• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

What we know so far about the Nintendo NX with sources

Status
Not open for further replies.
Metroid didn't completely skip Wii U, though. The Metroid Attraction in Nintendo Land offers more hours of content than plenty of games in the series, and is just as good as some entries IMO.
Nintendo land was pretty fun. I wonder if they'll make a sequel (probably not)
It was also super pretty and more of what I wanted out of HD Nintendo than I probably got.
Not sure if it's because of engineers at Nintendo that can only work on so many games (IIRC, they worked on MK8 later which might explain why it looks so amazing)
 
I'm not holding them to any standard. I'm just saying that if they announce the NX as being a platform with a handheld and a console, then I expect all software to work on either. I don't want to worry about buying multiple pieces of hardware. My rule for this is that either it's 100% or nothing, because the moment they start separating the platforms, then I feel like the whole thing is busted. The moment games won't work on either hand held or console, then they start introducing that doubt of whether or not the game will be released on either. It's like watching the Nintendo Directs and tuning out when a game for a platform you don't have gets announced. If it's all unified, then all games are coming to whatever flavor if NX you have and that's what makes the concept compelling.

Again, this is my point of view. Maybe people don't agree, but I want the ability to play the next Fire Emblem on the TV or on the go. This is compelling to me.
As someone who wants to only go portable and would be happy to never see a Nintendo home console ever again I 100% agree with this.

I'm pretty confident all games will work on both, but its still up in the air sadly. All I can do is hope.

If Nintendo is stupid enough tobsplit them...then while I will sill get the handheld, I may sadly miss out on the home system. I can't justify getting another home system anymore. The thought of being forced to miss a Nintendo system pains me but I just don't think playing home systems is worth it for much anymore.

I just don't want to be forced on the tv as someone who usually plays 15-30 minutes and then stops for awhile.
 

z0m3le

Banned
That's not the SoC alone. You cannot dissipate 4W from the surface of your average SoC without a sizable spreader. It's out of the question.
Quote me in 1 place where I said the SoC should be 4 watts. I was pointing to the entire device. Do you even read my posts?
Didn't you say 1.3W for the CPU block or did I misunderstand you there?
Yes, 2 A72 cores @ 1.5ghz + 4 A53 cores @ 1.5ghz would run 1.3watts for just the CPU. The GPU using only 2CUs clocked at 500mhz, polaris should fit under 1watt for that, given mullins was consuming 3.6watts for 4 jaguar cores (max turbo at 2.2ghz but base seems to be 1.4ghz which is what we should look at) and 2 CU GCN cores @ 500mhz.

Define 'few cores'.

Ok, let's take your hypothetical scenario - let's assume there are 2x A72 in the hh, of single-core performance within 30-ish percent from the console cores. What percentage of the home titles lineup do you expect to run fine on 2 cores, given the console will offer at least 4 to apps? Would that percentage justify the lower net CPU performance of the hh?
Considering that those 2 A72 cores at 1.5ghz would actually be faster than PS4's jaguar cores at 1.6ghz, why would you want to put out such a low 70% estimate, these threads would perform at or above the main threads devs would use in PS4. I also give the handheld 4 A53 cores at 1.5ghz, maybe they put the os on just a single core like PS4, this gives developers 3 more cores for run off. Many japanese developers would be very happy with these 2 threads being just as fast because main logic can get messy, having threads they know will do whatever they throw at 2 PS4 threads, gives developers room to target their most demanding functions at these threads, helping the game run smoothly and giving developers 3 more slower cores to work with that are about as good as espresso, gives them plenty of room for lesser functions. Trust me, as a developer, lesser, faster cores are more useful and much easier to code for. Why do you think it took so long to develop multithreaded applications?

No. If power consumption is the same and TDP is within expected margins (< 1W), cumulatively 4x A53 will be faster than 2x A72. That's the whole justification for the existence of big.LITTLE. Otherwise A72 runs fine at A53's upper power envelope - it's the higher efficiency that makes the LITTLE cores useful.

It isn't about being faster as a whole, it is about developers being able to actually run their games. A72 threads that I've outlined would actually out perform jaguar cores at 1.6ghz. Developers would be able to give these cores their main threads and use the other A53 cores to fill out their CPU performance.

Look if you had 7 cores that all run 60% of those jaguar cores, you'd have more processing power than 2 cores that do 110% and 3 cores that do 60%, but you need to do more work in 99% of the games you made with those 7 slower cores than in the 5 mixed cores. Developers would be able to target the handheld with the same level of complexity that they do the console, except for super multithreaded games, which might like the slight increase of having 7 slower cores. Nintendo is a Japanese company first, and when they are looking at what to do, this makes the most sense. I'm not entirely convinced that they couldn't fit 4 A72 cores and 4 A53 cores, but I'm trying to be reasonable with my expectations.

I started posting in here again because I wanted to point out that the handheld is going to need more performance in the CPU than the console will comparatively, and that is simply because CPU doesn't really scale with resolution, so cut down the GPU and give more CPU performance.
 

Josh5890

Member
Metroid didn't completely skip Wii U, though. The Metroid Attraction in Nintendo Land offers more hours of content than plenty of games in the series, and is just as good as some entries IMO.

That's a bit of a reach bro. I agree though that the Metroid attraction was a blast to play (no pun intended)
 
I think Metroid skipped Wii U because it was an expensive game to make well in HD for a niche audience on a system with an install base lower than some destination islands.
 
I think Metroid skipped Wii U because it was an expensive game to make well in HD for a niche audience on a system with an install base lower than some destination islands.
Yeah, that's pretty much the case. Zelda is still pretty popular so that'll sell. Nintendo is willing to invest more on titles like that and MK8, or at least they've said as much.
I do wonder if the NX would give a lot of these franchises a chance to come back.
Probably not on the console exclusively, but a smaller budget to fit the handheld's limitations. It should allow for more experimentation and for risks to not be as big as they were on Wii U.
 

Sterok

Member
I think Metroid skipped Wii U because it was an expensive game to make well in HD for a niche audience on a system with an install base lower than some destination islands.

Didn't stop W101 or Xenoblade X (yeah they were started before the Wii U was released, but they were still given the time and budget needed even after it was obvious the system was a bomb). Metroid's real issue is that there's seemingly no team that wants to make one, otherwise we probably would've gotten a regular entry on the 3DS by now.
 

KingSnake

The Birthday Skeleton
This whole theory with different versions of the games for handheld and console that are needed to make the console version with great graphics and not being held back by the handheld's lack of power ignores a very simple fact: there is a very high chance that hh sells much more units than the console and thus the hh games will get priority anyhow. And in this case a full library with some compromises on console is much better than few great games and droughts.
 
2D Metroid by Retro as NX launch title. I'm calling it now.

I'm gonna have to get to some replies tomorrow. Maybe we'll get a leak before then and I won't have to.

Yeah, right.
 

Hiltz

Member
I think the reality is Nintendo doesn't always, and should not be expected to release games based on every IP it has under its belt. Developers get presented with different opportunities to do sequels, spin-offs, collaborations and new IP in general. Retro Studios chose not to do Metroid because it had left over ideas that justified why it wanted to do an immediate sequel to DKC Returns. The first Metroid Prime game was initially only meant to be a one-time thing, but it ended up becoming a trilogy by the end of the first game's development. As much as we'd like to see Metroid 2D or Metroid Prime 4, there could always be an opportunity to experiment with the Metroid IP in a way that may not result in a traditional Metroid title. Kensuke Tanabe took the opportunity to do Metroid Prime: Federation Force because it was a new idea that used something familiar but introduced different concept for it to be based around. After all, Nintendo likes the idea of trying to offer something we didn't know we wanted. Of course, when you have things like Sticker Star, Super Paper Mario, Animal Crossing Amiibo Festival and Metroid: Other M, it isn't always met with praise from fans and gaming journalists.
 
I'm curious what people think about Super Smash Bros. for Wii U/3DS - if there are any hints that could be found in its development which may reflect how Nintendo will develop games for NX handheld and console.

Perhaps they were testing the waters of releasing the same game on two very different consoles without sacrificing the more powerful?

The 3DS arguably didn't hold back the Wii U version graphically, only possibly in features and certain characters (such as Ice Climbers). I don't see the NX console being held back by the mobile platform at all.
 
Didn't stop W101 or Xenoblade X (yeah they were started before the Wii U was released, but they were still given the time and budget needed even after it was obvious the system was a bomb). Metroid's real issue is that there's seemingly no team that wants to make one, otherwise we probably would've gotten a regular entry on the 3DS by now.

It probably would have been more expensive to cancel those games than to release them.
 

gulma1

Neo Member
On the topic of Retro, I want to see a 3D Donkey Kong game similar to DK64 from them.

As for Smash, I doubt they were testing. I think it had to do with Wii U sales and they wanted to have a portable smash game
 

10k

Banned
I think quite a few people in this thread are seriously overestimating the extent to which home titles will be able to run on the hh with a mere visuals downgrade - there won't be such a thing for anything outside of low-CPU-usage games like Mario Tennis and the likes. The only scenario where this would be feasible is if nintendo gimped the home unit significantly. That won't happen - it makes zero sense. HH titles running on the home unit - yes. Down-ports with decreased game logic complexity - yes. Full-blown home titles running on the hh with a mere IQ downgrade - no. And this is before considering factors like GPGPU which may or may not scale with res at all.
This is what I've been saying the last few days but because it comes from you it'll be read more lol. I'm a nobody.
 

KingSnake

The Birthday Skeleton
So instead of console downports for hh, there will be hh upports for console, because the hh will sell much more. Hopefully at least without extra charges.
 

Hiltz

Member
On the topic of Retro, I want to see a 3D Donkey Kong game similar to DK64 from them.

As for Smash, I doubt they were testing. I think it had to do with Wii U sales and they wanted to have a portable smash game

Nintendo's Kensuke Tanabe commented on considering the possibility of doing a 3D DK title from back in 2014:
Maybe the Donkey Kong series over time will also feature both 2D and 3D like as in the Super Mario series! The style of the next Donkey Kong Country title would also depend on the game design.
 

ReyVGM

Member
I'm not holding them to any standard. I'm just saying that if they announce the NX as being a platform with a handheld and a console, then I expect all software to work on either. I don't want to worry about buying multiple pieces of hardware. My rule for this is that either it's 100% or nothing, because the moment they start separating the platforms, then I feel like the whole thing is busted. The moment games won't work on either hand held or console, then they start introducing that doubt of whether or not the game will be released on either. It's like watching the Nintendo Directs and tuning out when a game for a platform you don't have gets announced. If it's all unified, then all games are coming to whatever flavor if NX you have and that's what makes the concept compelling.

Again, this is my point of view. Maybe people don't agree, but I want the ability to play the next Fire Emblem on the TV or on the go. This is compelling to me.

Prepare to be disappointed because not all games will be available for both NX console and handheld. Do not expect Zelda U/NX on the portable NX. Do not expect the NX console Final Fantasy 15 to run on the portable version. Do expect the likes of Mario Kart, Smash, etc. though.
 

Malus

Member
This is what I've been saying the last few days but because it comes from you it'll be read more lol. I'm a nobody.

Don't be so hard on yourself. Everyone here is a nobody, save for Ideaman. We all wait patiently for him to return and shine his holy wisdom upon us.

Nintendo's Kensuke Tanabe commented on considering the possibility of doing a 3D DK title from back in 2014:

The only Retro games I've played are Prime 3 and DKCTF, and I'd honestly much rather have a new 3D DK platformer than Metroid.
 
So instead of console downports for hh, there will be hh upports for console, because the hh will sell much more. Hopefully at least without extra charges.
I think we could very well get console down ports down the line like what happened to Xenoblade on the new 3DS. Maybe a few revisions down the line
But yeah, I'm expecting most titles to be built with the handheld in mind. I think Nintendo will try and put extra effort for the console ports tho
 
Some of the hardware predictions in here are just silly. 14nm Polaris chip for the handheld? You are going to get a 28nm battery-hungry mullins chip and you are going to like it! lol
 
Prepare to be disappointed because not all games will be available for both NX console and handheld. Do not expect Zelda U/NX on the portable NX. Do not expect the NX console Final Fantasy 15 to run on the portable version. Do expect the likes of Mario Kart, Smash, etc. though.
I fully expect Zelda U/NX on the portable. Absolutely no reason not to just because some people on the internet are like "but but its a home console style game" Can't wait for the meltdowns.

3rd parties are a toss up. A company like Square would push for handheld since they are more popular in Japan. (Of course FFXV might be too much for it)

It also depends on if Nintendo mandates both versions. I think they will, since it would just be dumb retail confusion otherwise on only some NX games working on both but we'll see. Just wish it was announced so my head can stop hurting from some of the suggestions in this thread. :p
 

z0m3le

Banned
I think quite a few people in this thread are seriously overestimating the extent to which home titles will be able to run on the hh with a mere visuals downgrade - there won't be such a thing for anything outside of low-CPU-usage games like Mario Tennis and the likes. The only scenario where this would be feasible is if nintendo gimped the home unit significantly. That won't happen - it makes zero sense. HH titles running on the home unit - yes. Down-ports with decreased game logic complexity - yes. Full-blown home titles running on the hh with a mere IQ downgrade - no. And this is before considering factors like GPGPU which may or may not scale with res at all.

Ridiculous, Jaguar cores perform about on par with A57 cores clock for clock, there is a 15% increase in performance clock for clock on A72 vs A57, so yes a 1.4ghz A72 core would perform slightly better than PS4's cores at the same clock. You can fit 4 of these at 1.4ghz at only 1.1watts, you add 4 A53 cores @ 1.2Ghz at only 0.5watts, 8 cpu cores, 4 of which are on par with PS4 cores and you can do parity with consoles with only 1.6watt CPU power consumption. It all depends on polaris here, but if they can get 128gflops out of that and still have the entire CPU + GPU at 2.5watts, they would be fine in that power envelope.

This isn't what I expect, it is simply what Nintendo should do.

EDIT: I'll also add that the n3DS has a much faster CPU than the original, a very similar thing can happen with the console and handheld.
 

Roo

Member
I fully expect Zelda U/NX on the portable. Absolutely no reason not to just because some people on the internet are like "but but its a home console style game" Can't wait for the meltdowns.

3rd parties are a toss up. A company like Square would push for handheld since they are more popular in Japan. (Of course FFXV might be too much for it)

It also depends on if Nintendo mandates both versions. I think they will, since it would just be dumb retail confusion otherwise on only some NX games working on both but we'll see. Just wish it was announced so my head can stop hurting from some of the suggestions in this thread. :p
No way they force/mandate a version for both consoles.

Companies would give Nintendo the middle finger if forced before supporting a console that doesn't meet their vision/criteria/business or whatever lame excuse you want to use.
 

Hydrus

Member
Just went to my local best buy, and the Wii U section was barren! They had maybe 6 or 7 different games and that was about it. So yea, the Wii U era is definitely winding down quick. Cant believe the die-hard's seriously believed a new console wasn't coming until "holiday 2017 at the earliest!"
 

Peru

Member
Just went to my local best buy, and the Wii U section was barren! They had maybe 6 or 7 different games and that was about it. So yea, the Wii U era is definitely winding down quick. Cant believe the die-hard's seriously believed a new console wasn't coming until "holiday 2017 at the earliest!"

I definitely don't think a new stationary console is coming before 2017.
 
So instead of console downports for hh, there will be hh upports for console, because the hh will sell much more. Hopefully at least without extra charges.
It doesn't matter which system sells "much" more, what matters is how many games are sold, since that's where most of Nintendo's profits come from. The 3DS has sold more hardware than the SNES for example, and is still significantly behind the SNES on software. Wii sold 50 million less systems than DS, and yet they sold roughly the same amount of software (both had heavy bundling of course). Mario Kart 8 (12 million install base) was able to outsell Double Dash (21 million install base) and SC (81 million install base) too. Of course, the Handheld will have the advantages of having a larger install base tovsell games to, but that doesn't mean Nintendo should just give up on their consoles.
 

The Giant

Banned
I'll rather have Retro make a new Ip (which was the last rumour I read that they were doing a new ip) rather than metroid. Retro's time with metroid is over. People need the figure that out.
 
I'm curious what people think about Super Smash Bros. for Wii U/3DS - if there are any hints that could be found in its development which may reflect how Nintendo will develop games for NX handheld and console.

Perhaps they were testing the waters of releasing the same game on two very different consoles without sacrificing the more powerful?

The 3DS arguably didn't hold back the Wii U version graphically, only possibly in features and certain characters (such as Ice Climbers). I don't see the NX console being held back by the mobile platform at all.

Graphically, I believe Sakurai said that the models were designed first for Wii U as the base, and then scaled down and optimized later for the 3DS. Though I'm sure that's probably standard in development, that it's easier to downsize and make less out of more than the other way around.
 

Astral Dog

Member
I think Metroid skipped Wii U because it was an expensive game to make well in HD for a niche audience on a system with an install base lower than some destination islands.

I think the success of DKCR, the constant milking of the Metroid Prime series (with decent but not Donkey Knog sales), Retros desire to do something a bit different and the total failure of Nintendos own Other M had more to do with Metroid needing a rest and not being on Wii U than the system itself.

They could have done a handheld game to compensate, but if the best they could come up with was Federation Force after all these years...
 

Roo

Member
Graphically, I believe Sakurai said that the models were designed first for Wii U as the base, and then scaled down and optimized later for the 3DS. Though I'm sure that's probably standard in development, that it's easier to downsize and make less out of more than the other way around.
That's correct.
 

AlStrong

Member
Was this posted? I have no idea what's going on with the arm spec comparisons these days:
http://www.anandtech.com/show/9878/the-huawei-mate-8-review/3
power-big.png
 

mrklaw

MrArseFace
It doesn't matter which system sells "much" more, what matters is how many games are sold, since that's where most of Nintendo's profits come from. The 3DS has sold more hardware than the SNES for example, and is still significantly behind the SNES on software. Wii sold 50 million less systems than DS, and yet they sold roughly the same amount of software (both had heavy bundling of course). Mario Kart 8 (12 million install base) was able to outsell Double Dash (21 million install base) and SC (81 million install base) too. Of course, the Handheld will have the advantages of having a larger install base tovsell games to, but that doesn't mean Nintendo should just give up on their consoles.


But if the handheld was selling disproportionately more globally, and you have a common architecture designed from the outset to ease development across both platforms - why would you invest in home console exclusive content? Logically you'd trim the scope in order to sell it on the much larger user base of handhelds, and also the home console.
 

Malus

Member
But if the handheld was selling disproportionately more globally, and you have a common architecture designed from the outset to ease development across both platforms - why would you invest in home console exclusive content? Logically you'd trim the scope in order to sell it on the much larger user base of handhelds, and also the home console.

The home console version of NX would lose a lot of its appeal if Nintendo decided to trim everything down for the handheld. Why not make a handful ambitous home console titles and try to appeal to that crowd. I imagine there are some devs at Nintendo that want to make something with a big scope just for the hell of it anyways *coughmonolithsoftcough.*

If it really got the the point where they decided it's not worth it to make games that take advantage of the home console because of declining sales, why not just scrap the whole thing? Iwata did mention that maybe the market would decide that Nintendo should narrow things down to 1 form factor.

HELLLOOOOOOO NINTENDO..???!!!!

ARE YOU THERE?

They can't hear you. They can't see you.
 
Just went to my local best buy, and the Wii U section was barren! They had maybe 6 or 7 different games and that was about it. So yea, the Wii U era is definitely winding down quick. Cant believe the die-hard's seriously believed a new console wasn't coming until "holiday 2017 at the earliest!"

Nintendo has released very few games on Wii U since it came out, I wouldn't put it past them to let the Wii U linger until holiday 2017.
 

blu

Wants the largest console games publisher to avoid Nintendo's platforms.
Quote me in 1 place where I said the SoC should be 4 watts. I was pointing to the entire device.
In a discussion about SoCs you were pointing the TDP of the entire vita. Gotcha.

Do you even read my posts?
Not only that but I even quote the parts I respond to.

Yes, 2 A72 cores @ 1.5ghz + 4 A53 cores @ 1.5ghz would run 1.3watts for just the CPU.
1.3W is more than half of 2.5W for the SoC. Which is what I said.

The GPU using only 2CUs clocked at 500mhz, polaris should fit under 1watt for that, given mullins was consuming 3.6watts for 4 jaguar cores (max turbo at 2.2ghz but base seems to be 1.4ghz which is what we should look at) and 2 CU GCN cores @ 500mhz.
Citation needed. The best-in-class 4-core A10 Micro-7600T is 4.5W part

Considering that those 2 A72 cores at 1.5ghz would actually be faster than PS4's jaguar cores at 1.6ghz, why would you want to put out such a low 70% estimate, these threads would perform at or above the main threads devs would use in PS4.
I was comparing to your home NX estimates, which will be the main source of down-ports for the hh.

I also give the handheld 4 A53 cores at 1.5ghz, maybe they put the os on just a single core like PS4, this gives developers 3 more cores for run off. Many japanese developers would be very happy with these 2 threads being just as fast because main logic can get messy, having threads they know will do whatever they throw at 2 PS4 threads, gives developers room to target their most demanding functions at these threads, helping the game run smoothly and giving developers 3 more slower cores to work with that are about as good as espresso, gives them plenty of room for lesser functions. Trust me, as a developer, lesser, faster cores are more useful and much easier to code for. Why do you think it took so long to develop multithreaded applications?
Trust me, as a former game developer, cumulative performance is more important a good deal of the time.

It isn't about being faster as a whole, it is about developers being able to actually run their games. A72 threads that I've outlined would actually out perform jaguar cores at 1.6ghz. Developers would be able to give these cores their main threads and use the other A53 cores to fill out their CPU performance.
Their main threads? What is this, the 90's? Developers have been complaining about the cumulative performance of the 6-7 cores the ps4/xbone have been providing them, and your're arguing about the feasibility of 2 cores or somewhat higher performance? Really?

Look if you had 7 cores that all run 60% of those jaguar cores, you'd have more processing power than 2 cores that do 110% and 3 cores that do 60%, but you need to do more work in 99% of the games you made with those 7 slower cores than in the 5 mixed cores. Developers would be able to target the handheld with the same level of complexity that they do the console, except for super multithreaded games, which might like the slight increase of having 7 slower cores. Nintendo is a Japanese company first, and when they are looking at what to do, this makes the most sense. I'm not entirely convinced that they couldn't fit 4 A72 cores and 4 A53 cores, but I'm trying to be reasonable with my expectations.

I started posting in here again because I wanted to point out that the handheld is going to need more performance in the CPU than the console will comparatively, and that is simply because CPU doesn't really scale with resolution, so cut down the GPU and give more CPU performance.
You're arguing for lost-cause setup where the hh has on-par CPU performance with he home units (ps4, xbone. nx home) - that ain't happening.

Ridiculous, Jaguar cores perform about on par with A57 cores clock for clock, there is a 15% increase in performance clock for clock on A72 vs A57, so yes a 1.5ghz A72 core would perform slightly better than PS4's cores at the same clock. You can fit 4 of these at 1.5ghz at only 1.2watts, you add 4 A53 cores @ 1.2Ghz at only 0.5watts, 8 cpu cores, 4 of which are on par with PS4 cores and you can do parity with consoles with only 1.7watt CPU power consumption. It all depends on polaris here, but if they can get 128gflops out of that and still have the entire CPU + GPU at 2.5watts, they would be fine in that power envelope.
'Only' 1.7W of CPU consumption? See?
 
But if the handheld was selling disproportionately more globally, and you have a common architecture designed from the outset to ease development across both platforms - why would you invest in home console exclusive content? Logically you'd trim the scope in order to sell it on the much larger user base of handhelds, and also the home console.
They should at least make an attempt at being successful again in the home console front. I find it hard to believe that Nintendo, with a rather large war chest, will just give up on the home console front like that, that is essentially what a 100% shared library will lead up to. Even Miyamoto hopes that the system after Wii U will be a bigger hit. If their next console plays the same games as the handheld, it will be the death of the Nintendo home console.

They can't hear you. They can't see you.
Oh they can probably hear and see us, they just wanna make it seem like they don't to mess with us!
 

z0m3le

Banned
In a discussion about SoCs you were pointing the TDP of the entire vita. Gotcha.


Not only that but I even quote the parts I respond to.


1.3W is more than half of 2.5W for the SoC. Which is what I said.


Citation needed. The best-in-class 4-core A10 Micro-7600T is 4.5W part


I was comparing to your home NX estimates, which will be the main source of down-ports for the hh.


Trust me, as a former game developer, cumulative performance is more important a good deal of the time.


Their main threads? What is this, the 90's? Developers have been complaining about the cumulative performance of the 6-7 cores the ps4/xbone have been providing them, and your're arguing about the feasibility of 2 cores or somewhat higher performance? Really?


You're arguing for lost-cause setup where the hh has on-par CPU performance with he home units (ps4, xbone. nx home) - that ain't happening.


'Only' 1.7W of CPU consumption? See?

The CPU should be where most resources are spent with my approach, you attack me with my 1.7w CPU design, but I see you are avoiding the 12 A53 cores clocked at 1ghz with 4CUs clocked at 400mhz, which should produce very similar SoC power consumption. Unless you honestly think polaris 4CUs clocked at 400mhz will draw drastically less than 1.5w.

http://cpuboss.com/cpus/Intel-Atom-Z3735F-vs-AMD-A10-Micro-6700T is the 3.66watts for the 6700t apu power consumption, yes at the lower typical clocks and not the turbo 2.2ghz boost that fits in the 4.5 watt thermal design. The gains from both polaris and A72 as well as trimming down to 14nm LPP should bring heavy reductions to the power consumption here. I'm not even sure what you are arguing when a quad core X86 chip based on jaguar's design with a GCN based 2CU GPU clocked at 500mhz can achieve such a low result on a process node 5 years old.

You believe the SoC limit would be 2.8watts assuming this is what you meant by Vita not pushing more than this, so my ~2.5watt SoC is impossible because?

I also over calculated Jaguar vs A72, you can get away with A72 cores clocked at 1.4ghz which for a quad core, would only be 1.1watts, adding 4 A53 cores at 1.2ghz would cost 0.5 watts, like I've said many times, it is up to polaris, but this CPU would handle console development with minimal rework as it is unlikely that most japanese developers will be pushing all 7 PS4 cores to their max, how is it even possible that XB1 games don't run like crap on PS4 thanks to the 1.75ghz clock of those cores is beyond me. ;)

As for you drawing from experience with your posts, that is great and all, but if you were working on frostbite, you might not be the Japanese developer Nintendo would be targeting with this approach. Nintendo's own internal teams would also benefit from this design and it isn't like I'm Nintendo, I can have an opinion about maximizing single thread vs throwing more threads at the tasks. At least the thread isn't talking about how impossible 14nm would be for Nintendo to use anymore, I swear if Nintendo announces big.LITTLE for NX this thread will reason it as the obvious choice. In the end I'm just giving my opinion on what they should do, correct me all you want, but if your solution is to release 12 A53 cores clocked at 1ghz instead as with Tharoktor's post, then good luck there, the headache of not having any performance for any threads is going to make it far worse...
 

The Giant

Banned
I sense that a direct is coming this week. All because this weeks pal eshop is barebones and normally when that happens, a direct happens.
 

Turrican3

Member
I think quite a few people in this thread are seriously overestimating the extent to which home titles will be able to run on the hh with a mere visuals downgrade - there won't be such a thing for anything outside of low-CPU-usage games like Mario Tennis and the likes. The only scenario where this would be feasible is if nintendo gimped the home unit significantly. That won't happen - it makes zero sense.
I guess it depends on what kind of strategy Nintendo has planned.
If they plan on getting decent support from western publisher of course at least XB1-level hardware is a must for the home. And so everything you say obviously seems reasonable.

But I wonder if we can really exclude they are targeting something totally different (for example, heavily depending on some gimmick), so that the handheld and the home end up having closer specification than we might expect.

If that's the case, full-blown home titles might basically cease to exist, and I guess everything would be tailored to the handheld specification (and would of course run the same way on the home, just at higher resolution).
 

z0m3le

Banned
I guess it depends on what kind of strategy Nintendo has planned.
If they plan on getting decent support from western publisher of course at least XB1-level hardware is a must for the home. And so everything you say obviously seems reasonable.

But I wonder if we can really exclude they are targeting something totally different (for example, heavily depending on some gimmick), so that the handheld and the home end up having closer specification than we might expect.

If that's the case, full-blown home titles might basically cease to exist, and I guess everything would be tailored to the handheld specification (and would of course run the same way on the home, just at higher resolution).

Yes, this is basically what I think Nintendo SHOULD do, I don't really expect them to, but it can be done.

Since I did all the work to prove that a 2.5w SoC can exist for the handheld that is 4 1.4ghz A72 cores (on par with PS4 cores) + 4 A53 cores clocked at 1.2ghz and 2 CUs clocked at 500mhz based on polaris for 128GFLOPs.

The console could use the same CPU with a higher clock on the A72 cores and 1.5ghz on the A53 cores, similar approach here to n3DS vs 3DS. The GPU could then be 8CUs clocked at 1GHz, if this is polaris and 14nm, these clocks should be reasonable and would give the same features as the handheld, while offering similar performance to the XB1 thanks to gains from Polaris over GCN, also that CPU should be on par with PS4 on the home console while being close enough on the handheld to give minimal effort to downporting PS4/NX games. Some people have argued that Polaris and 14nm might not happen in the console. I don't think it really needs to, to achieve this relationship with the handheld and these numbers, but if you were one team working on the next Nintendo devices, you'd probably focus on the handheld and use those same technologies on the console, as you can design once and have much more exact performance estimations since GCN and Polaris are not identical, with various upgrades to different components, such as the tessellation unit.
 

Rodin

Member
On the topic of Retro, I want to see a 3D Donkey Kong game similar to DK64 from them.

As for Smash, I doubt they were testing. I think it had to do with Wii U sales and they wanted to have a portable smash game
Smash 3DS was announced before the Wii U even came out, i think it was E3 2011.

As for Retro, if they'll ever make a 3D platform/adventure i want it to be Duck Hunt.

Around GDC we could see some "proper" leaks I guess.
That's in march right?
 

beril

Member
The CPU should be where most resources are spent with my approach, you attack me with my 1.7w CPU design, but I see you are avoiding the 12 A53 cores clocked at 1ghz with 4CUs clocked at 400mhz, which should produce very similar SoC power consumption. Unless you honestly think polaris 4CUs clocked at 400mhz will draw drastically less than 1.5w.

That's a terrible idea for a game console, and very unlike Nintendo especially.
I think you're vastly overrating the importance of the CPU for games in general. Sure there are a few outliers complaining about CPU performance on PS4/Xbox one, but for the vast majority of games the GPU is the bottleneck.

And while blu is right that it won't be possible to run demanding PS4 games on a handheld with just a graphical downgrade, I don't think it's necessarily easier to do a massive graphical downgrade than it is to free up CPU cycles, and a lot of the time it goes hand in hand.
 
At some point don't Wii U owners deserve to know what is next? We bought a Wii U and see the release schedule slowing down to a complete stop, without knowing what to expect. All I can say is, I hope Nintendo has something planned to create new Nintendo fans, because the last years of the Wii and the Wii U as a whole did a great job to narrow their consumers down to the hardcorest of hardcore fans.
 

z0m3le

Banned
That's a terrible idea for a game console, and very unlike Nintendo especially.
I think you're vastly overrating the importance of the CPU for games in general. Sure there are a few outliers complaining about CPU performance on PS4/Xbox one, but for the vast majority of games the GPU is the bottleneck.

And while blu is right that it won't be possible to run demanding PS4 games on a handheld with just a graphical downgrade, I don't think it's necessarily easier to do a massive graphical downgrade than it is to free up CPU cycles, and a lot of the time it goes hand in hand.

I'm not talking about the Console, I'm talking about a Handheld with a shared library with the Console. As you know, CPU performance doesn't really scale with resolution, so while you can produce similar performance across 1080p and 540p with 2 GPUs, one being 1/4th the performance or even less. The CPU needs to remain similar between these devices.

In fact everything you said, strengthens the idea that you could produce a handheld that can handle most console games, with mostly graphical downgrades and few concessions to game logic / AI. As you say, most PS4/XB1 developers are not maxing out the CPU, this is especially true of the Japanese developers, who are not so heavy into 7 thread cpu developments.

Your second point, where graphical downgrades would lend itself to freeing up CPU cycles, is also pointing to this approach being workable. A72 cores at 1.4ghz matching PS4 cores at 1.6ghz is the main reason this is even a reasonable suggestion, if they aren't maxing out the use of those extra 3 cores, a lot of games can probably work using 3 slower A53 cores to pick up that slack, especially if while you are downgrading the graphics, you are freeing up CPU cycles.
 
Status
Not open for further replies.
Top Bottom