Rumor: XBox dual SKU next-gen launch

What IF (pardon little provocative style)
Lockhart (which as codename imply will be locking hearts of casulas by)

-low price
-something in specs clearly handicaped entire generation like arcade lack of hdd or bd-drive

Anaconda (squeezing hearts the loudest , remeber your legacy next month games will give you another 3000)
-ohhh such 20%more pixels, not as such dirty ps5 , best place to play , let's play Gaaas atempt #14 BTW also available on windows store, where can be playd in better numbers.

Would this outcome be a fair good and ambitious Microsoft vision to console gaming? Will it reach nex generation ambitions? or do you feel something else?
 
What IF (pardon little provocative style)
Lockhart (which as codename imply will be locking hearts of casulas by)

-low price
-something in specs clearly handicaped entire generation like arcade lack of hdd or bd-drive

Anaconda (squeezing hearts the loudest , remeber your legacy next month games will give you another 3000)
-ohhh such 20%more pixels, not as such dirty ps5 , best place to play , let's play Gaaas atempt #14 BTW also available on windows store, where can be playd in better numbers.

Would this outcome be a fair good and ambitious Microsoft vision to console gaming? Will it reach nex generation ambitions? or do you feel something else?

If your starting premise is, "What if they implement the strategy poorly, will it be a good outcome?" isn't the answer going to usually be "No" no matter what the strategy is?
 
If your starting premise is, "What if they implement the strategy poorly, will it be a good outcome?" isn't the answer going to usually be "No" no matter what the strategy is?
Unless you have exceedingly that rare blind luck that results in the creation of penecillin, champagne or cornflakes. Go incompetence FTW! :nope:
 
If your starting premise is, "What if they implement the strategy poorly, will it be a good outcome?" isn't the answer going to usually be "No" no matter what the strategy is?

Starting premise would be this : New hardware budget for sizable jump is very tight in any case isn't it? That cheaper sku would need to have meaninufull cuts in hardware to reach that spacing and price difference, not some 200mhz here and there. That would have implications for rest of the pack in lowest common denominator development realities.
 
Starting premise would be this : New hardware budget for sizable jump is very tight in any case isn't it? That cheaper sku would need to have meaninufull cuts in hardware to reach that spacing and price difference, not some 200mhz here and there. That would have implications for rest of the pack in lowest common denominator development realities.

Not necessarily. In the past, $100 price differences were justified by increasing hard drive capacity. The price difference isn't necessarily dictated by an increase in the BoM. In the case of the split SKUs here, I think the premium SKU won't even try to justify itself from a price/performance standpoint. It will be expensive because it's the best and that price will serve to move it out of the volume sales slot in the product stack. Since the higher SKU won't have to sell in volume, they can use more aggressive binning and deliver a higher-clocked part than they would otherwise. I don't expect the basic building blocks of the two SKUs to be meaningfully different.

Think more of the difference between the $799 (at launch) 2080 vs the $599 (at launch) 2070. Did the 2070 have $200 worth of an actual hardware difference with the 2080? Doesn't seem like it.
 
Think more of the difference between the $799 (at launch) 2080 vs the $599 (at launch) 2070. Did the 2070 have $200 worth of an actual hardware difference with the 2080? Doesn't seem like it.

Indeed, which is a fair analogy and to some extent can also apply to mobile phone SKUs in that sense as well. I think folks on the internet are overblowing the need for the SKUs to be wildly different for that sort of reason; we are enthusiast elitist (as can be for console peasant) techheads here versus the mass market, so it'll be more about how MS would deliver the message.

I can see the basis for the concern being the existence of the mid-gen twins, which are an example going against the pricing model we see in other markets, however, it's ignoring that the midgen twins arrived 3-4 years after the base launch. One could also argue it's then a comparison between the Slim models and the Pro/X, but again, we have to look at the context of the market introduction there. Tens of millions of folks already bought into the platform for 3 years, so the cost of folks upgrading is the initial price they paid plus the cost of the midgen twins. Folks new to the platform have the choice of it, but they also waited years before jumping in at either the cost-reduced slim or the higher priced HW upgrade.

We certainly weren't going to get a Scorpio in 2014 or a 4Pro in 2013.

Early adopters have an option to get the best of the best or they can spend less at launch and be secure in the messaging of future compatibility with HW updates. If the option didn't exist, it makes things a lot trickier for the HW at launch because of how expensive things get for higher performance when we have inflation, slowing down of fab nodes, periphery components that are difficult to cost-reduce.

If the higher end SKU didn't exist, we'd probably still just see the lower end SKU anyway with zero option for those who are willing to go the extra $ for a higher performant device on their platform of choice for the next 3-4 years, so it's a question of whether that extra amountis worth it now or pay XXX dollars for the pleb base SKU then spend XXX again in 3-4 years vs somewhere a little in between at launch or simply not join the generation at all for another 3-4 years if they're that concerned about the money situation.

Nothing is saying one has to buy the higher end SKU at launch - it probably wouldn't have been offered otherwise, so make some Adult Decisions.
 
Last edited:
If thats the case then ok. Question remain, if is simple upselling sku, why different codename? I don't recall precedent in console space for such thing, sounds like something of different target. What exactly would they accomplish with overclocked parts on launch to bother with it ?Maybe things will come down to resolution again 1080p for lockheart and 4k for anaconda. It's consumer market and perceivable diffrence must be there and certainly it would be hard to reach difference similar to midgen upgrades at launch.

To be clear, I am all for optional models but working as extension of solid base platform , but definitely not it's if cheaper one is cheap attempt to undercut competition, muddy the water and reframe foundations of console market, cuz fk instead of dedicated platform to push and optimize for it's all pc with windows store whenever we need next one now and next 5 forza games are already on assembly line all for the greater good of console gamers. This one will grab casuals and this one will shoot from the rooftops games plays best on x, do you have problem with sellling your gems now sony?, future without generation :S
 
1440p is passable on a 4K screen and looks great on a 1080p screen.

If 1440p is the baseline of the lower end SKU, then a PS4->PS4Pro power difference makes the most sense IMO. Same with the silicon: double up the GPU and clock it a touch higher.
 
If thats the case then ok. Question remain, if is simple upselling sku, why different codename? I don't recall precedent in console space for such thing, sounds like something of different target. What exactly would they accomplish with overclocked parts on launch to bother with it ?Maybe things will come down to resolution again 1080p for lockheart and 4k for anaconda. It's consumer market and perceivable diffrence must be there and certainly it would be hard to reach difference similar to midgen upgrades at launch.

To be clear, I am all for optional models but working as extension of solid base platform , but definitely not it's if cheaper one is cheap attempt to undercut competition, muddy the water and reframe foundations of console market, cuz fk instead of dedicated platform to push and optimize for it's all pc with windows store whenever we need next one now and next 5 forza games are already on assembly line all for the greater good of console gamers. This one will grab casuals and this one will shoot from the rooftops games plays best on x, do you have problem with sellling your gems now sony?, future without generation :S

When console makers are determining their specs, yields are a factor. With a single SKU, if the testing of the capabilities of dies that come out of test runs result in a bell curve that says, "this % of chips can satisfy at least this performance requirement, any that can't must be thrown away" you are going to have two factors gating your performance targets. "Can we produce enough dies at this performance level to meet production targets?" and "Is the number of discarded dies tolerable?" There will be some dies produced that are exceptional. They will be fully functional and/or will be able to clock higher at lower power. The additional capabilities of these dies cannot be exploited in the device they will be installed in. Having a higher-end SKU with a lower production target, though, creates a path to more fully exploit the capabilities of these exceptional dies.
 
Having a higher-end SKU with a lower production target, though, creates a path to more fully exploit the capabilities of these exceptional dies.
Indeed, but the question really is how much more power does that actually give you, and how much can you charge for that box to make the business move sensible.

What is the maximal performance delta between a typical console die with a bit of redundancy and a perfect die with no redundancy and better clocking? I'm thinking 20%?
 
Indeed, but the question really is how much more power does that actually give you, and how much can you charge for that box to make the business move sensible.

What is the maximal performance delta between a typical console die with a bit of redundancy and a perfect die with no redundancy and better clocking? I'm thinking 20%?

I'll quote myself from the last time this came up.

MS managed with the One X to get 43% more performance vs. what Sony managed on the Pro on the same process with just 4 extra CUs (the same number of CUs that are disabled on the Scorpio die) and increased clocks enabled by the extra cooling and power delivery added to the design.

And supplement that by saying the price difference doesn't need to have a direct relationship with the performance delta any more than the price difference between a 2070 and 2080 has a direct relationship to the performance delta between those two parts.
 
Indeed, but the question really is how much more power does that actually give you, and how much can you charge for that box to make the business move sensible.

What is the maximal performance delta between a typical console die with a bit of redundancy and a perfect die with no redundancy and better clocking? I'm thinking 20%?
Considering this midgen saw the market hesitate between skus offering 2x to 4x more power for $100 more, I am a bit skeptic about using the same die for both, unless it's really designed for the low sku in mind and the second one is just because they can, selling a very small number.

In order to have continuous stock of both skus they need to match the binning based on expected demand split. Too agressive and they get stock issues, too conservative and the high sku is too small of an improvement.

If the cloud is important, they might instead have two different dies. Basically, the big one primarily developed for the racks, and then they might as well use it also for a high end sku. The low end sku precisely developped for console efficiency. It prevents one sku demand affecting the other. They can adjust production at will with no stock issues. It also keeps all skus at a good efficency point, versus the clusterfuck that happened with vega64 TDP. And of course it allows to precisely target performance and cost individually.
 
In order to have continuous stock of both skus they need to match the binning based on expected demand split. Too agressive and they get stock issues, too conservative and the high sku is too small of an improvement.
That's a very good point. If you offer a PS5+ that everyone wants instead of the PS5-, demand for that will outstrip supply and your platform sales will be impacted. If you offer a PS5+ that not many want, I guess you could always start disabling good chips to make PS5-, but then you've had all that faf and cost of developing and positioning the two SKUs. It's actually quite the gamble to go dual-SKUs.
 
I'll quote myself from the last time this came up.

MS managed with the One X to get 43% more performance vs. what Sony managed on the Pro on the same process with just 4 extra CUs (the same number of CUs that are disabled on the Scorpio die) and increased clocks enabled by the extra cooling and power delivery added to the design.
Is it me or does the math there fail? 28.6% higher clocks * 111% CU count == 32% faster theoretical GPU flops.

Taking PS4 Pro's 4.12 TFs, * 128.6 for higher clocks, and * 1.11 for more CUs, gives a flop count for XB1X of 5.887 TFs instead of 6.

Still, going by this example it looks like ~30% is the difference between a base SKU and enhanced. It's probably most economical to not worry about CUs and just clock higher, at which point aren't you just needing a better (more expensive) cooling solution? Woudln't it be better to clock the base unit higher and just have a 30% faster machine overall?
 
One way of keeping costs low on the higher SKU is to not bundle it with a Controller. I expect Xbox One controllers to be compatible with NextGen Xbox. Even if the controller has new functionality, sell that seperate. Yes, the overall price to the consumer is higher, the price of the console SKU itself is lower, and it raises the profit of the higher SKU when the consuner buys the controller seperately.

I think that is acceptable if they're trying to hit some magic threshold price point (for instance only $100 or $200 more).
 
Is it me or does the math there fail? 28.6% higher clocks * 111% CU count == 32% faster theoretical GPU flops.

Taking PS4 Pro's 4.12 TFs, * 128.6 for higher clocks, and * 1.11 for more CUs, gives a flop count for XB1X of 5.887 TFs instead of 6.

Still, going by this example it looks like ~30% is the difference between a base SKU and enhanced. It's probably most economical to not worry about CUs and just clock higher, at which point aren't you just needing a better (more expensive) cooling solution? Woudln't it be better to clock the base unit higher and just have a 30% faster machine overall?

What did you do there? You somehow got all of the numbers wrong.

PS4 has 4.2 TF : 36 * 64 * 2 * 911 / 1000000 = 4.2

Xbox One X has 6 TF : 40 * 64 * 2 * 1172 / 1000000 = 6.0

(6 - 4.2) / 4.2 * 100 = 43% increase

Edit: Well starting from 4.12 for PS4 seems to be most of where you went wrong, and the rest is down to rounding.
 
What did you do there? You somehow got all of the numbers wrong.
Dumb old me added instead of multiplied the two percentage increases. :oops:

1.286 * 1.1111 == 1.43; 43% increase.

Edit: Not quite; I wasn't that dumb. Did 28.6% x 1.11% but represented the 28.6% wrong....28.6 x 1.11 instead of 1.286 x 1.11
 
Last edited:
Considering this midgen saw the market hesitate between skus offering 2x to 4x more power for $100 more, I am a bit skeptic about using the same die for both, unless it's really designed for the low sku in mind and the second one is just because they can, selling a very small number.

In order to have continuous stock of both skus they need to match the binning based on expected demand split. Too agressive and they get stock issues, too conservative and the high sku is too small of an improvement.

If the cloud is important, they might instead have two different dies. Basically, the big one primarily developed for the racks, and then they might as well use it also for a high end sku. The low end sku precisely developped for console efficiency. It prevents one sku demand affecting the other. They can adjust production at will with no stock issues. It also keeps all skus at a good efficency point, versus the clusterfuck that happened with vega64 TDP. And of course it allows to precisely target performance and cost individually.

That makes sense, too. It'll be interesting to see how it plays out.
 
Back
Top