Predict: The Next Generation Console Tech

Status
Not open for further replies.
from september 2010?
fake then
19-97027c6d13.jpg

nukezilla should remember change ATI logo to AMD logo next time
 
from september 2010?
fake then
19-97027c6d13.jpg

nukezilla should remember change ATI logo to AMD logo next time

Nope. Try again. That's how the actual demo shows it (~ 1:20 mark)

 
Nope. Try again. That's how the actual demo shows it.


a 2008 video?well,more thanks to prove it for fake

and can't believe some people think nukezilla thing is real
oh and the upload date is almost same as nukezilla "leaked news" day
 
Last edited by a moderator:
Then you're seriously, SERIOUSLY above the norm, even for industrialized western countries.

Building truly fast broadband connections throughout entire nations (IE, fiberoptic links) will take at least a decade.

Here in Finland, there is a serious push for fiber-to-the-home. Basically all new properties get fibre links pretty much by default, and where they don't already exist, new ones are getting built up as fast as they can be. My dad who lives "out there by the sticks" just got an offer for fibre broadband, and there are less than 50k people living within 50km of him. The goal set by the state is 100Mbit for everyone by 2015, but we aren't quite going to make that. The long-term plan is to replace the copper networks with fibre everywhere with a high population density, and push 4G and successors everywhere that doesn't. At present speed, it will be done at some point in 2017-2018. Similar development programs exist in most western countries. Fiber to the home will be reality for the majority of city-dwellers before the end of this decade. (The exception being the Americans -- apparently it's not okay for the government to push for infrastructure development there... :p)

An interesting added incentive is that the market price for copper is now so high, that collecting old copper bundles and replacing them with fiber can actually *make money* for the operator, so long as it's air cables with minimal work required.

Ahaha... Ok, and while we're dreaming let's make that gigabit connection cost no more than $10 a month too so that most people can afford it, and cover the entire world...

After it's built, there's really no fundamental cost difference between a 10Mbit fibre link and a 1Gbit fibre link to the operator. Actual connection speeds far over the internet will of course be less. (And will likely be limited somehow to reduce costs to the ISPs). But that's why CDNs exist -- my roundtrip ping to youtube is 13ms. Heavy content is moved near to the customer, so the ISP doesn't have to pay for access to it. And MS has one of the most extensive CDNs in existence.

Right now, I can get a 10Mbit speed for 10€/month, 50Mbit/s for 20€/month, 100Mbit for 30€/month, 250Mbit for 60€/month, or the 1Gbit for 90€/month. I expect the prices for the upper tiers to fall at least 20% yearly.

I'm not saying that on-demand-only will be realistic for everyone within the current gen. However, there will be tens of millions (and hundreds of millions at the end of the generation) of potential customers who have really fast connections, and would be better served by an on-demand console, because getting the data from the CDN will simply be faster than getting it from the drive.
 
You would need a 500 megabit line to beat that.

If most people have that in the next five years i'll eat my hat, jacket and router:)

Most city-dwellers won't have that. They'll have a Gbit line, because it's a nice round number that's easy to sell, and there is no significant cost difference to the operator.
 
Do you really think that is a fake?
That's must be the best fake in the entire history. It' really well done and extremely long. And they show SmartGlass at page 20.. and the document was uploaded before the E3.

I think it's a really old marketing/technologic research, but it think that's true. Some strange rumors now fit nicely.. like the dual-GPU ones.

My interpretation:
Yukon is the architecture: one core-SoC, one application "module" for next-gen gaming+backward compatibility.
Oban is the name of the core-SoC.
Durango is the final project name of Xbox 720.


The core-SoC won't be used by games, but by the OS. It's probably a Steamroller module.
The application "module" is made out of 16 Jaguar-core, and CapeVerde+eDRAM (48-64mb?), plus the 3PPC cores. Maybe they are divided in 2 chips (CPUs + GPU/eDRAM).
The entire thing has 8 Gb of DDR4 (that require less power than DDR3 at the same freq).

Since the presentation was made in Sept 2010, they didn't know what kind of architecture AMD was going to have in the future. If they used the R7x one, the 64ALUs fit well with an APU-like solution (LLano/Trinity).
 
Last edited by a moderator:
Most city-dwellers won't have that. They'll have a Gbit line, because it's a nice round number that's easy to sell, and there is no significant cost difference to the operator.

It depends. Though areas that have above ground utilities will likely convert reasonably quick as the overall cost of FTTH is fairly low. The issue really comes about in those areas with underground utilities, as the cost to run fiber line underground is significantly higher. American cities seem to have a lot more underground utilities than most of the rest of the developed world. It is part of the reason where I live (SF) with a relatively high population density and a very high per capita income and per capita disposable income is still relatively stuck in the dark ages internet wise while my parents in Chattanooga, TN which at best has a light dense suburban/rural population and both much less per capita income, lower disposable income, AND less inclination towards high speed internet has FTTH across a multi-county several hundred square mile footprint and 1+ Gb/s(fully symmetrical) capability to every dwelling.
 
Yea, and its most likely fake :smile: Can't believe someone would go through all this trouble for this.

I have no no big opinions on the content, but I think it looks very much like an internal MS presentation, from words and idioms used to visual style. Then again, there are probably at least hundreds of thousands of people out there who can make that distinction, and who could make it to look like a real one if they wanted.

seems the faker wasn't technically inclined...

The one who put that together no doubt wasn't technically inclined. Doesn't mean he's a faker. That kind of presentations are not done by the engineers. They employ the engineers. For the one who made it, the tech spec slide was the least important and interesting in the deck.
 
Last edited by a moderator:
Well I have always paid for the fastest line I can have. I was one of the first 4k people to have adsl in the uk at home after winning a competition. That was in Docklands London.

I now live in a 4 year old housing development 25km from the center of Madrid. Fastest I can get is 10 Meg, I may be able to get a 100 Meg fibre connection in the next year but that would cost close to 80euros a month. How many takers do you think there will be for that in Spain in the current economic climate?

I would be very surprised if most houses have 100 Mega links in the cities here in the next 5 years. Whatever you might say Madrid isn't not some technology backwater.

My parents and brother's houses in rural England (glos) have less than 1 meg and it will be years and years before they get anything better.
 
you Europeans crack me up

Well yeah, what high and low population density means depends on your point of view. :)

The issue really comes about in those areas with underground utilities, as the cost to run fiber line underground is significantly higher.
Yeah. It also depends on how old the underground lines are -- most reasonably modern ones have cable trenches that you can draw the fiber bundle trough, without having to dig anything. The places where the cables went underground earlier are pretty much SOL.

American cities seem to have a lot more underground utilities than most of the rest of the developed world.
And a lot of those underground lines are very old, and cannot be reasonably upgraded without a whole lot of digging. To be fair, you guys had cable TV back when most of Finland was first getting fridges, and being the first mover has disadvantages.
 
Whatever happened to the fiber through the sewer concept ... does it have some unexpected complications?

PS. other than everyone wanting a piece of the pie and driving up costs unnecessarily of course, but there is where (local) government should step in.
 
Whatever happened to the fiber through the sewer concept ... does it have some unexpected complications?

PS. other than everyone wanting a piece of the pie and driving up costs unnecessarily of course, but there is where (local) government should step in.

getting slightly off topic, but one of the major issues at least in the US is private companies (TWC, Comcast, ATT, et al) actively preventing it through lawsuits and lobbing.

The other issue with sewers is physical security of the lines vs pests and debris.
 
It has flash storage, pcie, 3 CPUs and 2'GPUs, flash cache, EDRam, and they expect $299 with margins? I'm trying to figure out why pcie would even be listed... This is closed system... No need for highly bandwidth limited standards designed for modularity within it

Uhhhhhhhhhhhhhh

Scalability, OEM, cellular, glasses?

This is so fake

This is Microsoft, not Nintendo. They have never sold a console at launch for profit. That's what license fees on games are for.
 
2- "DDR4", why microsoft engineers would choose a non existent technology (that could never see the day of light in 2013) over a proven successful one a la GDDR ? it is like saying sony would use xdr2 ram for its ps4 : commercially non sense.

The consoles are designed with a 5-10 year road map. Even if DDR4 is in tight supply in late 2013 it will be *the* standard the majority of it's lifetime. DDR3 of equal speeds is hard to come by and has power issues and, very importantly, EOL in terms of technology. As DDR4 ramps up over the next couple years DDR4 prices will drop and availability rise while DDR3 will be on the reverse coarse. This is very similar to MS's selection of GDDR3 700MHz for the Xbox 360.

I would love to hear your sales pitch on GDDR. It is more expensive and has some performance and power trade offs. If MS is looking to spool off high bandwidth clients (eDRAM, a pool of stacked memory, etc) the benefits of GDDR may not be as appealing and the downside, cost and power as well as long term AFFORDABLE availability, could be big negatives.

4- process : 22 nm, really ? :rolleyes:

If this was made in 2010 then 22nm on the roadmap for a quick shrink in 2014 was a totally reasonable game plan.

1- specifying the exact frequency of the GPU (and the CPU) this early (2010), and saying ">32 mb edram", is just non coherent. it is more difficult to decide early on the frequency of the GPU than on the quantity of edram. The guy faking the document obviously wanted to play it safe.

iirc the Xbox 360 block diagram used similar nomenclature, e.g. "256MB+ RAM". Especially if the author/audience is not a technical design group (the document looks like a big picture product plan) giving general data like frequencies, performance range, memory ball parks, etc are all reasonable -- especially if things are in flux and other things nailed down. They may have already spec'd a CPU but markets and contracts are dictating memory, etc. And of course 3+ years out everything is subject to change. Expecting such a document to "qualify" comment is ridiculous.
 
Status
Not open for further replies.
Back
Top