Rumor: XBox dual SKU next-gen launch

[Jay]

{{ And this really seems to be rehashing a bit of the rolling-generation / no-generation discussions. }}

It's possible that from next gen onwards could get rolling generations.
Just not from this one imo, the cpu in 1X is to far out of the performance window.
Not sure how rolling would work if you start with both sku's though as, if you buy the lower one then it would only have life of 3 years not 6 for example.

 

[Silent_Buddha]

Well, the thing that I think will be the most important is to maintain the option to go either way WRT starting an new generation or continuing the current one when there is an opportunity to do an upgraded hardware release as the technology and market conditions dictate. That's how you have the best of both worlds.

Yes, it's entirely possible (roll with me for a second on the rolling generation idea) that once people get used to a faster cadence (currently called mid gen refresh) that it becomes the expected cadence.

Throw in backwards and forwards compatibility (IQ settings, AI settings, NPC density settings, whatever) to have continuous support from one generation to it's immediate predecessor or successor.

Then, once people are comfortable with that (perhaps they already are), you can throw in the occasional exclusive that requires some new hardware feature and the developer doesn't want to spend the time to make it work on the immediate predecessor until the PC version releases.

I mention PC version, because a PC version is likely to work on even worse hardware than the immediate predecessor, so a version based off of that could always be ported to the previous console (for example, the XBO-X).

Which has the result of the previous generation console getting support for the "next generation" games, but potentially after a delay that coincides with a PC version being released.

Of course if no PC version is ever released than it may never get that "next generation" title. However, as MS appears committed to supporting PC, this is likely to happen with all their exclusives. And most multiplatform developers already support PC, so again, not an issue there.

Regards,
SB

 

[mrcorbo]

I still think applying the concept of rolling generations to the console market is ill-conceived and that having defined product families where new, upgraded designs can either become a part of the existing family or will instead be the start of a new product family is a better approach. Not looking to argue that here, though.

 

[Tkumpathenurple]

What if they do the following?

1. Next-Gen Launch Model A at X TF
2. Next-Gen Launch Model B at 1.5X TF
3. Mid-Next-Gen Refresh Launch Model C at 2X TF

Plug in what numbers you'd like (6TF, 9TF, 12TF) ; (7TF, 10.5TF, 14TF) ; (8TF, 12TF, 16TF)

{{ And this really seems to be rehashing a bit of the rolling-generation / no-generation discussions. }}

I think something along those lines is ideal. IMO, the following:

1) XBoxTwo Mini.

• AAA demanding titles generally run between 1440p and 1800p.
• Approximately 7TF of GPU performance, in order to incontrovertibly declare that even their budget console is more powerful than their high end previous generation console.
• Architectural improvements too, but no RT. Maybe some rudimentary support, but not as capable as the RTX2070. Enough for things like shadows and AO.
• A primary focus on CPU and memory, in order to ensure that it can run games for a great many years and basically only face fidelity reductions.

2) XBoxTwoX.

• AAA demanding titles generally run 1800p-4K.
• Approximately 14TF of GPU. Maybe a bit more than double performance so that it can comfortably run at double the resolution, and also dial up some other effects.
• Dialed up RT hardware. Still not enough for reflections in reflections in reflections, but enough for developers to dabble.
• An appropriate increase in CPU and memory clockspeeds, and more memory.

3) XBoxTwoXXX.

• Released 3-5 years in.
• The RTX2080 to the GTX1080.
• Focused on RTRT, and getting the hardware and SDK's in dev's hands a good few years before the big RTRT generation, much like the PS4Pro and the XBoxOneX have gotten developers used to 4K before the 4K generation hits.

modedit : readability. There are lots of formatting options, people.

 

[Silent_Buddha]

I still think applying the concept of rolling generations to the console market is ill-conceived and that having defined product families where new, upgraded designs can either become a part of the existing family or will instead be the start of a new product family is a better approach. Not looking to argue that here, though.

Sure it makes sense for the console maker in order to move new hardware, but makes almost no sense for a 3rd party developer.

Now envision the scenario that MS is likely moving towards. The Xbox, just being a cheaper alternative to a PC with a more limited scope and one where they control the store.

In such a scenario, MS would "in theory" be able to always push better graphics then the competition (make the best version of a "MS exclusive" run on the top of the line PC hardware available, then scale that down to the Xbox with similar specs to the competition) if they desired.

Meanwhile, they'll also be able to address all segments of the console market at all times. No need to wait until later in the generation for that cost reduced "slim" model to attract more budget conscious buyers.

No more resetting of generations with a potential fall-off in software licensing revenue and then having to recoup that later in the generation when sales finally take off.

3rd parties being less at risk during generation transitions as they're always targeting the latest hardware (PC with only light modifications needed to run on this hypothetical PC lite Xbox game console). They can still target the console first (less hardware needed to test for QA) while using whatever level of assets from the "uber" version of the game is appropriate for whatever level of console they are releasing on.

Imagine if MS had the balls to become the graphics pushing publisher analogous to how Crytek used to push the PC. Always push the envelope on PC and port what you can to the console. Only in this case, they know what the consoles will contain ahead of time and so can have their studios tailor the engine to be more friendly towards that (not limited by it, but not engineered in such a way that you can't easily port it to those consoles).

In such a scenario, even if the next console doesn't feature RT hardware, they're already prepared for console level RT hardware because they'd have been pushing it on the PC versions of their games.

Regards,
SB

 

[vipa899]

@Silent_Buddha

I think your on to something there, Halo Infinite is supposedly being designed with the pc in mind, the pc platform being threated as a class platform. Maybe thats why the helmet has some RT-like elements?
Halo being developed with pc in mind confirms your suspicions, they could have AAA pc games with the latest tech avaible and use that further on their xbox platform, in perhaps lower grade settings, but up that ante when the next one X comes.

A link to the video

https://www.facebook.com/

 

[eastmen]

perhaps the next gen xbox will have RT tech in it . Rumors say navi but that can just be early dev kits . Perhaps what will actually appear in the console is something further ahead in the pipe line. I mean we are still expecting navi in 2019. So a 2020 console could come with a new amd design

 

[Shifty Geezer]

There's a thread for predicting console hardware that has already engaged in the possibilities of raytracing hardware. This thread is about a two-tier launch.

 

[Mobius1aic]

A two SKU launch using a Standard/Pro model seems unnecessary to me. A low tier streaming only box makes some sense, but for local rendering I think devs probably want a chance to get truly familiar with hardware before another higher end SKU gets thrown into the mix. Having a mid tier refresh also gives gamers something to look forward to. In a two tier launch, a midgen refresh might be seen as out of the question, or less likely, and any improvements may not be as drastic. Having two production lines with their own specific tooling and BOMs makes things more complicated and possibly expensive too.

 

[AlNom]

A low tier streaming only box makes some sense, but for local rendering I think devs probably want a chance to get truly familiar with hardware before another higher end SKU gets thrown into the mix. Having a mid tier refresh also gives gamers something to look forward to.

Devs should be more than familiar with scaling workloads for performance capability if they're remotely considerate of what's going on currently with multiplatform while architectural features is important in the long term.

A generation can be defined by the latter while performance can scale over time. This should be clear just looking at the nature of development and obviously the current generation setup. The existence of a higher performance tier at launch just targets the folks willing to pay more for an early performance bump where it otherwise couldn't exist while not forcing the price-sensitive folks away, potentially losing said customers to the competition while the promise of compatibility can lock them in somewhat for the future.

It simply comes down to giving the consumer choices.

Even if a console vendor has only one SKU, the multiplatform dev still has to deal with a console from another vendor with a highly likely performance differential anyway, and if they're also hitting PC, anything is fair game. The important thing is to have a baseline of architectural features.

 

[Shifty Geezer]

Yeah. The days of targeting hardware specifically ended with PS360. Now everything is cross platform and cross generation factoring in mid-gen, and all PC hardware based. Scaling down a SKU in the same family is no different to scaling down for a weaker console in the same gen or the lower of the mid-gens, or a lower spec PC. Only platform exclusives will care about highly optimised, device specific code. Heck, not even that now if mid gen is going to occur again.

 

[mrcorbo]

Having two SKUs will let them serve different segments of the market more effectively. They can have the bar-none most powerful console with the retail price to accommodate it and a console intended to sell in volume. A single SKU couldn't do the same as effectively.

 

[Mobius1aic]

Trying to teeth out and optimize the production of two different machines just seems like a costly extra pain in the ass. You're doubling the list of many materials and possibly the support base needed handle these machines. Only the software is really the same. And how would refreshes be handled? Cut the barebones model, send the high tier launch model to low end and introduce a new high tier SKU? Also, for launch do you possibly artificially limit the number of one SKU available to encourage extra sales of the other more profitable one?

Consumers say they want choice, but hand holding can be a good thing. Less confusion, less uncertainty. I'd rather see a staggered upgrade path with fairly high end launch system, and an upgraded refresh a couple years later.

 

[BRiT]

Trying to teeth out and optimize the production of two different machines just seems like a costly extra pain in the ass.

Dynamic Scaling. If your engine doesn't support it in 2020, you're still doing something wrong. Your engine should be dynamic enough to support that in 2017.

 

[dobwal]

I still think applying the concept of rolling generations to the console market is ill-conceived and that having defined product families where new, upgraded designs can either become a part of the existing family or will instead be the start of a new product family is a better approach. Not looking to argue that here, though.

Not in a reality where the console market readily accepts thin client streaming. You can have distinct hardware gens with forward compatibility offered through a streaming feature. Want to play current gen titles locally then buy the latest hardware? Don’t want to incur the hardware cost of staying current, find thin client streaming as an acceptable experience and still want to play current games then old hardware that maintains compatibility through streaming becomes a perfect way to accommodate your needs.

Release each gen with two skus. A full on console to serve core gamers and a thin client with enough juice to support the latest GUI/app features as well as all the other doodads like the latest HDMI specs or whatever is current in the TV world (8K or WCG in the future).

 

[Mobius1aic]

Dynamic Scaling. If your engine doesn't support it in 2020, you're still doing something wrong. Your engine should be dynamic enough to support that in 2017.

I was talking about actual hardware production, not the software.

 

[BRiT]

I was talking about actual hardware production, not the software.

Ah, got it. Thought you were talking about game production.

Though the production of console hardware, it doesn't seem like too much of a hassle depending on how early on you plan to support 2 models. They wont be making 2 consoles with 100% different components. You will have very large component overlaps. Overall, there's what, maybe 30% difference in components used in the console between Base and Pro levels?

I think with a 75% : 25% product split between base and professional levels (assuming slight improvement on mid-gen refresh product split numbers), you would only need slightly different product lines with slight variations for the main PCB, SOC and RAM, heatsink, and PSU. Even then you might possibly make out better through scaling up using the same heatsink and power supply blocks to larger quantities. Hell, maybe you even have identical amounts of RAM too and only scale the SOC CPU/GPU. You'll end up using the same internal storage and optical units and power buttons and other components.

 

[Mobius1aic]

Scaling only the SoC and a few other components seems like a waste of going through the trouble of having two different SKUs. What could be a savings for the higher end SKU still incurs an extra cost in the weaker by virtue of supporting both for certain components. One console is going to subsidize the other it seems like unless there is that much more money in developing two seperate SoCs or using binning from one manufacturing source (which makes more sense to me). Though the binning process could yield enough savings to make having a two SKU approach actually quite worth it when yields are such a big issue. That's alot of chips potentially getting a second chance if they are truly console custom and not part of the overall upcoming AMD chiplet ecosystem.

 

[mrcorbo]

Trying to teeth out and optimize the production of two different machines just seems like a costly extra pain in the ass. You're doubling the list of many materials and possibly the support base needed handle these machines. Only the software is really the same. And how would refreshes be handled? Cut the barebones model, send the high tier launch model to low end and introduce a new high tier SKU? Also, for launch do you possibly artificially limit the number of one SKU available to encourage extra sales of the other more profitable one?

Consumers say they want choice, but hand holding can be a good thing. Less confusion, less uncertainty. I'd rather see a staggered upgrade path with fairly high end launch system, and an upgraded refresh a couple years later.

Nearly every new smartphone that launches does so with at least 2 SKUs and every smartphone manufacturer refreshes their product line each year.

 

[Mobius1aic]

Nearly every new smartphone that launches does so with at least 2 SKUs and every smartphone manufacturer refreshes their product line each year.

I considered that, but smartphones (at least the ones people pay attention to) can command prices close to $1000 and actually have real profit margins. I see one console SKU carrying the other in a dual SKU launch.