Predict: The Next Generation Console Tech

Status
Not open for further replies.
If not for the whole DLC and digital download thing, it would actually be much, much more cheaper to have 8-16GB of cache ram instead of a traditional HDD or even a SSD, not to mention a between 10 to a 1000 times faster.

Could for example 64GB 'slow' (ie. really cheap) solid state memory for DLC and digital download games, coupled with 8GB cache ram be cheaper than a mechanical hard drive?
 
Could for example 64GB 'slow' (ie. really cheap) solid state memory for DLC and digital download games, coupled with 8GB cache ram be cheaper than a mechanical hard drive?

Looking at the cheapest possible present options, the raw flash for 64GB would cost ~$60, to which you have to add a controller, (idk? ~$5?), and the lines and space on the board. A hdd would still be cheaper.

However, 32GB would already be cheaper than a hdd, and if you wait just a single year, the 64GB of flash is likely to become cheaper (by a nice margin).

Also, there might not be need for that cache ram -- the price of "fast" mlc flash has more or less converged with "slow" mlc flash in the most traded chips, so the only thing determining the speed is the controller. There's no sensible reason you couldn't put 64GB of "faster-than-a-mechanical-drive" flash on a console shipping next year.

I think it's a safe bet that at least some consoles will ship with just flash as opposed to hard drives, as this will allow much cheaper minimum prices. There might be sense in allowing hdd upgrades for media center use and for downloaded games.
 
I expect every console to come equipped with at least a 250GB HD next gen. I'd be incredibly surprised if that wasn't the case. Downloadable titles is becoming more prevalent each year. Downloadable titles are only going to get bigger and more ambitious, and there will also be the option to download full retail titles. Can't imagine there will be consoles without adequate mass storage. It was different at the beginning of this gen when they really didn't know how big the sales of download titles would turn out, and the expectation was that download titles would be small in the hundreds of megabytes range.
 
Specs even less RAM are not what makes a system lasts. A system lasts because it is successful.
If either the PS3 or 360 have been a run away success (even without considering wii like level of success) I'm pretty sure that one of the two system may have lived a shorter life or have met with its successor already. My belief is that the main ingredient for success are brand strength, game catalog and price. That doesn't mean that I believe that a cripple system is not important but I believe that this take back seat vs the previous ingredients.

I also belief that launch period have a greater importance than the amount of ram the system embark. So for me the focus and so how money is spent should be not on lasting but making the most successful launch as possible, so quick results on the system and lowest price as possible while offering a clear upgrade.

Launch date will be important I believe that sooner would better than latter. If somebody actually manage to launch in the same time frame as Nintendo next system it will be a huge win for them. Actually we hear rumors about Ms but it would be fun if Sony managed to keep secrecy and surprise everybody by announcing a new system for fall 2012 (even if it's not a world wide launch).
Still I believe that MS would gain the most by managing to launch pretty early almost along with N next system. They're investing a lot of time negotiating with tv channels, possibly ISP and other on-line services. Kinect is a pretty successful device, if they offer a clear upgrade path for their consumers with a reasonably priced system they may manage a really successful launch. My belief is that they should do so soon, before Kinect and the 360 starts to really run out of steam. A launch by fall 2012 in US alone would have been their best, too it appears they are likely to miss this windows. For their sake Sony better not take out an ace from their sleeves and achieve what everybody deems unthinkable, launch before MS.

Regarding hardware and its price for me a Soc is a good news and no matter what most people say here I believe 2 GB with a 256 bits bus would be a great news. A HD6850 ships around 160$ a hHD6950 anywhere between 230 and 280$. The difference can be as low 70$ for a way bigger chip, more expansive cooling, twice as much RAM and the extra on the mobo + being a performance leading card by it-self induce a premium. As Blackowicz reminds us some pages ago some flash storage on board is not a magic bullet without a proper controller, caching may not be an option without significantly raising the cost of this investment. So a more reasonable bet is that flash will be used for OS, kinect, save and some downloadable content. In this context 16GB could very well be enough. In the same time HDD less system prove to set a lot of limits for games but also considering future usages of the system. altogether I might be reasonable to pass on flash and simply invest into the cheapest hdd available. Cheapest HDD available and providing enough capacity are 3.5" in form factor as sell as low as 40$. It might be one best bet. 2.5" are more expansive, less reliable and offer in the same price range a lot less storage.

Speaking of the SoC costs and future shrink, I wonder about the option of MS using a real Soc not an APU/fusion chip. Clearly an extra cost in R&D and an investment in die size but in the long run it could prove cost efficient. See the price (not the cost) of HD6950 with a proper SoC one would just have to add a HDD, some ports and minor chips (like wifi) to make a compliant system.
The win? really simple mobo design a chip, 8 memory modules, a cooler, some room and a connector for the HDD, an optical drive, the casing, external PSU. I can see MS launching such a system @399$ including kinect without loosing much money, as long a the system encounters no production problem and meet success the loss will be offset rapidly by games sales and live subscription.
Lean system may also be lean manufacturing, they buy the SoC both from IBM and GF (possibly putting them in competition), they won't run out of memory providers, neither out of HDD provider or optical driver. There are less limiting factors, with more chips you may have one that does well the other not that much etc. which prevent you to achieve your production goals.
Lean system (1 chip 1 memory pool) as well as swallowing the fixed cost of a HDD means that devs and editors will get a lot out the system really soon into the system life, which can be a bless for a successful launch.
Imho it's more the whole package than specs per self, if MS has the deal ready in regard to on-line service and content, get exclusive and third parties to really quick start the system due to extremely lean design, and they manage to launch as a conservative price, and launch asap, they can achieve something bigger than through better specs more RAM, etc.
Also I believe that they should not do a world wide launch if they risk to face production constrains better build a big user base as fast as possible in US, and delay EU launch by a quarter or two, Japan should come last / when there are no longer inventory constrain.
 
Last edited by a moderator:
As shrinks happen and it becomes viable it is entirely possible the Southbridge and other components are integrated into a single die. I think the GPU for the GameCube and Wii implemented the southbridge onto its die for example.

If not possible straight away then in a shrink or two I would not put it beyond the expertise of whoever it is that designs the southbridge. Who did the Xbox360 southbridge? Wasn't it SIS?

In any case - sounds like we have a ready system:

Option 1:

Multicore CPU
Dx11 equivalent GPU @ current levels of performance
256bit memory bus
2 - 8 GB of GDDR5
16GB of flash memory
Optical disk drive (Blu-ray)
HDD (different SKU's get different sizes)
Wifi
Ethernet
USB3

Option 2:

Multicore CPU
Dx11 equivalent GPU @ current levels of performance
eDRAM
128bit memory bus
2 - 8 GB of GDDR5
16GB of flash memory
Optical disk drive (Blu-ray)
HDD (different SKU's get different sizes)
Wifi
Ethernet
USB3

Option 3:

Multicore CPU
Dx11 equivalent GPU @ current levels of performance
128bit memory bus
2 - 8 GB of GDDR5
50GB of flash memory
Optical disk drive (Blu-ray)
HDD (different SKU's get different sizes)
Wifi
Ethernet
USB3

Option 4: (my favourite)

Multicore CPU with OpenCL GPU co-processor for physics
Dx11 equivalent GPU @ current levels of performance
128bit memory bus
2 - 8 GB of GDDR5
16GB of flash memory
Optical disk drive (Blu-ray)
Opional HDD (different SKU's get different sizes)
Wifi
Ethernet
USB3

Or Option 5 for PS4

Super CELL2 Multi PPC and SPU design with its own pool of XDR2/GDDR5 memory 1GB - 4GB
2GB GDDR5 GPU @ current levels of performance
128bit memory bus x 2
16GB of flash memory
Optical disk drive (Blu-ray)
Opional HDD (different SKU's get different sizes)
Wifi
Bluetooth
Thunderbolt
Ethernet
USB3

Bake in the over for 12 months.. voila. :)

Don't forget the dev tools, cooling, design, advertising and make it as cheap as possible without breaking it.
 
Well, my main beef here is the 8GB vs 4GB, which for Microsoft, would be akin to them including 1GB in the 360, which would have made the console quite a bit larger and more complicated with the tracing on the mainboard for years to come, not to mention even much more supply constrained as GDDR3 700MHz was the cutting edge in 2005 and one of the major reasons the console was supply limited in the first place.

I do realize that PS3 represented a 16x increase, but again, there's the physical I/O involved here as well. PS2 was smaller than Xbox.



-------

One thing I would like to point out is how the Xbox 360 dev kits still only used 512MB RAM for years because they used identical hardware. There was just no room for double the RAM chips. It wasn't until the 1Gbit GDDR3 chips were in mass production that they finally updated them.

Will devs need that doubling for dev kits next gen? Maybe not, but they'll be wanting to make use of the whole amount of memory without the headaches of fitting dev tools in memory simultaneously with the game content.

Doable... it's just there are a number of factors involved.
Nobody said Sony, Microsoft, etc, should be over ambitious. But if you realize what happened to Wii, the catastrophe this console is, you can understand that good software sometimes requires fine hardware.

Take into account how much time the Wii lasted, not even 4 years and a half since it has been launched, they announced a new console. The Wii U launch is nearing..( and I feel horrible for Nintendo and the mistakes they've made ).

Having a Wii right now is having a beautiful paperweight at home. It does nothing but playing games with outdated technology, visuals and possibilities. No online updates in most cases, limited support for storage.., and the list goes on and on. :oops:

Their hardware strategy has been a success in the beginning. The Wii has been prohibitively expensive, taking into account what its hardware offered, since the beginning, making a HUGE profit for every sold unit, but this strategy backfired on them.

Now Nintendo find themselves in a complicated situation, launching two consoles in an interval of a year and a half, having to deal with the colossal and immense investment the launch of a new console requires, and an uncertain future.

If this wasn't enought, there are rumours that the Wii U is underpowered already...

PS3 and 360 launched with a decent amount of RAM for the time, and are probably the most ambitious consoles ever made -more than the Neo Geo, for instance-, :smile: and the results are there. I mean that, while they show signs of aging, we can play these consoles until 2014 without being in a boring perpetual state of feeling that we need a new generation the next week or so.

4 years has been the age of death for the Wii.

Nintendo accomplished a lot, selling hardware at a huge profit, bringing new people to gaming, and it's funny, and great. But it feels like an entry level game console which wasn't ambitious enough and it paid the consequences.

Ageing so soon for a console is a bad sign really. For people, well, your looks go downhill. Your figure. Your youth. And you can't go into clubs full of early 20 somethings without feeling icky. But when it's about hardware, innovation, processing power, etc, makes time to be on your side, and helps you to keep abreast of the times. :smile:

Wii missed a lot of software because developers can't squeeze anything more out of the system, I think, and Nintendo had to reassess the situation of the console. It will be interesting to know the success of the new Wii U but I believe they will fare as well as the Wii -I expect 100% backwards compatibility- and they will have all developers and traditional gamers on board too.

I don't see adding at least 4GB of RAM as any sort of problem, so see no reason to even contemplate changing the natural progress of hardware it if I were given the option.
 
Indeed. The Wii has been a catastrophic failure for Nintendo.
Indeed the Wii is still selling at a pretty profitable price :LOL:

If anything Nintendo failed to transition from the Wii to something else, the same is happening with the DS and the 3DS. Clearly spec have nothing to do with what is happening to Nintendo right now, Exec and share holders are responsible not specs.

EDIT
I take that your post was a sarcasm so I agree with it. I'm not sure it was clear.
 
Last edited by a moderator:
Yeah I don't get it either. How much more money has Nintendo made with the Wii, ten times or more than what Sony or MS has? Then again they both haven't really recouped their losses yet, so I think the Wii has actually made infinitely more money compared to the HD consoles.
 
If you include all ram (and not just main ram):
- PS1 3.5 MB
- PS2 40 MB
- PS3 512 MB

PS1 -> PS2 increase of 11.4x
PS2 -> PS3 increase of 12.8x

I know this means nothing for next generation of course; Sony will use whatever amount makes sense and not use a formula based on PS1 to PS3.
 
Indeed the Wii is still selling at a pretty profitable price :LOL:

If anything Nintendo failed to transition from the Wii to something else, the same is happening with the DS and the 3DS. Clearly spec have nothing to do with what is happening to Nintendo right now, Exec and share holders are responsible not specs.
The main problem with Wii is that it was interesting, seemingly fun and really cheap. That meant tons of people bought it but after you realized how little it really offered it just staid in a corner collecting dust. I wouldn't be surprised if software attachment rate for it was several times lower than either PS3 or XB360.
 
Looking at the cheapest possible present options, the raw flash for 64GB would cost ~$60, to which you have to add a controller, (idk? ~$5?), and the lines and space on the board. A hdd would still be cheaper.

However, 32GB would already be cheaper than a hdd, and if you wait just a single year, the 64GB of flash is likely to become cheaper (by a nice margin).

Also, there might not be need for that cache ram -- the price of "fast" mlc flash has more or less converged with "slow" mlc flash in the most traded chips, so the only thing determining the speed is the controller. There's no sensible reason you couldn't put 64GB of "faster-than-a-mechanical-drive" flash on a console shipping next year.

I think it's a safe bet that at least some consoles will ship with just flash as opposed to hard drives, as this will allow much cheaper minimum prices. There might be sense in allowing hdd upgrades for media center use and for downloaded games.


Can someone exlain the difference between flash and and an SSD to me? I dont quite get why SSD are so expensive per GB, but flash isn't. Especially because flash is apparently fine for consoles aka the 4GB 360.

Is it speed and durability? How much faster is an SSD than typical 4GB 360 flash? Which is faster, 4GB 360 style flash or a mechanical HDD and by how much?
 
If you include all ram (and not just main ram):
- PS1 3.5 MB
- PS2 40 MB
- PS3 512 MB

PS1 -> PS2 increase of 11.4x
PS2 -> PS3 increase of 12.8x

I know this means nothing for next generation of course; Sony will use whatever amount makes sense and not use a formula based on PS1 to PS3.

Yes, but you could also take time into account. Ps2>ps3 6 years. Xbox>360 4 years. 360>nextbox=7? 8? PS3>PS4=?

You could argue we should see a bigger ram jump this time, 360 will be looking at at least 7 year gap, probably 8 or more.

Yeah I don't get it either. How much more money has Nintendo made with the Wii, ten times or more than what Sony or MS has? Then again they both haven't really recouped their losses yet, so I think the Wii has actually made infinitely more money compared to the HD consoles.

I would guess 360 is already profitable as a project for a while. This is based on if you add up all the division P&L, they were getting pretty close (think it was ~-$600 million? but this was 2-3 Q's ago I havent kept up). But of course thats the raw numbers, and I recall MS saying in some quarterly conference call that most of their EDD profits, not revenue, were based on 360. I wish I had bookmarked that. Basically they were confirming that 360 alone was more profitable than the division it was in. So, if the division is at ~-600m, my guess is the 360 alone is already profitable.

Would be nice for some MS guy to drop that in interview one of these days to confirm. Something like "well you know 360's been profitable for us this generation blah blah"
 
Last edited by a moderator:
Can someone exlain the difference between flash and and an SSD to me? I dont quite get why SSD are so expensive per GB, but flash isn't. Especially because flash is apparently fine for consoles aka the 4GB 360.

Is it speed and durability? How much faster is an SSD than typical 4GB 360 flash? Which is faster, 4GB 360 style flash or a mechanical HDD and by how much?

Prices are comparable. The cheapest SSD I can get here in Denmark is the OCZ Vertex+ 120GB with a per GB price of $1.13 (ex VAT). Mainstream price for SSD is $1.5 / GB, spot price for flash is $1 / GB

You pay extra for the controller, PCB. The SSD vendor also has to turn a profit.

Cheers
 
The PS1 had 3.5 MB, the Saturn had 4 MB - that Saturn lasted 5 years and the PS1 lasted 10. The PS2 had 40 MB, the Xbox had 64 (+1 scratchpad on the HDD) - the Xbox lasted 5 years and the PS2 is 11 and still going!

The relationship between longevity and specification seems ... complicated.

Some of these RAM counts where you straight add in the EDRAM seem disingenuous...it's not like the 360's 10MB of EDRAM effectively gives it 522 MB of RAM, from what I can gather, that memory is special purpose and doesn't add to main RAM. It's not used in the same way. Likewise I dont consider PS2 had 40MB of RAM, but 32MB. Seems you added 2MB of "I/O memory", 2 MB sound memory, 4MB of EDRAM, etc.

The same as people say the GC had 40MB of RAM (you would say 43) but 16MB of that was effectively useless. 81 mhz on an 8 bit (!) bus. In reality GC had 24MB and PS2 32 MB, Xbox 64 MB.
 
The main problem with Wii is that it was interesting, seemingly fun and really cheap. That meant tons of people bought it but after you realized how little it really offered it just staid in a corner collecting dust. I wouldn't be surprised if software attachment rate for it was several times lower than either PS3 or XB360.

Think again, Nintendo sold a LOT of software. About as many units as the other two combined, at least, and most of the big sellers were first party, with pretty low budgets compared to the top selling HD franchises.
 
I would guess 360 is already profitable as a project for a while.

Might be profitable per quarter, but there's a multi-billion dollar deep hole it has to climb out, of with those few hundred millions per quarter it might not even happen within the system's life cycle. Then there are the Xbox1 losses to recover, too.

Nintendo on the other hand has been profitable all the time. Maybe not that much with the GC, but they made no losses at all, only profits...
 
Status
Not open for further replies.
Back
Top