Predict: The Next Generation Console Tech

[Squilliam]

This could be good for next gen consoles, esp Xbox with Kinect!

Arstechnica: Thunderbolt smokes USB

Thunderbolt can use either copper or fiber connections for 10Gbps bidirectional communication. That speed is 20 times faster than the theoretical limit of USB 2.0, 12 times faster than FireWire 800, and twice as fast as USB3. According to Intel, however, the 10Gbps isn't just a theoretical peak speed, but usable bandwidth. This allows a single port to communicate with multiple devices simultaneously for a combined throughput of 10Gbps.

...Intel said that adapters can be made using a Thunderbolt controller and common PCI bridges to adapt existing FireWire, USB, eSATA, and even Ethernet connectors...

...Connected devices can be clock-synchronized to within 8 nanoseconds.

This is plenty, right? Also apparantly it can carry 10W of power AND it covers multiple protocols for both display port and PCI-E with much lower latency and CPU useage compared to USB.

So potentially a console can use one of these ports in the back to support video, peripherals, ethernet and audio. One port to delete about 4-5 ports?! Not bad! They could probably even use the standard internally and have a second Thunderbolt port to cover any applicable optical drives, flash, sata HDDs and USB ports.

This sounds like a really really good thing to simplify the base console hardware. It also gives a lot more opportunities for expansion than USB 3.0 AND it also means that supporting cartridges, SSDs/flash or whatever shouldn't be a problem from an I/O perspective given the availability of bandwidth.

 

[Shifty Geezer]

So potentially a console can use one of these ports in the back to support video, peripherals, ethernet and audio. One port to delete about 4-5 ports?!

Not sure that's a good idea, daisy chaining. And even if that work well in practice, having to follow a chain of devices to see who's connected to who, if you only supported this in your console, you wouldn't be able to sell to people still using USB and HDMI techs, meaning pretty much everyone! So this sounds like an awesome connection tech, but it can't be a simplification for next gen. It'd have to be in addition to, and extra cost rather than a cost reduction. Although as I understand it it's a single controller that supports all the IO protocols, which at least keeps that simple.

 

[Squilliam]

Not sure that's a good idea, daisy chaining. And even if that work well in practice, having to follow a chain of devices to see who's connected to who, if you only supported this in your console, you wouldn't be able to sell to people still using USB and HDMI techs, meaning pretty much everyone! So this sounds like an awesome connection tech, but it can't be a simplification for next gen. It'd have to be in addition to, and extra cost rather than a cost reduction. Although as I understand it it's a single controller that supports all the IO protocols, which at least keeps that simple.

I was thinking something like:

Lightpeak connected to -> Various HDMI / USB / Audio ports / Next gen camera interface. That way they can shift the board complexity for supporting these various standards from the motherboard to an external adapter and it would give them the bandwidth to support high speed media such as cartridges etc should they so desire to as well as all the throughput they need for an HD camera. Over and above that I believe it would also give a next generation console flexibility and the ability to better adapt to changing conditions over the course of the generation, unlike unfortunately something like Kinect which couldn't get the bandwidth nor power it needed directly from a launch console.

 

[fehu]

so i buy a 200€ console and then go to search for an adapter just to connect it to the tv?
and if i'm really unlucky another to connect my usb pendrive?

and btw... how much complex must be the board layout to converge all the internal output such as video, usb, net etc to the lightpeak chipset, and then go to the real ports/adapters?
and isn't more expensive to put all this stuff LP+adapters compared to simply integrate the vanilla controller?

 

[-tkf-]

Thunderbolt requires a chip in both ends, which have to be bought, add to that the chip with whatever function it is you need and you add even more cost.

Besides that it´s an extender to PCIe and a Displayport.

So you need both PCIe and or a Displayport capable device onboard.

If you already have the Display device then you really aren´t saving much on that part.

That leaves the PCIe extender function and bandwidth to burn, what do we need that for?

It´s perfect for laptops and to a certain degree for stationary PC´s

 

[eastmen]

Except they won't be competing against the 360 or ps3.....

...anyone with a 360 or ps3 has no motivation to buy the console, genius. The 1st party titles aren't the crutch they use to be.

There are 80million wii owners who are going to be looking to upgrade. Some of them may own an xbox 360 or a ps3 but many of them wont. So nintendo being able to get ports of the current big named games esp if they are enhanced in some ways it will help them sell.

There are people like some of my friends who have had the xbox 360 since its launch in 2005 and are looking for a better experiance. Smoothing out frame rates offering a higher resolution and better textures in a cheap small console might be what they are looking for. Many of them will buy a nintendo console just for mario , zelda , metriod and the other nintendo games so the early ports will be a bonus .

http://www.youtube.com/watch?v=mdPi4GPEI74&feature=player_embedded

I'm not saying llano is the only choice but it is a choice and it might be a smart move for nintendo to look at if they want a small console with little power usage that can get ports of current gen games and future games.

Llano would put nintendo in a much better postion than what the wii was able to do inregard to future competition.

obviously if nintendo is launching in 2012 or 2013 they would go with something diffrent. AMD would be off 32nm and most other companys will be off 32/28 nm and most likely on 22nm perhaps a bulldozer with a 7x00 chip intergrated into it might be smarter . Or mabye they would stay with a quad core llano and have a diffrent intergrated gpu inside of it .

 

[Prophecy2k]

They will if they launch first.

It's not just as simple as launching first. By that reckoning alone it should be the 360 that has sold 80 million by now and should be in first place. Ostensibly that's not the case. Wii captured the casual gaming audience that the PS2 had, the same that were alienated by the very "hardcore" premise of the xbox360 and PS3. The Wii also expanded that audience by capturing a section of non-traditional ganers and non-gamers, an audience who had no intention of ever buying into the Xbox or PS3.

Tell that to the 80 million Wii owners. They certainly didn't buy for the performance advantages.

Again, Nintendo captured a completely different audience. One that shuns non-Nintendo traditional core games (a notion supported by software sales) and laps up stuff like Wii-Fit and Just Dance by the container-loads.

So Nintendo weren't even in direct competition with the Xbox360 and PS3, and they haven't been until now that those two companies have released their own motion peripherals.

There's also a point to be made that whether Nintendo launches first with a new console or not, it's not guaranteed that they'll be able to sway PS3 & Xbox 360 gamers away because of the existence of the online platforms on those consoles and the lack of such for Ninendo. They had nothing in the way of online integration this gen, and even if they pull something out with their next console they'll be trying to persuade the current hardcore userbase to abandon the services they're already acustomed, and all the content they've already invested in to for an almost invariably immature service.

Not an easy problem to overcome for Nintendo at this point.

I think Nintendo should continue targeting the Core and Casuals with the 3DS and focus on their current demographic with their next home console. After all they of all other game publishers are probably the only publisher that has been able to successfully figure out how to make software that appeals to that userbase. All the 3rd parties tried and failed and then gave up with the Wii, and even the ones that found some success with individual titles like Ubisoft and Just Dance, they'd thrown so much faeces at the wall that something was bound to stick eventually.

 

[liolio]

Llano doesn't looks like a competitive product imho, I anticipate poor CPU and GPU performances either in perf per Watts or mm².
Overclocked whatever AMD pushes as Bobcat.V2 would be a better match for usual Nintendo philosophy in regard to hardware.

 

[fehu]

performance / watt / die size for the comparison between espected bobcat2 and quad A9 or dual A15 with sgx6?

 

[eastmen]

Llano doesn't looks like a competitive product imho, I anticipate poor CPU and GPU performances either in perf per Watts or mm².
Overclocked whatever AMD pushes as Bobcat.V2 would be a better match for usual Nintendo philosophy in regard to hardware.

looks pretty good from what i see , 50watts when running multiple applications including the new FF online .

Remember Llano is 50watts on 32nm another micron drop and it will be down quite alot in power.

 

[Mobius1aic]

Llano will be awesome for the low end desktop and mobile market as of course it gives people a number of decent capabilities for real cheap with a relatively tiny TDP. While, yes it could provide some decent gaming on the small budget, it is severely under equipped bandwidth wise for a console. Unless you get the bandwidth up to 50 GB/s minimum, I would consider the chip a waste of good silicon if you're not going to give it the communication speed it needs. 128 bit bus + GDDR5 or XDR @ 1000 MHz+ pretty much would be the best solution to the issue unless you're willing to kick up the memory speed to more than 1200 at the cost of more energy consumption or perhaps willing to move to XDR2 (at higher cost of course).

 

[eastmen]

Llano will be awesome for the low end desktop and mobile market as of course it gives people a number of decent capabilities for real cheap with a relatively tiny TDP. While, yes it could provide some decent gaming on the small budget, it is severely under equipped bandwidth wise for a console. Unless you get the bandwidth up to 50 GB/s minimum, I would consider the chip a waste of good silicon if you're not going to give it the communication speed it needs. 128 bit bus + GDDR5 or XDR @ 1000 MHz+ pretty much would be the best solution to the issue unless you're willing to kick up the memory speed to more than 1200 at the cost of more energy consumption or perhaps willing to move to XDR2 (at higher cost of course).

I'm comparing Llano to the wii and to xbox 360. Compared to those its very capable and very cheap and would allow nintendo to go the wii route again.

Of course if you wanted to build a power house consle that blows Llano out of the park you could. But it be much more expensive.

 

[corduroygt]

I'm comparing Llano to the wii and to xbox 360. Compared to those its very capable and very cheap and would allow nintendo to go the wii route again.

So you got any pricing info on Llano since you insist it's cheap?

 

[eastmen]

So you got any pricing info on Llano since you insist it's cheap?

Just going by the markets amd wants to enter with it .

 

[corduroygt]

Just going by the markets amd wants to enter with it .

If everyone's going to be clamoring for these Llano laptops as you say, why should AMD sell it at very little profit to a console maker?

 

[AlphaWolf]

If everyone's going to be clamoring for these Llano laptops as you say, why should AMD sell it at very little profit to a console maker?

because its 5-10 years of guaranteed sales regardless of incoming competition?

and who said very little profit?

 

[eastmen]

If everyone's going to be clamoring for these Llano laptops as you say, why should AMD sell it at very little profit to a console maker?

why is it allwys the same shit with you ?


80 million consoles or so over 5 years is alot of chips and a great amount of volume that amd wont be able to duplicate in the pc market.

Llano with 4 cores and a 6620 wont be a compelling product for very long in the pc market , perhaps 8-12 months before its replaced by a Llano with more cores or a next gen chip.

Getting into the nintendo 6 or whatever will allow amd to produce Llano with 4 cores and a 6620 for years to come in the tens of millions . Two things will happen. 1) Nintendo wont want die shrinks and the product will remain on 32nm for years to come which will become an unused process for ms with tons of capacity to spare. or 2) nintendo will want die shrinks where as each time it costs amd less and takes less capcity from other products to make a chip for nintendo.

Even a fixed profit per chip is good for AMD $20 per chip over 80m consoles will bring in 1.6B or so which last i check is the settlement money intel gave them. Not bad for producing a chip that they already designed and sunk R&D research into .


Why would an ARM company sell chips to nintendo when they could be selling them in tablets and phones ?


The answer for all these questions is simply because they make money !

Of course Llano is a good processor for a 2011 console. I'm sure next year nintendo will want ot look at something more powerful but i'm sure another fusion part would make sense in the future if nintendo once again wants to go the wii route with a console that is just slightly above the last gen consoles in graphics. It offers a single chip solution , acess to tons of tools and complete engines and will have been made a few million times before nintendo starts using it thus lowering prices and increasing yields.

 

[MfA]

My guess, probably ~125$.

 

[eastmen]

My guess, probably ~125$.

You really think its going to be almost 4 times more expensive than brazos ?

 

[corduroygt]

Why would an ARM company sell chips to nintendo when they could be selling them in tablets and phones ?

You don't have to buy chips from a company that makes ARM, you can just license it yourself. You cannot do the same with an x86 chip unless Intel agrees. Besides, a quad core x86 chip with IGP would cost at least $100, while those ARM chips in tablets are only a few bucks a piece. An order of magnitude difference there.