Predict: The Next Generation Console Tech

[Squilliam]

Interesting thought but could play havok if there was a lazy dev that just relied on the fix speed of the console to do their timing for them. Not that I imagine any dev in this day and age would do something like that.

Added to that, it's not like that would really benefit MS or gamer's really. It isn't like it would be cheaper for MS. And it isn't as if dev's would be able to take advantage of any increase in capability as they still have to consider existing consoles.

I really wouldn't expect MS to do any changes with the X360 that weren't related solely to the UI, reliability or cost of the machine.

As to predicting the future. I'm currently not expecting anything new from anyone except possibly Nintendo before 2012 at the very very earliest. And most probably not for a few years after that.

And thus the only thing I can think of would be more programmable general purpose GPUs that are more like Larrabe.

Possibly with Sony/MS both looking at a CPU/GPU integrated on the same chip.

Regards,
SB

With the more complicated hardware, I don't think developers could code based on strict timings anyway. They would have to code to a best and worst case scenario between cache misses, memory vaguaries and multi-threading. With increasing the hardware performance a little, Microsoft may have to give their console a small clock bump anyway to control leakage (This was explained to me, so im not the source of this) and in the end all I was thinking was making a refresh which always gives the software running a best case scenario level of performance to keep games running a little smoother.

Beyond that, with a slim console one of my most wanted features would be the ability to form an ad-hoc local lan connection between two or more wireless equipped consoles so having a few more processor cycles could help here as well. Furthermore as the consoles are the local hosts in games it would help to improve performance for other consoles connected to the local host.

As for new consoles, between Microsoft and Sony we're seeing a proportion of the console market being locked into a particular brand of console by nothing more than peer pressure and staying with their friends. As the networks are not interoperatable it seems to me that the incentive is to release earlier to lock in people to a certain console and reap the network effects with people wanting/needing to buy a console that their friends own. In this case it matters even more than ever what consoles your friends have so an early start and lead becomes even more valuable. This is why I wonder if a late 2010 early 2011 start by Microsoft should be expected?

 

[ban25]

As for new consoles, between Microsoft and Sony we're seeing a proportion of the console market being locked into a particular brand of console by nothing more than peer pressure and staying with their friends. As the networks are not interoperatable it seems to me that the incentive is to release earlier to lock in people to a certain console and reap the network effects with people wanting/needing to buy a console that their friends own. In this case it matters even more than ever what consoles your friends have so an early start and lead becomes even more valuable. This is why I wonder if a late 2010 early 2011 start by Microsoft should be expected?

Microsoft already has the network with Xbox Live. Sony's mistake was shipping without something equivalent, let alone a year later.

 

[TheChefO]

As for new consoles, between Microsoft and Sony we're seeing a proportion of the console market being locked into a particular brand of console by nothing more than peer pressure and staying with their friends. As the networks are not interoperatable it seems to me that the incentive is to release earlier to lock in people to a certain console and reap the network effects with people wanting/needing to buy a console that their friends own. In this case it matters even more than ever what consoles your friends have so an early start and lead becomes even more valuable. This is why I wonder if a late 2010 early 2011 start by Microsoft should be expected?

Very good point.

Not only for initial purchase of a console, but also for future software purchases. The people that I know that own both a ps3 and xb360 all buy games for the xb360 for the online component to play together or for their gamerscore in single player games...

MS needs to keep their release close to the competition if they expect to keep the lead ... and rolling over the gamerscore will help.

 

[Squilliam]

Very good point.

Not only for initial purchase of a console, but also for future software purchases. The people that I know that own both a ps3 and xb360 all buy games for the xb360 for the online component to play together or for their gamerscore in single player games...

MS needs to keep their release close to the competition if they expect to keep the lead ... and rolling over the gamerscore will help.

Releasing early, so long as they get the games ported may be an incredible boon for them simply because if you take current generation graphics levels, but add say 2-4 player coop in the next GTA game on the local console at an increased framerate its something that would get a whole bunch of people together playing the new console, its simply free marketing if they can nail it right.

But yeah, rolling over gamerscore and giving an incentive to current Gold members to upgrade could get them running out the gate really quickly unlike the slow start of this generation.

I kind of wonder if it would make perfect sense to release in early 2011 if they can get say GTA IV xxx city as a near launch game for example. Or something like World of Warcraft and lock in that massive userbase quickly.

 

[TheChefO]

I kind of wonder if it would make perfect sense to release in early 2011 if they can get say GTA IV xxx city as a near launch game for example.

Whenever they release, it will be sometime in the Fall/Winter.

I'd expect them to hit 2011 at this point, but I wouldn't be shocked to see 2010.

I'm sure they are keeping an ear to the ground WRT the competition and if any devs are working on "next gen" titles for competing systems.

They can not afford to be last out of the gate, but it also doesn't help if you're the lone soldier and the competition isn't coming out for a couple years.

Nintendo has the difficult task of following up the uber popular Wii with something that will catch the imagination of the casual gaming public .... again. HD graphics may just do the trick. Nintendo obviously has the technological disadvantage this gen, so logic says they will release first, but they are still selling rather well at $250, years after it's been available.

They may just overlap the consoles and let them both occupy store shelves at the same time, but this would risk losing the base of customers they currently have by confusing the market.

One simple solution would be to borrow the PC business model and just have WiiHD play the same Wii games, but in HD. No improved textures, or memory, or shaders ... just rendering at 720 (or even 1080).

Developers could keep on with work as usual and customers would have less confusion. Charge a premium for the kit ($350) and be done.


Sony IMO has the easiest road forward ... I'm not sure why there was/is word of them abandoning Cell as it is the key to future success for them. IF Nvidia is no longer a good business partner to work with, they can get ATI to whip something up.

Developing the extended PS4 should cost significantly less than ps3 and the initial price shouldn't be a deterrent this time around. It's all about timing and dev support ...

 

[Brimstone]

I'm waiting on solid news on Charlie's last news bombshell concerning a Larrabee in PS4 before buying into a speculation piece where he doesn't have a source other than himself.

Well it does look like Sony is decreasing funding to nVidia regarding certain development contracts.

CPB. CPB revenue decreased by 72% to $11.9 million in the first quarter of fiscal year 2010, compared to $42.5 million for the first quarter of fiscal year 2009. The overall decrease in CPB revenue is primarily driven by a combination of decreases in revenue from our cell phone products, decreases in revenue from certain contractual development arrangements with Sony Computer Entertainment, or SCE, and a drop in royalties from SCE due to lower seasonal sales of PS3 Playstations.

http://www.gurufocus.com/news.php?id=57112



To me it seems likely the PS4 will be Larrabee2 based. Intel obviously desires a console contract to help get more game developers on-board with their software model.

 

[Shifty Geezer]

Why woudn't they get more luck with MS? The way things are set up now, Sony would prefer Cell+Stuff, whereas MS are more freeroaming. So Intel would do well to court MS with more chance of success, no?

 

[3dilettante]

There's not much substance to a statement that was purposefully made to give nothing other than the information that there was a revenue decrease.

Nothing was confirmed to have ended, nothing was confirmed as being unplanned or planned.

A more definite hint would be discussion of the ending of certain contracts, but that unfortunately would be too specific.

 

[Brimstone]

When you take into account the rumors swirling about Larrabee going into the PS4, and now a decrease in revenue from undisclosed contracts at nVidia, a breadcrumb trail is starting to manifest on the path to next-gen Playstation hardware. I thought Sony already pulled their engineers from the Austin CELL lab?


Next, Sony needs to split the sticks on the controller, improve the triggers, and put in next-gen rumble technology from Immersion (they paid for it afterall). This would help them get back onto the warpath.

 

[nAo]

Does the CELL lab still exist?

 

[3dilettante]

When you take into account the rumors swirling about Larrabee going into the PS4, and now a decrease in revenue from undisclosed contracts at nVidia, a breadcrumb trail is starting to manifest on the path to next-gen Playstation hardware. I thought Sony already pulled their engineers from the Austin CELL lab?

There is chatter that Intel has been courting all console makers for a chance to get Larrabee into the next gen. That was to be expected.

What rumors are there of Larrabee in the PS4, other than the ones spawned by Charlie's own article?

The most substance thus far is that Charlie heard it from a guy. (Still better than the apple console piece where he didn't even have that).

I have to balance that with other Sony statements stating otherwise, some decent reasons why Sony would not want to depend on Intel, and the possibility that Charlie was being duped by a source.

There may very well be truth to the claim that Larrabee will pop up in a PS4, but no substance as of yet.

 

[ban25]

From what I've heard, no decisions have been made, but unsuprisingly, Larrabee is being strongly considered by Sony.

 

[nAo]

I am more concerned about the fact that we haven't heard anything about CELL2 yet.

 

[ban25]

There's no commitment to a future Cell by either Sony or IBM. It's a one-off, essentially.

 

[Tahir2]

That is a very sad development if Cell is canned.
It seems a terrible waste of resources to design it and then not go forward with it.

EDIT: I feel bad for the software developers that had to 1. build tools to deal with a new architecture and 2. the developers that had to then use these tools to make games. Seems quite a feat to relearn how to program every 4-5 years.

 

[NRP]

Perhaps a question for nAo and/or other devs:
Would use of Larabee in one of the nextgen consoles force a new, difficult programming paradigm similar to Cell this gen?

 

[ban25]

Perhaps a question for nAo and/or other devs:
Would use of Larabee in one of the nextgen consoles force a new, difficult programming paradigm similar to Cell this gen?

Intel has committed to supporting standard graphics APIs like DirectX and OpenGL for Larrabee. They are also talking up their new ISA extensions and the GPGPU potential for Larrabee. I would say it's up to the platform designer and Intel to determine what kind of system is built around the chip.

Even if it's a single-chip solution, I'd say development might not be all that different from today if you are leveraging 3 or 4 LRB cores for game logic and treating the rest as a traditional GPU with DX as the API. It might be more dynamic than that -- imagine if cores could be scheduled on the fly, alternating between graphics and game code mid-frame. Of course, it could be a pain to develop, but there are a lot of possibilities for a highly parallel symmetric multi-core architecture.

Of course, I wouldn't mind a traditional less-parallel architecture, say Nehalem, DX11 GPU, and a very large unified memory pool. Ah, to dream...

 

[brain_stew]

You know with all the chatter about Intel courting the console makers, I'm beginning to think that Nintendo would be the best fit and most likely suitor.

We know that Nintendo are going to have to significantly upgrade their graphic capabilities next gen. if only for the fact that they need something with reasonable performance in HD resolutions, I believe Iwata has even been quoted saying that their next console will target HD resolutions. The reason I think LRB fits is the fact that if anyone is going to be enticed by a potential cost saving then its surely Nintendo.

How would a 12/16 core LRB sound as the setup for the Wii's successor? Surely that could be delivered for a very reasonable price in 2012/2013 and yet still represent one of the most significant leaps forward from one manufacturer's console to the next.

Coming from a console with such low CPU performance would LRB be able to cope with the expected CPU tasks on its own? I know LRB was never designed to run an OS or replace a CPU but this is Nintendo we're talking about if there's a chance to forgoe a separate CPU entirely to save a few pennies I could see them considering it. It may require tweaks to the LRB architecture but I'm not knowledgable enough in this area to comment.

There's the issue of BC of course, would this sort of setup be capable of emulating the Wii in software? I think it must be possible, with close to no documentation homebrew developers are making great progress with getting Dolphin upto scratch on current hardware.

If LRB's single threaded performance for CPU like tasks isn't enough to emulate the Wii's PowerPC CPU, could Nintendo simply add a die shrunk version of the Wii's CPU for compatabilities sake? I can't imagine the cost of such a chip being at all significant if produced on the latest process node.

Would adding say a 2ghz dual core out of order Power PC chip make a lot more sense, it'd certainly help ensure BC went smoothly and wouldn't be a particularly large die all things considered but obviously it'd be pricier than a single chip solution.

Really, I think my main reason for suggesting Nintendo will consider LRB is simply because I'm sure they'll cheap out! Nintendo have seemingly got to drop the old ArtX design of the Cube/Wii next gen, so any BC would likely have to come through emulation, even with ATI at the helm, so I don't see anything tieing them to their current contractors.

Getting the chip into the successor of what is looking to be the most successful home console of all time has got to sound mighty appealing to Intel, so the contracts could very well be rather lucrative. If Intel want LRB to really take off it seems like a smart investment.

 

[Silent_Buddha]

I currently don't see Nintendo having a reason to move away from ATI with an updated dx10.1/11 capable chip.

Unless something changes, ATI is very attractive from a cost control standpoint for console makers. They are willing to design the chip and then sign off ownership to the console maker. Nintendo and MS currently. This allows them to control their own costs by contracting manufacturing, etc. without regards to a 3rd party.

I'm not sure Intel would be willing to release complete control of one of their designs. So it would be a similar situation to what a console maker would have with Nvidia.

For both Nintendo and MS, it would be more of a gamble to go with Larrabee than for Sony, IMO.

Then again, Intel's manufacturing prowess may give it a bit of an edge if they want to undercut anything TSMC could offer. But would Intel be willing to cut margins that much just to gain market penetration for Larrabee?

Regards,
SB

 

[Entropy]

I am more concerned about the fact that we haven't heard anything about CELL2 yet.

There could be many explanations for that. Some disappointing, some just boring.
What I'm interested in is what application programmers feel are areas that need to be addressed with the next iteration. Are there any gotchas with the current Cell that can't be coded around once familiarity with the platform has been gained?
Almost all of the early comments was of the "this is new!" variety, and it would be very interesting to hear how the people who have a couple of years of experience programming the BE look upon it today.