Yeah I'm the one who asked him the question lol. No idea what CMT is, but he pretty much shot down Bristol ridge. So maybe it is a Puma+ APU?
Reddit [verified] User shares NX info: x86 Architecture, Second screen support etc.
- Thread starter DaSorcerer7
- Start date
- Status
DaSorcerer7
Member
James Scott
Banned
They can pump out a ton of mid tier games which is what made the 3DS so compelling to most.
I'm not entirely sure on the 3rd party situation, but if they can secure the exclusives the handhelds usually get (monster hunter, Yokai watch, Atlus games) and a few major 3rd party titles like DQXI, KH3, COD, and madden them maybe?
I think the new ecosystem will attract a lot of Japanese developers, but it's a bit of a wait and see situation
Let it goooo let it gooooo.
It's not efficient for gaming apparently.
Puma+ or it's not x86 at all and will go with A72's
Schnozberry
Member
CMT is cluster based multi-threading. It was AMD's attempt to compete with Intel's implementation of simultaneous multi-threading, better known as Hyperthreading. It failed miserably, and is a poor implementation of the technology that required a lot of extra die space and hardware and ended up handicapping AMD's performance for the better part of a decade.
Bristol Ridge wouldn't make sense next to A72 cores. Too much heat, too little performance advantage, and it wouldn't scale nearly as neatly to a handheld.
DaSorcerer7
Member
No x86 is good for handhelds. Likely the handheld uses ARM and console uses ARM or x86.
I'm hoping for 8 A72's at 2.2Ghz and 8GB of GDDR5x.
James Scott
Banned
Here's how much room the disc drive takes up on the tiny wii U
Less moving parts means less heat, energy/power needed, less potential damaged units, etc
Not sure if it's worth the limitations of cartridge games
Plinko
Wildcard berths that can't beat teams without a winning record should have homefield advantage
That's the thing--if they go to a small cartridge slot, that's a decent amount of space that can be used for better things. I know Nintendo has been adamant about not putting out a gigantic XB1-sized box, and no disc drive can help that a ton while still managing a decent amount of power.
Dropping the disc drive means more space, means better ventilation, means less heat, means less need for a fan (or a less powerful one at least... mayve a heatsink?), hence less noise and less power (though i imagine a fan doesn't use that much energy in the WiiU). So it certainly makes some sense.
James Scott
Banned
Would the lack of a disc drive actually work?
I imagine the only reason one would want that is for the handheld games to work on the console, but to do that you'd need to put the console assets into the tiny cartridge.
That would either not work, require you to connect to the internet (which people won't be too fond of) or increase the cost of producing a retail game.
I acknowledge the benefits to the hardware, though. But, the benefits of loading wouldn't necessarily be there, right? Don't most games install to the hard drive now a days with minimal need to load?
I imagine the only reason one would want that is for the handheld games to work on the console, but to do that you'd need to put the console assets into the tiny cartridge.
That would either not work, require you to connect to the internet (which people won't be too fond of) or increase the cost of producing a retail game.
I acknowledge the benefits to the hardware, though. But, the benefits of loading wouldn't necessarily be there, right? Don't most games install to the hard drive now a days with minimal need to load?
Just to play devil's advocate, if the HH is coming and plays all the same games I wonder if they shouldn't just launch it with the console. But with the lack of HH leaks I'm skeptical it's coming this year. The Trev leaks also points IIRC to only one form of NX coming this year, but I guess it doesn't explicitly do so. Hmm. I do wonder of they'll really go with a fully shared library. I'm not convinced like I used to be that they'll do so. I still think there's a good chance for a partially shared library and major exclusives on each.
I couldn't agree more.
I thought NLG is pretty small. I assumed they were mainly tied up with Fed Force.
Based on Shikamaru's posts, I'm skeptical we will see MK9 so soon, unless having the MK8 engine already helps them to make it more quickly.
Gotcha. I appreciate the analysis, thank you. Some excellent points in here for sure.
frankie_baby
Member
It would fit assets for both. Faster loading too. Less space necessary, less moving parts, better ventilation and heat dispersal. The increased costs could be worth it in the long run.
R
Rösti
Unconfirmed Member
Feel free to take a look at the original thread about a Nintendo console without optical disk: http://www.neogaf.com/forum/showthread.php?t=1099932
Dark Cloud
Member
Having the cartridge slots sounds more worth it than the disc slots. I think being cheaper for the console and having the capability to put the same game into console or handheld is big if that's what they're going for.
If you look at the space the drive takes up in the WiiU, it's not just for fast loading and cross platform.
Hey, don't count those chickens just yet, Mr. Mike. They haven't revealed the system yet. It's possible it won't be revealed and released in the same year. 😜
I think even with just a partial shared library (as I suspect they might go with), the other factors you mentioned as well as unknown features of the system or systems can help make it a success. I'm just trying to temper my expectations on that shared library front I suppose.
Schnozberry
Member
Also a disc drive can't even run modern games, so they'd also need a big HDD to install everything to. Whereas if they go with fast enough carts they can again get away with limited internal NAND memory for some DLC, patches and the odd eshop/VC game, and make the all digital minority buy an external drive.
That's a lot of cost savings, and given the average attach rate for a console is around 10:1 at best it might be worth eating some higher production costs for software to reduce teh hardware cost.
I do think that we'll see a shared library to some extent, though certain games that are too ambitious for the handheld will likely remain console-exclusive (like how the n3DS has Xenoblade exclusive to it).
Which reminds me, did the x86 Reddit guy get unverified?
Nx is 2016, 14nm won't be ready for nx.
Nicer shadows (in the sense of higher resolution shadow maps) are something that scales extremely easily, but having "improved lighting" on the home console relative to the handheld generally means having to maintain two different lighting systems, and having to optimise two different lighting systems, which is often the most challenging aspect of engine development. At the easiest end of the spectrum you can simply disable features of your engine on the handheld, but you risk having the handheld versions of games look like low-spec PC titles; flat and devoid of the game's aesthetic character. At the other end, if you're using deferred rendering on the home console and simply can't manage that on the handheld, then you have to rewrite huge chunks of your game engine under a completely different paradigm.
Obviously it is possible to develop games across highly divergent hardware capabilities; look at PS4/Vita cross-play games or even Smash 4. The further the two are apart, though, the less you can share between them in terms of code, assets, and even areas like QA and optimisation. These are all things which cost money to duplicate, and which Nintendo have talked about not wanting to have to duplicate with future games.
Notice that, for the first time, AMD are releasing their entry and mid-range cards first, and not their high-end. The reasons for this would be two fold. The first is that, on a low-yield process, smaller dies are much better value (as the probability of a fault increases exponentially with die size). This is why we're seeing so many phone & tablet SoCs on 14nm & 16nm before much bigger CPU/GPU dies. The second reason is that low and mid-range chips are also used as laptop GPUs, where margins are far higher.
Traditionally for desktop chips you would want to move your production over to a node when the cost per unit performance is lower than your existing node (which factors yields, dies per wafer, cost per wafer, clock increases, etc.). In this case, though, I don't believe cost per unit performance on 14nm or 16nm will be lower than 28nm for desktops/consoles until next year, which is why we're not seeing Vega (and likely Nvidia's high end chips) until then. However, with a likely big jump in energy efficiency with the new nodes, the cost per unit performance will be much better for laptops this year, when AMD launch their first new cards.
AMD will probably launch the desktop 470/470X using Polaris 11 and 480/480X for Polaris 10, but it's unlikely that they'll make much money off them initially. Like the Pitcairn chip used in the 370(X) and Tonga in 380(X), though, they'll command far more money as laptop GPUs. In that sector they can, with relatively small 14nm dies, considerably outperform the existing options in terms of both maximum performance and performance/Watt, and there's scope for AMD to make a lot of high-margin sales and potentially grab a bigger chunk of the laptop market (or even lose what they have if Nvidia beats them to the market with 16nm chips).
I wouldn't be surprised if AMD have been sitting on Polaris for a long time, and I don't think it's the architecture that's the issue for Nintendo, but rather the maturity of the node. Particularly when you consider that they would have had to make this decision in late 2014, it would have been very risky of them to commit to 14nm at that stage purely on the basis of predicted yields. At that stage they probably would have been particularly keen to get a successor to Wii U out in a timely fashion, and I doubt they would have had much of a tolerance for risk in that regard.
All the above points, incidentally, feed into why I believe it would be a possibility (although still unlikely) for Nintendo to go with a 14nm Polaris based SoC for the handheld. The handheld is more heavily dependent on perf/W, so the cost per unit performance differential will make it viable well before it makes a home console viable. It's also a very small die, which means even on a low-yield process you'll still get a relatively good number of well-functioning dies. On the last point, Nintendo would probably have been fairly confident about the 3DS's longevity in late 2014, so the risk of a handheld launch being pushed back to 2017 might not have worried them so much. Then, of course, there's the fact that a higher-performing handheld (and a GCN-based one at that) would save them money in development costs over its life by allowing them to share more code and assets with the home console.
While I also think a return to a friend code-like system is unlikely, it's worth noting that Miitomo has an odd friend system that doesn't allow direct adding except via other social networks. There *is* a precedent for Nintendo using unusual methods even today.
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
Re fab nodes, 14nm is nowhere near 'fine' ATM.
There are 3 sub-20nm SoCs by chinese vendors, and all of those are absolutely top of their respective vendors' crops. I'm referring to Kirin950 (16FF, out now), Spreadtrum SC9860 (16FF, Q2) and Helio P20 scheduled for Q3 this year (16FF). That's it.
Also, chances are NX console's SoC/APU will be far larger than any of those, and thus, suffering much worse yields than any of those mobile SoCs if it was produced this year.
James Scott
Banned
A lot of Nintendo games are very tiny, even on Wii U a lot of games are sub 5GB.
I think 3D World is less than 2, so it's possible that even the HD versions of 1st party cross buy games will be on the smaller sized carts with smaller carts=smaller costs.
This would work perfectly with most of Nintendo's output, but games like Xenoblade and potentially Zelda might be too big to be worth the cost.
Big issue with multiplatform games that won't even be on the handheld like Call of Duty or Assassin's creed which are around 50GB in size.
Unless they can convince consumers to go digital in most cases which is unlikely.
Being able to play the single game on either NX device would be a big benefit, though.
You don't have to worry about overproducing one version while the other will sell more since it's just one copy of the game, it increases value of the software, production lines at Nintendo would only have to press one type of format, you don't have to worry about retailers not wanting to give a lot of shelf space to the Wii U.
So...I'm not sure. Even if they can get away with putting in less HD space, it'll look bad if it's less than 500 and the 1TB ones will look more appealing to consumers.
I'd be up for it, but If they follow through with it I'd hope they handle it well.
Probably won't be an N64 type scenario, but less risks might be better
Unless for console exclusive games you'd need to download a big patch to make them function? Would people be ok with that? Or maybe Nintendo can choose not to collect royalties which would be unfortunate
Additionally, I might just be overthinking things and it won't be nearly as expensive as I'm thinking
I don't know why you're discussing Polaris.
Good luck building and testing a console when one of its component isn't even supposed to hit market once production has started.
NX home built has supposedly started in late 2014, it can't rely on a component nearly 36-48 months away. Unless it's 2017. Don't forget there is big contracts and Nintendo will ask for some custom chip. Those things needs time and planning far ahead.
"Industry leading chips" probably meant "Industry leading 2015 chips on the consumer market". This is where tech-GAF should look IMO.
Good luck building and testing a console when one of its component isn't even supposed to hit market once production has started.
NX home built has supposedly started in late 2014, it can't rely on a component nearly 36-48 months away. Unless it's 2017. Don't forget there is big contracts and Nintendo will ask for some custom chip. Those things needs time and planning far ahead.
"Industry leading chips" probably meant "Industry leading 2015 chips on the consumer market". This is where tech-GAF should look IMO.
I was about to post this. What were industry leading chips in fall of 2015? Industry leading CPU. Industry leading GPU. Industry leading ram. Industry leading fabrication.
What APU was industry leading in 2015 that could fit into a console for $399?
Puma+, Bristol Ridge, Intel Atom CPU, GDDR5, etc.
That's what i want to know. Even though industry leading is vague. For all we know it was a PC with a 980Ti and 16GB of ram running an unoptimized game in a SDK.
I admit I have no idea what the actual costs would be.
But you had 4GB carts on 3DS 4 years ago. Technically the limit was 8GB but no games went that high and admittedly only a handful of games went as high as 4GB but we don't really know if that's because of the manufacturing costs or the fact that very few 3DS games look like RE:Revelations, and the limited hardware and resolution meant that very few games needed that much space.
Just a lazy application of Moores law (which I don't know if it's applicable to mask rom at all should give you close to a Bluray disc by the time NX is released. It'll be more expensive than a disc for sure but maybe not as much as one might think.
I'm not saying I expect it to happen, but it may not be completely impossible and part of me does long for the return of carts on consoles (even though I'll likely buy nearly everything digital anyway)
An AMD Carrizo based CPU and a Tonga based GPU? An 8 core 28nm Carrizo should draw less power and improved performance over the Jaguar cores in the PS4/XB1 and the improved triangle throughput and bandwidth saving features of Tonga would allow Nintendo to go with something cheaper than GDDR5 but still remain competitive in performance.
Schnozberry
Member
Industry leading is probably a cagey way of saying "in game consoles". If it was slightly more powerful than the Xbox One or PS4, that would qualify.
It seems like the biggest hindrance to performance on the PS4 has been the CPU. If Nintendo can get close to the PS4's GPU performance while offering a modest boost in CPU power, it will be a much more balanced system overall.
Even if they went conservative with Puma+, they could still hit higher clocks than were possible with Jaguar on the same fabrication node. Moving up to a more modern GCN design should offer the same advantages for the GPU, with the more efficient instruction set and better tessellation performance. They probably wouldn't need as many compute units to achieve the same performance.
It's not that cut and dry. Whatever ends up in NX, it will have been designed and built alongside whatever AMD have had planned for the last few years.
That's what AMD will have pushed when pitching as that's what will be most profitable/beneficial for them & their fabs to produce over the next 3/4 years.
D
Deleted member 465307
Unconfirmed Member
No idea, but he said he made a mistake and that it's her today in a tweet.
DaSorcerer7
Member
So Polaris is still a possibility?
D
Deleted member 465307
Unconfirmed Member
This what I'm hoping. Since Nintendo tended to have crazy small file sizes in their first HD generation, I would expect a 64 GB cap on console carts while most of their games are on 32 GB, 16 GB, or 8 GB ones. I expect their file sizes to increase a little with more power and a 1080p resolution standard, so maybe they'd still some 4 GB and 2 GB carts but not as many as Wii U would have had. For the handheld, I think a 8 or 16 GB cap is reasonable, but I could see it going up to 32 GB cap for cross-platform purposes (like Smash).
So thinking about Moores law and cartridges I started doing some calculations
I still have no idea how applicable it really is for mask rom, but doing 2x increase every 18 months, 8MB in 1996 (Super Mario 64) would be equivalent of 80GB today. So if it's anywhere close to comparable we're probably not going to see an N64 situation with super expensive carts that don't come close to the storage capacity of discs.
edited because I calculate like an idiot and it got 3x too high.. but still a decent chunk of storage
I still have no idea how applicable it really is for mask rom, but doing 2x increase every 18 months, 8MB in 1996 (Super Mario 64) would be equivalent of 80GB today. So if it's anywhere close to comparable we're probably not going to see an N64 situation with super expensive carts that don't come close to the storage capacity of discs.
edited because I calculate like an idiot and it got 3x too high.. but still a decent chunk of storage
I honestly don't see them releasing both versions (handheld and console) of a game in the same cartridge. I fully understand there are benefits by doing so but at least I can think of one or two odds by pursuing such goal and most of them come from the handheld division.
1. With the current pricing model, either the console or handheld will have to give in.
Some time ago someone mentioned about a flexible pricing as: X game can be $40, Y game can be $50 and Z game can be $60 depending on the work done and IP.
Say Mario Kart 9 is one of the $60 games, it would negatively impact their business from people who have been playing exclusively on handhelds and refuse to pay $20 more for the assets of a game (console version) they will not play at all.
2. Packaging both versions in the same cartridge would mean they share the same box and retail space. Even if said box has neon lights shining bright saying X game is strictly console exclusive (because there definitely will be) there's a really really high chance people (parents mostly) will ignore that and still purchase the game for little Timmy's handheld, leading to a lot of problems between parents, retail stores and Nintendo itself.
The same will happen for handheld exclusive games.
Well, ask the cashier if a game works for X device first before buying!!! - Some of them are even clueless than the costumer, don't care or simply will say anything in order to sell games.
3. By releasing both versions in the same cartridge Nintendo basically would cut a big chunk (over half) of their revenue from those who actually bother to purchase both of their platforms and wouldn't do it anymore because the game is basically the same between devices**. Nintendo first and foremost is a business so I don't think it's the smartest thing to do.
Now, I know they already have quite a few titles that allow cross buy where you purchase one version and get the other one for free but that's mostly for smaller projects. We have yet to see it working grand-scale with all their AAA IPs.
**This is more of a side effect of unifying their development teams as those who usually get both platforms get them because of the different experiences each game has to offer compared to the handheld/console counterpart.
. e.g. 3DS Super Mario 3D Land -- Wii U Super Mario 3D World.
Now it would simply be let's say... Super Mario 3D Universe for both platforms.
This is basically when exclusive games for X device come out to play.
IMO they'd be better off releasing disc versions of games for the console and cartridges for the handheld. It really boils down to how clear their message is, how the games work and how their marketing handles the promo.
For some reason, some people have come to the conclusion that merging their development teams instantly means cross-buy//pay one get two kind of deal when in reality (and what we know so far) is merely to make it easier for them to create and manage new projects across devices after struggling to support both with a healthy and steady amount of games for years.
They way I see it, Nintendo will offer the same game for both platforms but ultimately it will be the costumer who decides what versions makes the most sense for him/her.
Now, I think however, they won't have any problem offering a digital version for free when you purchase the other one and/or offer X digital version at a heavily discounted price when you purchase the other one for some of their games (if not all of them). After all, that's the whole point of "brother/family of devices" or the infamous "shared library" concept but from that to offer both versions in the same physical medium is quite a stretch imo.
EDIT: In case it comes the wrong way, I'm not opposed to the console using cartridges AT ALL, if they can do it then that's great. My post is more towards sharing a single game/cartridge between both platforms.
On the topic of pricing, it could be mitigated by simplifying their logistics which is one of the biggest money hogs for a company.
Terrell has this excellent write-up about this. It's a very long and dense read. http://www.neogaf.com/forum/showpost.php?p=192728798&postcount=557
Terrell has this excellent write-up about this. It's a very long and dense read. http://www.neogaf.com/forum/showpost.php?p=192728798&postcount=557
Pokemaniac
Member
Nintendo typically does a pretty good job of distinguishing which systems carts are compatible with, often using physical properties of the carts to prevent insertion on incompatible platforms. They could do fine with a single cart format.
Also, shipping two versions of lots of their games seems a bit wasteful.
Also, free digital copies for the other platform tends to be a bit open to abuse unless they lock it down like Xbone tried to.
On the other hand I imagine a lot of people just buy one or the other, but would purchase the software of the other platform if it was made available on their platform of choice. There's a lot of 3DS games I would've bought if cross play were a thing.
Spyder_Monkey
Banned
What if it's digital only?
they won't destroy their relationship with retailers like that. I like the idea of a digital console personally but I don't think the market is ready for it.
Would be bad. They'd need to give everyone a very good reason for that decision. Right now the internet sucks, and they also need to maintain their relationships with brick and mortar stores.If they plan on removing the optical drive it leaves little options for an alternative to physical storage. That's why everyone came to the conclusion for cartridges.
The funniest thing though is that CDs are starting to feel more like relics of the past because read and write speeds are not fast enough to read all the data on a 25GB-50GB disc, and trying to go faster produces more noise and generates more heat. Hence why they've been relegated to physical storage with installs doing more of the work with assets.
Interesting how things have been turning around with CD being the future. What's old is new again.
What if a cartridge contains the whole handheld version and either the whole console version, or part of it depending on the size, then the rest you download?
- Status