He only talked about the Aula thing, right? Which wouldn't be a scoop, but repackaged old info about something unrelated to a revision (because it's still the Mariko Switch with nothing added).
-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
-
We have made minor adjustments to how the search bar works on ResetEra. You can read about the changes here.
Based on how tech has progressed since the Switch, how powerful could a Switch Pro in early 2021 actually be?
- Thread starter ItWasMeantToBe19
- Start date
- Status
And that's what I was thinking. If the talkis about Mariko being discontinued are true, they would have to take that into account by booking enough supply of this new chip as the replacement for standard Mariko.
I expect 128GB but not 256GB, as that is possibly enough for the Nintendo games, meanwhile expandable option for the very large third parties like Bioshock on the switch for example, or NBA2K
It works something like this "lets see what we want/need and how much that cost", if it's more expensive than they want offocurse they will cut some things, and if they have room for higher cost with targeted selling price point and profit, they can make it stronger than initially planned.
Nintendo hardware depends from potential selling price point (they dont want higher selling price point) and profit (they want profit from day one on hardware).
What backward you talking about?
We dont know for sure what Nintendo wanted/wants for this revision, we still speculate based on rumors, we still don't have official infos.
It is, but no one counts Virtual Boy, its discontinued only after one year..
I already told you, spending on a product then deciding the price. That is not it.
Also no..
They have the set planned price of the product and set the profit margins and the margins they want after x amount of time on the market. You have it backwards, it's not "let's see what we want and how much it costs" it is "built within this constraints that leads to this likely outcome in the end. What you are describing is not real.
Nintendo isn't effectively making the SoC, Nvidia is. They are going bang for buck.
It is "we want the most powerful within this price range, that we are selling at this P amount and that the cost of manufacturing lowers to this Q amount after X amount of time on the market" not "we want a 10 TFLOP machine with 16 GB of RAM, super fast SSD and a desktop class CPU for this price" (for example sake). After is when they negotiate and work out further business aspects of this that would raise the price or lower it if they so choose it is necessary for what they need. Billing everything at once for what it can be for what it can do. Not what it can do for a price they think it'll cost.
It is always bang for buck first and foremost, not the other way around. This is a mass produced item, and does not translate 1 to 1 for something you'd buy in a store for instance which you do as you are describing and then put an item back if it out of your budget range.
Yeah, and based on what we know of Orin S, available ARM CPUs and RAM Configurations able to be produced on Samsung 8nm, and how DLSS, and at least Ampere works, we can tell if they those terms were
"We want the most powerful device within 275-325 dollars that we can sell for 300-350 at launch and the cost of manufacturing lowers to 200-250 after a year on market, we would like this device to be able to run games at 4k using DLSS after developers add it to their games."
That is likely the pitch Nintendo went to NVIDIA with, or at least something close.
I think we dont understand each other (at least I have problem follow you) or simple we dont agree,
In any case my main point is that Nintendo with their hardware dont want higher selling price point and that they want solid profit on hardware from day one,
so people should have that point on mind when they talk about possible power and price point of this Switch revision.
Last edited:
So if the Mariko model's power consumption is the rough target (and thus, just entirely skip a die shrunk 2023 model), my quick and dirty way to guestimate would be...
Assume that the battery is the same size, CPU's power budget is roughly the same, GPU's power budget is roughly the same, and that any increase in power draw due to increase in memory throughput is compensated by efficiency improvement in the display (be it mini-LED or that Sharp IGZO screen or whatever)...
Then assume that TSMC 16/12->Samsung 8LPP/8LPU roughly doubles your CPU/GPU power budget... (TSMC 16/12->10 nm should be -40% power, or +2/3 power budget. IIRC, Samsung's better in 10 nm generation, then 8LPP further refined that)
Then your options are basically a mix and match of:
1. Keep core/SM count the same, increase frequency by sqrt(2), or ~41%
2. Increase core/SM by 50% (from 4 to 6), increase frequency by sqrt(2/1.5), or ~15.5%
3. Double core/SM, keep frequency the same
BUT, this doesn't take into account the changes in power efficiency from one architecture to another. Any A7_ architecture will be more efficient than the A57 used in the Switch. Thus, with the CPU, you can take whatever frequency you ended up with from above, then raise it a little bit. As for the GPU, I have no clue about the efficiency differences between Maxwell and Ampere.
But the big thing here is we are talking about a system at the level of the 8 year old Xbox One S in GPU power when docked, and pushing to PS4 level when accounting the CPU.
And we have to consider NVIDIA and Nintendo are collaborating on the Switch Platform, meaning costs on the SoC end could be eaten by NVIDIA taking a minor loss on some parts earlier on.
Either way, a One S level Ampere/Lovelace GPU with 8 A78 or A78C cores (As they are the ones that can be used by NVIDIA on Samsung Fabs) Clocked in a manner for 7 usable higher performance cores with a weaker lower clocked core for OS Functions, 8 GB of either LPDDR4 or LPDDR5, and 128gb of emmc or UFS 2.1 storage likely is within a reasonable margin for Nintendo and NVIDIA.
Now, complicated parts come on the housing/screen for portable mode which that is likely the thing that will influence price/power/poduction cost the most funnily enough as NVIDIA could eat the production cost for the CPU/GPU/IO of the SoC earlier on if needed. But the Screen, Battery, Shell.etc will be all on Nintendo.
Although I think all of that shouldn't be too hard to achieve for a 300-350$ MSRP to the end-user, especially as Nintendo got a shitload of profit on the OG Switch and still has older models in stock to burn out in order to buy time for production costs to go down on the new one.
Seriously, are we having hot takes in this thread as well? Geez...
Nintendo being one company didn't stop them from becoming the longest surviving soley-video-game-entertainment out there. Also they have several other industries that they are now involved in.
Yeah, they could make money forever off of the Zelda and Mario Franchises alone, even by licensing them out to other developers/studios .etc
The something was never for this week, it was supposedly around next week.
Sorry for the old quote. Do you have a link to your analysis?
I'm very excited to see what Nintendo can achieve when targeting Switch instead of Wii U for BotW 2.
A full 4*A77+4*A55 cluster is 12mm2 on 7 nm. I would expect a 8*A78 to be only 30~40% bigger due to A78 being smaller than A77.
TX1 cores have the same clocks due to it using the first Big.LITTLE implementation. DynamIQ cores can run at different clocks.
Exynos 9820 was ~600 GFLOPS on galaxy S9. It was the first 8LPP SOC. It is actually the only real high end SOC on this node as it is the only > 100 mm2 SOC (127mm2) compared to E880/S690/S750 that are way smaller SOCs with only 2 big cores. While the E9820 has 2 big A75 cores and 2 way bigger M4 cores (easily 3 or 4 times bigger than the 2 A75).
Looking forward to it, I can't promise anything but I'll see what I can do!
Sure:
Zelda: Breath of the Wild: The Technical Analysis
Posted on behalf of brainchild courtesy of the Adopt-A-User program, Gaming Edition. NOTE: Do you also have a thread you felt like that is worth posting, but you lack the posting capabilities to do so? Give the volunteers at Adopt-A-User a call, and we will assess whether or not your thread is...
www.resetera.com
I also compared the difference in tech between the Smash BOTW stage and the original game here:
I also did a casual analysis of the BOTW 2 trailer but there wasn't much new there; it mostly seemed to be using the same tech with some tweaks to the cel-shading (they added an AO-like pass on the characters) and shadow casting for local light sources which was absent in the first game. It did appear to be in-engine from what I can tell, based on rendering quirks I spotted.
I wouldn't draw many conclusions from my quick BOTW 2 analysis though. A lot has likely changed since then.
Awesome!
Nothing specific as I said and not related to the hard... software related...
Trademark ..
NSO ...
Collections ...
Im sorry, would it be really that far out to target base PS4 specs? (docked)
It would be 7.5 Years from the PS4 release, that would need an increase of 3-4 times the power of switch as far as i remember?
if its "just" a pro i could see them not wanting to go there, to leaver room for improvement for switch 2.
But if its an iterative model, where we get a switch every ~4 years and have 4 year cross gen, then why not.
The only real reason why a want a switch pro is that BotW2 can use increased ram/bandwidth/ etc for stable 30fps@1080
with higher resolution textures and higher LoD (And HDR)
It would be 7.5 Years from the PS4 release, that would need an increase of 3-4 times the power of switch as far as i remember?
if its "just" a pro i could see them not wanting to go there, to leaver room for improvement for switch 2.
But if its an iterative model, where we get a switch every ~4 years and have 4 year cross gen, then why not.
The only real reason why a want a switch pro is that BotW2 can use increased ram/bandwidth/ etc for stable 30fps@1080
with higher resolution textures and higher LoD (And HDR)
Spoken in riddles like a true prophet. Whatever happens, that thread has been a great time killer.
Not sure if you are serious, but botw 2 will run like a charm on this no matter what.
Botw 1 was ported in a relative hurry from a completely different architecture with incomplete tools etc. Botw 2 will be developed in tandem with the new chip which is (if our speculation is half right) a lot more powerful across the board and an iteration on the same architecture.
I assumed Zelda was what you were discussing for a while but I didn't want to speak for you.
Still think it's likely next week or around there?
The only way to get 8 homogeneous Cortex-A78 cores is go for the octa-core configuration of the Cortex-A78C.
2 clusters of the 4 Cortex-A78 cores and the 4 Cortex-A55 cores configuration will also give 8 Cortex-A78 cores, but in a heterogeneous configuration rather than a homogeneous configuration.
Yeah, I was meaning 7 clocked at like say 1.5 ghz when docked and the other clocked at ~1.0ghz for the OS
No worry....
things can change
But the starting point is the NSO
Cannot talk bout that. Just wait and real infos are incoming
March Switch 2021 reveal confirmed.
Would line up with Nintendo seemingly hyping the Anniversary of the Switch in advertising.
Was making a minor joke, because of
That's all
Even at work and you send an email .... people always take it as a harsh one 😂😂😂
What do you mean? They give the price they want the product to be to make, and want the most out of that price(can be specified in what areas need improvement but not a literal exact spec), in a way that after a certain amount if time the profit made per unit that they are selling increases. The partner presents config, they are all estimated to be the same price or around the same and are different in actual spec. N decides from there and thats history on that.
that's literally it.
Last edited:
But what you said doesn't make sense because that's not really how that happens.
But thats how Nintendo operates and those are some ofe their goals when they releasing hardware, again,
Nintendo with their hardware dont want higher selling price point and that they want solid profit on hardware from day one.
I wouldn't say that they made a solid profit on Switch sold from day one tho. Profit margin was pretty thin in March 2017.
Yes, we dont know how big profit were made with Switch on day one, but its certain they are making quite solid now when they were selling with profit from day one. My point is that Nintendo would want to have similar profit with new revision/s from day one like they having with current models.
For instance, when Switch Lite was launched in September 2019. Nintendo said they have similar profit margin with Lite like they have with main Switch model.
I have no idea what these "changes" could be for NSO, or what exactly is planned, but all I hope is that whatever it is, it signals that Nintendo is finally ready to start taking their online services(s) much more seriously.
It's just embarassing at this point. NSO is objectively worse than Xbox Live from almost two decades ago.
Switch BOM was 257$ at launch:
Nintendo Switch costs $257 to make, won’t line Nintendo’s pockets suggests report - KitGuru
Despite a number of teardown sites already having published complete guides on the Nintendo Switch,
www.kitguru.net
Factoring in the retailer's cut and distribution costs, Nintendo made very little profit day one on the console.
As a comparison, PS4's BOM was 381$ (399$ retail price) and was sold at a loss at launch.
Switch BOM was 257$ at launch:
Nintendo Switch costs $257 to make, won’t line Nintendo’s pockets suggests report - KitGuru
Despite a number of teardown sites already having published complete guides on the Nintendo Switch,
Factoring in the retailer's cut and distribution costs, Nintendo made very little profit day one on the console.
As a comparison, PS4's BOM was 381$ (399$ retail price) and was sold at a loss at launch.
Nintendo itself said they are selling Switch hardware with profit from day one, but based on that BOM estimate that profit wasnt big..
Yeah, in my opinion, they don't want to take a loss on HW day one but the initial profit on hardware is not a priority for them.
More often than not, it is in your best interest to maximize the value of your product given the price enveloppe (and then margin will increase with time and SW will benefit too).
Reggie had included an attach rate of 1 1P software in that profitability calculation in prior interviews. We really don't know on what basis that tweet is referring to.
Also The $270 bom estimate doesn't contradict the tweet especially with softwate included. Botw briefly had a greater than 1 tie ratio at launch. Meaning they sold more Zelda games than Switch hardware
Last edited:
Switch BOM was 257$ at launch:
Nintendo Switch costs $257 to make, won’t line Nintendo’s pockets suggests report - KitGuru
Despite a number of teardown sites already having published complete guides on the Nintendo Switch,
Factoring in the retailer's cut and distribution costs, Nintendo made very little profit day one on the console.
As a comparison, PS4's BOM was 381$ (399$ retail price) and was sold at a loss at launch.
I'm curious to what accounts for $45 in technology licensing costs per unit alone. Immersion patents?
If there's an alternative to paying per-unit, they may have taken it even if that would be a very expensive deal up front.
With that said, if they are paying a $22 licence for each joy con, it would somewhat explain why the Switch Lite is $100 less with a similar profit margin.
Last edited:
- Status