard storage for media/content through a home network or the Internet will be fundamental to the business models of both consoles. Personally I think that they're going to keep a mass storage option onboard, though it's just a random stab. Probably they'll just run the numbers on the advantages of the incremental revenues derived from inclusion vs the costs of inclusion averaged out over the ownership base - whichever makes the most sense they'll go for.
Agreed.Hope the rumour is not true. I want consoles running games at 1080p and even more next gen.
If true, they should Wiifi the PS3 by 2010. I mean, PS4 for power and a Wiified PS3 for mass market.
They just need to hit that price sweetspot [<400$] and to design their OS from grund up to support EVERYTHING that competition offers [we waited 2.5 years for in game XMB]. Any new hardware from 2011 will be more than enough to future proof ps4 untill next console version in distant 2016+
DuckThor Evil
Legend
But still even with a good 32nm shrink we are getting closer to the point where process shrinks are going to get harder and harder, atleast that's how I have understood the situation, so if the next gen Cell is a big chip at 32nm, I'm wondering how much smaller can it even get during its lifetime unless some major breakthroughs are achieved in processor manufacturing 22nm? 15nm?
I'm sure cost effectivenes is going to be a major element in Sony's strategy with PS4, so how much room is there for Sony to maneuver? Maybe the chip can't be that big in the begin with?
Last edited by a moderator:
So you're saying the Wii launched with conventional controller, no Wii Sports or Wii Play and the usual selection of Mario and Zelda, with same marketing (people looking happy playing Mario and Zelda) would have sold about as well as Wii has?
Come off it! Wii is a great hardware seller, but the actual money is on software as consumers appreciate more software variety. It makes a lot of sense with an unpirateable console like PS, though.Cheers
Goto in his article went the route of 45nm ~150mm^2. Now, Cell's already taped out on 45nm at ~115mm^2. I suppose you could do the new revision for a PS3.5 on 45nm, but... I don't know. I think it'll be on 32nm one way or the other, either IBM's new HKMG process, an SCE/Toshiba CMOS process as Pana suggested in another thread, or maybe those are one and the same thing come 32nm...
But I just think it makes sense to go into all the labor intensive re-working of this chip that is required on the I/O and process transition (no matter the process) at the same time, and 32nm would be perfect for 2011. I agree that it will be a modestly sized chip though, because yeah future shrinks (beyond 22nm on the HKMG) seem nebulous. I think between the die reclamation on the re-working wrt to the 'dead space' now, 32nm, and still modest die budgets, we could see a 'Cell evolved' concept at 32 SPUs and decent clocks under 200mm^2. If they went 16 SPUs (which they might), they could probably get it under 100mm^2.
They should make a newer version of the eye camera and have it included with every PS3 edit: I meant PS4.
Someone posted an old Sony patent 2 years ago that involved some gloves which enabled the user to hold virtual objects and move them. That looked awesome. I d love to see that take fruition.
I wouldnt say "appealing titles". I think its titles are just as appealing as those of the GC if not less. I feel that the appeal for these games comes from the controller alone.
Last edited by a moderator:
True it very much is...but I think even that is better served with optional storage, since people that really want to download movies, etc, will probably want a hard drive that is much bigger than what these machines ship with stock anyways. Plus, supporting external plug in drives (which so many people have) means 1tb+ of storage will be cheap and easy for anyone to add on.
That was probably just speculation on his part, and doesn't really relate to final performance anyways. Doubling cpu or gpu power sometimes opens up new techniques which can result in far more than a 2x perceived improvement. I guess what I mean to say is that how good something will ultimately look on screen is not necessarily linearly related to the quantified power increase of the hardware.
Not sure why people are so hung up on 1080p. Put it this way, hook up a modern pc via HDMI to a nice hdtv, and run the latest games at 720p on them with an Ati 4870 and 4xmsaa. You'd be amazed at how far better they look than any current 1080p console game.
Last edited by a moderator:
Not to mention that way, a fossil HDD from the Cretacic won't be mandatory for Sony and MS and they can defray the costs of such expensive -and not much useful in the future as it is old fashioned for a console with large flash drives- hardware.
It's getting late so I will be brief, if only this once.
Well, Grid is a good example of a 720p, 4xMSAA game and it looks great, not a jaggie in sight, most of the time. Your suggestion makes me reflect on the subject, and I wonder if I'm not becoming someone who values graphics above all, though I don't think so.
My 22" HDTV is 1680x1050 and we probably need SSAA, FSAA and MSAA -or other AA created by developers-
combined in the future, to get rid of any sort of jagginess, excuse me if I'm wrong.Love the 4870, despite not being a PC gamer I can appreciate it as it's a great piece of hardware (sometimes I silently voice my opinion in the 3D Architecture forums)
. Good evening....
Similar threads
- Locked
