Support NeoGAF

[lwilliams3]

What kind of comments do we have from devs by the way?

Outside of what insiders have reported, all I can recall that they seem to be optimistic in general. They can't say much of anything due to the NDA

I take your point but looking at what we have to go on - MK8, Splatoon and the new 3D Mario footage, do they look like they're a quantum leap over the WiiU versions ?

I suppose there is also the chance that this could very well be the first of many devices in the future so the reason Nintendo went with a custom, brand new architecture is maybe for the future, much more powerful devices to give them lots of options with regards to cores, clocks etc.

Honestly, a console which runs better looking WiiU games on the go with good battery life will be a big selling point to a lot of people, esp those who will be upgrading their 3DS to Switch and using it mostly as a handheld device.

They can always release a much more powerful dedicated standalone console sometime in 2019 based on the Switch architecture.

NVIDIA and Nintendo are definitely planning things for the long term, but I wouldn't expect to hear from them anytime soon.

I think people need to taper expectations to between Wii U and X1 at best. People are going to be severely disappointed otherwise. Even if the CPU is stronger than the ps4 / x1 the GPU isnt going to eclipse what is in those systems and the chips wont be nearly clocked as high. So it seems like a fairly moot point when put into perspective.

Honestly, I don't know who you are referring to. Afaik, everyone expectations appear to be reasonable and are within the range you stated. Well, outside of that 4K BOTW question, but no one agreed with it.

Btw, the CPU and GPU in the XB1 and PS4 are actually not "clocked" that high. The chipset in the Tegra TV is even clocked higher, but that shouldn't be compared due to the different architecture. The 256-core GPU architecture that will likely be in the Switch would have to be clocked well beyond 2GHz to reach the power of PS4's 800 MHz GPU.

 

[MacTag]

Nintendo doesn't have access of unique LPDDR.

So that's kind of impossible.

Embedded SRAM, 128-bit bus, 3200 LPPDR4 x4, even crazy HBM2 dreams. And that's even before we get into things like tile based rasterization, inherent bandwidth efficiencies for A57/A72/Denver over Jaguar, Pascal over Liverpool/Durango, etc.

The reality is Switch is unlikely to be as bandwidth constrained as you're so desperately hoping for and Nintendo has a long history of customization, embedded solutions and high spending in this particular area. There's more to architecture efficiency and bandwidth than one number on your RAM and Wii U is a pretty perfect example of that.

 

[Gotdatmoney]

Some previous posts are insinuating it could be or alluding to possible performance that just isn't likely with limitations of battery power.

Where has this been said? Seriously, I have mostly been reading this thread because I have nothing tech wise to contribute but I have no read anyone even kind of alluding out performing X1 not even mentioning getting to PS4 level. Who was saying this?

The system is in practice going to be weaker than the X1. This is a portable, you aren't getting console performance. I haven't seen anyone saying greater than X1 is on the table.

 

[MuchoMalo]

Who's arguing ?Well everyone who's saying it's damn impossible that the Switch is "just" WiiU power and is at minimum 2 or 3 times more powerful.

I don't get your second point. And if the kit is actively cooled, and bigger than the real thing.. wouldn't that mean they can't have the active cooling in the Switch and it has to run slower ? Tell me if i'm wrong.

And yeah i'm asking what is the battery life of a maxed out shield, annnd with what battery size.

If the kit is actively cooled, what it means is that either the final version is also actively cooled (and we have evidence that this is the case), or the final version is using a newer, more power-efficient chip than the dev kit (which has been rumored by multiple people). You appear to be under the impression that a chip must be customized before it can be underclocked; that is not at all the case. As for battery life, until we know for certain what Nintendo is targeting in terms of battery life, this is a moot point. The Shield Tablet is using an older chip than Switch as well.

You will not find a post by me claiming the Switch will be so powerful as the WiiU or Vita, so that's all rather pointless.

But when I talked about the problems of getting fast enough ram I onlyl get "It's Nintendo" and "there are options!" posts.

I wasn't around for any conversation on memory bandwidth, but the answer to that one is pretty simple: on paper, it'll be behind Xbone. The memory controllers for the Tegra X1 and Parker are designed for LPDDR4, so that's what Nintendo will use. Putting in a custom controller to use LPDDR3 wouldn't make sense because the cost difference between LPDDR3 and LPDDR4 isn't that big, and LPDDR4 prices will drop more over time. For bandwidth, the choices are 25GB/s on a 128-bit bus or 50GB/s on a 256-bit bus. Nvidia GPUs tend to utilize bandwidth more efficiently, so the latter option would effectively be relatively close to Xbone. There's also the option of eSRAM on the chip, though that might increase costs a bit too much depending on how much they'd have to use. That said, it's important to keep in mind that we're looking at a 720p machine with less RAM here.

Any more questions?

 

[Panajev2001a]

Shield tablet uses a older tetra chip as well, older and less power efficient than X1. Switch will likely use something newer and more efficient than X1, far more power efficient for both GPU and CPU than the K1 chip used in Shield. No way is a WiiU in Switches form factor any kind of achievement today.

The K1 mini tablet is still more than £249 without controller and there are not many games that really push it, the review I found (http://www.techradar.com/reviews/pc-mac/tablets/nvidia-shield-tablet-k1-1318312/review/2) is quite recent, and 5 hours battery life while gaming is a peak if not, as stated, optimistic figure. I do not think Nintendo is asking for simpler Android games optimised to run even for iPad 2 level HW.

 

[MuchoMalo]

The K1 mini tablet is still more than £249 without controller and there are not many games that really push it, the review I found (http://www.techradar.com/reviews/pc-mac/tablets/nvidia-shield-tablet-k1-1318312/review/2) is quite recent, and 5 hours battery life while gaming is a peak if not, as stated, optimistic figure. I do not think Nintendo is asking for simpler Android games optimised to run even for iPad 2 level HW.

The Tablet K1 is over priced/being sold at a high margin, so I wouldn't use that as a reference point.

 

[Donnie]

The K1 mini tablet is still more than £249 without controller and there are not many games that really push it, the review I found (http://www.techradar.com/reviews/pc-mac/tablets/nvidia-shield-tablet-k1-1318312/review/2) is quite recent, and 5 hours battery life while gaming is a peak if not, as stated, optimistic figure. I do not think Nintendo is asking for simpler Android games optimised to run even for iPad 2 level HW.

Like I said its old harder, newer chips are not only far more power efficient but also far more powerful and cheaper. That chip is using old A15 CPUs and Kepler 28nm GPU. Still comfortably more powerful than WiiU but not the most power efficient, yet it still gets 5 hour battery life. Switch will likely be a custom version of Tetra a generation or two ahead of that on a 16nm process and rumours suggest around 3 hours battery life.

 

[PdotMichael]

Embedded SRAM, 128-bit bus, 3200 LPPDR4 x4, even crazy HBM2 dreams. And that's even before we get into things like tile based rasterization, inherent bandwidth efficiencies for A57/A72/Denver over Jaguar, Pascal over Liverpool/Durango, etc.

The reality is Switch is unlikely to be as bandwidth constrained as you're so desperately hoping for and Nintendo has a long history of customization, embedded solutions and high spending in this particular area. There's more to architecture efficiency and bandwidth than one number on your RAM and Wii U is a pretty perfect example of that.

I don't care about crazy dreams.
I want some meaningfull options for a mobile device in the expected price class.

I also don't care that much about insiders, I know one got banned because he claimed that the infamous fake controller was real.

 

[ShadowFox08]

Right ?

Also nobody never talks about the fact that there is 3 batteries in the switch. Both joycon needs a battery to. Nobody is never taking into account the many constraints induced by the whole concept.

i'm just saying this is a possibility, that it's nothing more than a WiiU. Some people are in denial when they're all "SHUT UP YOU FOOL IT'S NOT POSSIBLE, WE KNOW IT! WE KNOW EVERYTHING!!".

If the RAM is really 4GB as rumored, then its pretty much confirmed that that its at least 2-4x as powerful as the Wii U.

Being two Wii u consoles ducktaped together isn't going to go well with many. Especially console core gamers and third party devs. Remember guys that this is a hybrid and not a home console. Some up lock performance when docked could be possible(I'm guessing more of a resolution boost though)

This sudden wave of pessimism is really weird.

 

[lwilliams3]

I don't care about crazy dreams.
I want some meaningfull options for a mobile device in the expected price class.

I also don't care that much about insiders, I know one got banned because he claimed that the infamous fake controller was real.

Except for maybe Matt, who has been careful and vague with all of his contributions, I don't think any of us are claiming that they are 100% infallible.

Having said that, if numerous people are reporting the similar news, that may something to look into. SMD went against the "insiders", and it didn't do him much good.

 

[Gaspard]

Considering the emphasis of combat on this title I would prefer if the Switch version went the framerate route rather than bells and whistles of lighting and anti aliasing.

 

[MuchoMalo]

I don't care about crazy dreams.
I want some meaningfull options for a mobile device in the expected price class.

I also don't care that much about insiders, I know one got banned because he claimed that the infamous fake controller was real.

... Which I literally just gave you.

 

[EDarkness]

I would personally prefer a boost to 1080p resolution with anti aliasing, anisotropic filtering and improved textures at a more consistent 30fps over the same WiiU visuals at 60fps but it would be nice if Nintendo offered the option like some other developers are now doing.

I imagine the game will be 1080p when docked. When not docked most likely will be 720p 30 FPS. I don't know if Nintendo would go 60 FPS, but who knows. Maybe...

 

[Boney]

So they did a full reverse from Wii U's pathetic CPU? Kewl

 

[EDarkness]

You know, guys. I think you're wasting your time trying to convince these guys that the NS will be more capable than the Wii U. If they want to be that cynical, then let them. Obviously they aren't looking at reason and with all the explanations you guys have given them, I just feel like they're trolling at this point. They don't want to accept even the basic facts, then that's on them. The very notion that Nvidia and Nintendo would make a system purposefully gimped is downright absurd. If they want to have that view, then fine.

Besides, it doesn't matter anyway because Nintendo will just be Nintendo. Right?

 

[PdotMichael]

... Which I literally just gave you.

128bit is the only possible solution because that's part of the standard specification, which means that the Switch (even ignoring the lack of ram if we accept the 4GB claim) can't match the performance of the Xbox One with its DDR3 + static ram solution.

 

[MuchoMalo]

128bit is the only possible solution because that's part of the standard specification, which means that the Switch (even ignoring the lack of ram if we accept the 4GB claim) can't match the performance of the Xbox One with its DDR3 + static ram solution.

That's assuming that the final chip is definitely based on the Tegra X1 with little to no changes. Parker (the Pascal Tegra) has a 256-bit bus. If you feel that Pascal is definitely ruled out and that the only customization to the X1 was the removal of unnecessary components, then you are right. That's a big "if" though.

 

[TunaLover]

That's assuming that the final chip is definitely based on the Tegra X1 with little to no changes. Parker (the Pascal Tegra) has a 256-bit bus. If you feel that Pascal is definitely ruled out and that the only customization to the X1 was the removal of unnecessary components, then you are right. That's a big "if" though.

Do you have any thoughs about what customizations would be done to the X1?

 

[Thraktor]

I don't care about crazy dreams.
I want some meaningfull options for a mobile device in the expected price class.

I also don't care that much about insiders, I know one got banned because he claimed that the infamous fake controller was real.

A 128-bit LPDDR4 setup could give them anywhere up to around 60GB/s of bandwidth depending on what clock speed they go with. This is the same bus width they used on the 3DS and would be plenty given the expected performance level and bandwidth efficiency of Pascal, even ignoring the benefits of altering the cache layout to maximise the benefits from tile-based rendering (which is something I'd expect them to do, Nintendo doesn't exactly shy away from large on-die memory pools).

There's really no reason to believe, or any evidence to suggest, that Switch will be in any way bandwidth constrained.

 

[PdotMichael]

That's assuming that the final chip is definitely based on the Tegra X1 with little to no changes. Parker (the Pascal Tegra) has a 256-bit bus. If you feel that Pascal is definitely ruled out and that the only customization to the X1 was the removal of unnecessary components, then you are right. That's a big "if" though.

The Pascal Tegra doesn't use a 256-bit bus.

nVidia claims 50GB/s for Parker. But sure Nintendo can use a supercharged Parker I mean what could stop them except price, battery, heat.

 

[Donnie]

128bit is the only possible solution because that's part of the standard specification, which means that the Switch (even ignoring the lack of ram if we accept the 4GB claim) can't match the performance of the Xbox One with its DDR3 + static ram solution.

Actually 128bit bus could offer the same main memory bandwidth as Xbox One and TBR would give a very similar bandwidth saving to Xbox One embedded memory. So it could certainly be matched or at least extremely close to matched. Then again I don't expect Switch to be 1.4Tflops so why match bandwidth anyway?

 

[MDave]

We know with almost certainty from reputable sources that the dev kits are Tegra X1 clocked to the point it needs active cooling. We know that Nintendo and Nvidia have designed a custom Tegra based SoC for the Switch. It would have to be at least as powerful as the dev kit X1 chip, so it being custom leads us to two paths they could of gone with it:

1: More battery efficient / less heat but the same power as the X1 dev kits at full power.
2: More performance at the same battery / thermal limits of the X1.

Maybe a bit of both: This is all certainly possible with even just a smaller node process to 16nm. Expected for a mobile device launching in 2017.

The 3+ hour battery life makes it sound like they are not making many compromises to performance for portable mode.

 

[MacTag]

I don't care about crazy dreams.
I want some meaningfull options for a mobile device in the expected price class.

I also don't care that much about insiders, I know one got banned because he claimed that the infamous fake controller was real.

HBM2 is out there but it destroys even GDDR5 in speed. Other than that everything I mentioned was realistic (high bus, embedded mem, tile based, APU efficiencies) and nothing to do with insiders. All this corncern trolling over LPPDR4 is really nothing, I'd suggest finding a new line of attack.

 

[mrklaw]

You know, the wording of many sources implies that the system has some considerable power despite its form factor:

ShockingAlberto: around XB1 power

Emily Rodgers: beyond the Wii U, but will not reach XB1 raw power

Laura: Vouches for Emily

NateDrake: Around 3x-4x the Wii U, the full power TX1 is a good estimate.

Matt: Not weaker than Wii U, not as "CPU limited" as the XB1/PS4, porting will not be a portable issue.

OB: Nintendo will be fine. Porting from the other consoles will not be a technical issue.

LCGeek: CPU stronger than XB1/PS4.

Did I miss anyone major?

Compare this to what was said about the Wii U from lherre, Akram, Ideaman, the devs, Epic, insider reports, and etc prior to the release. There was a very big "lower your expectations" flag coming from them (Ideaman tried to sugarcoat it, but it was obvious). I think we can worry a little less about the power of the Switch.

In portable or docked mode? Could be a big difference between the two.

While > WiiU power in portable mode is a huge leap forward for Nintendo and fantastic news for Nintendo handheld fans, it is much less interesting for those who manly want a home console. That's the tightrope that Nintendo have to walk.

 

[LordOfChaos]

A 128-bit LPDDR4 setup could give them anywhere up to around 60GB/s of bandwidth depending on what clock speed they go with. This is the same bus width they used on the 3DS and would be plenty given the expected performance level and bandwidth efficiency of Pascal, even ignoring the benefits of altering the cache layout to maximise the benefits from tile-based rendering (which is something I'd expect them to do, Nintendo doesn't exactly shy away from large on-die memory pools).

There's really no reason to believe, or any evidence to suggest, that Switch will be in any way bandwidth constrained.

Nintendo also has a history of making sure memory bandwidth isn't a bottleneck for the last several products, so I'm excited to learn if they did any customization in this regard.

If they use any novel approach to adding memory bandwidth it could be a real leg up against most mobile chips, which really are memory bandwidth constrained, only getting past the 360/PS3 bandwidth level somewhat recently, and not of course including the eDRAM on the 360. Even the excellent Apple A10 is around 25GB/s, while only the 13" iPad Pro goes with double the pinout for 50GB/s.

 

[PdotMichael]

Actually 128bit bus could offer the same main memory bandwidth as Xbox One and TBR would give a very similar bandwidth saving to Xbox One embedded memory. So it could certainly be matched or at least extremely close to matched. Then again I don't expect Switch to be 1.4Tflops so why match bandwidth anyway?

could, could, could

Is it possible to have people using things based on official published data instead going full fantasy.

Stuff like HBM2 on Tegra Parker for a device that is supposed to be released in a few months (not even sure if it part of any roadmap for the Tegra series in the near future) is fantasy and one of the things people starting to turn into a "possibility" because of such circlejerking threads.

 

[MacTag]

could, could, could

Is it possible to have people using things based on official published data instead going full fantasy.

Stuff like HBM2 on Tegra Parker for a device that is supposed to be released in a few months (not even sure if it part of any roadmap for the Tegra series in the near future) is fantasy and one of the things people starting to turn into a "possibility" because of such circlejerking threads.

He's not talking about HBM2. 128-bit bus is far from "full fantasy", even 3DS used a 128-bit bus.

 

[PetrCobra]

I take your point but looking at what we have to go on - MK8, Splatoon and the new 3D Mario footage, do they look like they're a quantum leap over the WiiU versions ?

Yes they do.

But that still doesn't mean the Switch won't be significantly more powerful. A port is cheaper if you don't upgrade it too much, those games already looked good enough. And we need to see more of the new 3D Mario to judge the technical side.

 

[PdotMichael]

He's not talking about HBM2. 128-bit bus is far from "full fantasy", even 3DS used a 128-bit bus.

Do I have a stroke right now?

128bit is what Tegra Pascal uses.

Which means (nVidia data) 50GB/s transfer rate. Which is significantly slower than what we have in the Xbox One.

 

[Hero of Legend]

You know, the wording of many sources implies that the system has some considerable power despite its form factor:

ShockingAlberto: around XB1 power

Emily Rodgers: beyond the Wii U, but will not reach XB1 raw power

Laura: Vouches for Emily

NateDrake: Around 3x-4x the Wii U, the full power TX1 is a good estimate.

Matt: Not weaker than Wii U, not as "CPU limited" as the XB1/PS4, porting will not be a portable issue.

OB: Nintendo will be fine. Porting from the other consoles will not be a technical issue.

LCGeek: CPU stronger than XB1/PS4.

Did I miss anyone major?

Compare this to what was said about the Wii U from lherre, Akram, Ideaman, the devs, Epic, insider reports, and etc prior to the release. There was a very big "lower your expectations" flag coming from them (Ideaman tried to sugarcoat it, but it was obvious). I think we can worry a little less about the power of the Switch.

Don't forget Pascal from NateDrake.

But yeah I hope it's using the latest stuff it can, even if it's not uber powerful, but just enough to be modern and not considered to be dated.

Plus 4K-upscaling (but up to native 1080p) and HDR ala XB1S would be REALLY amazing.

 

[MacTag]

Do I have a stroke right now?

128bit is what Tegra Pascal uses.

Which means (nVidia data) 50GB/s transfer rate. Which is significantly slower than what we have in the Xbox One.

All donny mentioned was the bus and tile based rendering. I'm sorry I ever mentioned HBM2 as you seem to have seized on it in some insane attempt to discredit any realistic talk at all. HBM2 isn't a realistic solution as it's too new/expensive, I just threw it out casually to show LPDDR4 isn't the only option.

You still need a new line of attack though.

 

[Nostremitus]

Do I have a stroke right now?

128bit is what Tegra Pascal uses.

Which means (nVidia data) 50GB/s transfer rate. Which is significantly slower than what we have in the Xbox One.

I think you've had several over the course of the day considering how you've chosen to misinterpret damned near anything anyone has said to you. And seem to lack any short term memory as you keep bringing stuff up that has already been addressed.

Hell, I just had back surgery and I'm able to follow these conversations better than you... Either that, or you're intentionally misrepresenting what others are saying....

 

[PdotMichael]

All donny mentioned was the bus and tile based rendering. I'm sorry I ever mentioned HBM2 as you seem to have seized on it in some insane attempt to discredit any realistic talk at all. HBM2 isn't a realistic solution as it's too new/expensive, I just threw it out casually to show LPDDR4 isn't the only option.

You still need a new line of attack though.

So I asked for realistic solutions based on the goals for price, form factor, battery, heat. And you are just pulling buzzwords because why not?

 

[MacTag]

So I asked for realistic solutions based on the goals for price, form factor, battery, heat. And you are just pulling buzzwords because why not?

I and multiple posters have given you several; a high bus, embedded memory pools, tile based rendering, bandwidth efficiences in the APU. And it bears repeating that Nintendo's track record in post-N64 architecture design in this exact area is worth looking at. Feel free to continue keeping your head in the sand at 50GB/s though.

 

[EDarkness]

I and multiple posters have given you several; a high bus, embedded memory pools, tile based rendering, bandwidth efficiences in the APU. And it bears repeating that Nintendo's track record in post-N64 architecture design in this exact area is worth looking at. Feel free to continue keeping your head in the sand at 50GB/s though.

You guys should just stop responding to this guy. Seems like obvious trolling at this point.

 

[PdotMichael]

I and multiple posters have given you several; a high bus, embedded memory pools, tile based rendering, bandwidth efficiences in the APU. And it bears repeating that Nintendo's track record in post-N64 architecture design in this exact area is worth looking at. Feel free to continue keeping your head in the sand at 50GB/s though.

I have just nVidia on my side.

But who cares about nVidia I bet Matt or random insider know more.

 

[MacTag]

I have just nVidia on my side.

But who cares about nVidia I bet Matt or random insider know more.

Matt hasn't said a word on memory bandwidth. Are you really that butthurt over ARM vs Jaguar?

 

[lwilliams3]

I have just nVidia on my side.

But who cares about nVidia I bet Matt or random insider know more.

NVIDIA on your side about what? NVidia calls it a customized chipset, so Nintendo can request a number of changes and it will still be under the same label.

 

[MuchoMalo]

The Pascal Tegra doesn't use a 256-bit bus.

nVidia claims 50GB/s for Parker. But sure Nintendo can use a supercharged Parker I mean what could stop them except price, battery, heat.

It does use a 256-bit bus. You need a 256-bit bus to hit 50GB/s. A 128-bit bus like the Tegra X1 is generally limited to 25GB/s with LPDDR4, with a theoretical max of 30GB/s.

Edit: I may have remembered incorrectly. It doesn't really make a difference either way. The point is that 50GB/s is what to expect if there's no eSRAM. You're still clinging on to this point way too much, and you're trying too hard to rule things out with no proof.

 

[Donnie]

could, could, could

Is it possible to have people using things based on official published data instead going full fantasy.

Stuff like HBM2 on Tegra Parker for a device that is supposed to be released in a few months (not even sure if it part of any roadmap for the Tegra series in the near future) is fantasy and one of the things people starting to turn into a "possibility" because of such circlejerking threads.

Is it possible for you to calm down and read my post before replying? I'm not talking about HBM2. Just about LPDDR4 on a 128bit bus, which is already earmarked for Tegra Parker.

 

[Donnie]

Do I have a stroke right now?

128bit is what Tegra Pascal uses.

Which means (nVidia data) 50GB/s transfer rate. Which is significantly slower than what we have in the Xbox One.

You haven't had a stroke, you just need to calm down and read people's posts. 128 LPDDR4 can provide around 60GB/s+, which isn't significantly under Xbox One.

 

[blu]

Wasn't the line "supercharged pc architecture"?

A 4-QPI Xeon allowing for fully-connected, fully-coherent quad-socket topologies with a 25+GB/s spare link to a local accelerator (say, a GPU) is a 'supercharged pc architecture'. An APU with an extra bus where a cluster experiences coherency issues talking to its peer sitting on the same die is a supercharged, well, APU.

 

[Mr. Wonderful]

I would personally prefer a boost to 1080p resolution with anti aliasing, anisotropic filtering and improved textures at a more consistent 30fps over the same WiiU visuals at 60fps but it would be nice if Nintendo offered the option like some other developers are now doing.

I feel similarly.

If Nintendo gave me Wii U 1.5 level of graphics plus dynamic resolution between 800-1080p when docked, I'd be happy. Nintendo is going to be able to do fantastic things with first-party games on this system, given how impressed I was with what they did with the Wii U. It's third-party games that are going to be rough looking.

 

[Polygonal_Sprite]

Yes they do.



But that still doesn't mean the Switch won't be significantly more powerful. A port is cheaper if you don't upgrade it too much, those games already looked good enough. And we need to see more of the new 3D Mario to judge the technical side.

Can you show me in screenshots what improvements Splatoon and MK8 have over the WiiU versions ?. Mario Switch certainly looks nice but I don't think it looks like a generational leap over 3D World esp when you consider that 3D World had to account for a possible four playable characters on screen when designing it's environments.

From what Nintendo have shown so far (which is all we really have to go on outside of vague insider info), Switch certainly looks like a modest leap over WiiU in terms of specs which lines up with what Emily has said (falls short of the XB1 GPU, half XB1's memory) and also falls in line with the limitations a system which is designed to be played on the go would have due to heat and battery life.

 

[Skittzo0413]

Can you show me in screenshots what improvements Splatoon and MK8 have over the WiiU versions ?. Mario Switch certainly looks nice but I don't think it looks like a generational leap over 3D World esp when you consider that 3D World had to account for a possible four playable characters on screen when designing it's environments.

From what Nintendo have shown so far (which is all we really have to go on outside of vague insider info), Switch certainly looks like a modest leap over WiiU in terms of specs which lines up with what Emily has said (falls short of the XB1 GPU, half XB1's memory) and also falls in line with the limitations a system which is designed to be played on the go would have due to heat and battery life.

I just don't think any footage we've seen has been running on anything close to the Switch hardware, so it's impossible to judge hardware power based on that. I'm sure everything in trailer was running on development PCs.

 

[EDarkness]

Can you show me in screenshots what improvements Splatoon and MK8 have over the WiiU versions ?. Mario Switch certainly looks nice but I don't think it looks like a generational leap over 3D World esp when you consider that 3D World had to account for a possible four playable characters on screen when designing it's environments.

From what Nintendo have shown so far (which is all we really have to go on outside of vague insider info), Switch certainly looks like a modest leap over WiiU in terms of specs which lines up with what Emily has said (falls short of the XB1 GPU, half XB1's memory) and also falls in line with the limitations a system which is designed to be played on the go would have due to heat and battery life.

I'm pretty sure we haven't seen a game running on an actual NS yet. They're reserving that for January when they can pull out the stops and from there we'll get an idea of what the system can do. Right now, we only have a basic idea of specs, but no complete details. No doubt this upcoming event is going to be important. They seem to be confident about what it can do as they're working hard trying to bring in everyone to take a look at it in the days following the event in Tokyo.

 

[test_account]

Yes they do.

But that still doesn't mean the Switch won't be significantly more powerful. A port is cheaper if you don't upgrade it too much, those games already looked good enough. And we need to see more of the new 3D Mario to judge the technical side.

Regarding the comic, what makes it small?

 

[Nostremitus]

Regarding the comic, what makes it small?

Quantum refers to things smaller than atoms.
Ie quantum physics or a quanta of light.
It's theoretically the smallest possible area of study


Like, the opposite of astronomical.

 

[test_account]

A 4-QPI Xeon allowing for fully-connected, fully-coherent quad-socket topologies with a 25+GB/s spare link to a local accelerator (say, a GPU) is a 'supercharged pc architecture'. An APU with an extra bus where a cluster experiences coherency issues talking to its peer sitting on the same die is a supercharged, well, APU.

Both things are architectures used for PC, arent they? Or whats the difference in defining them as architectures?

Quantum refers to things smaller than atoms.
Ie quantum physics or a quanta of light.
It's theoretically the smallest possible area of study


Like, the opposite of astronomical.

Sure, but the leap is referring to an event, isnt it? An event can still be concidered very important/big and significant even if the elements are very tiny. This is also the first thing i think of when i see quantum leap:

 

[Nostremitus]

Both things are architectures used for PC, arent they? Or whats the difference in defining them as architectures?



Sure, but the leap is referring to an event, isnt it? An event can still be concidered very important/big and significant even if the elements are very tiny. This is also the first thing i think of when i see quantum leap:

A quantum leap would be an event so insignificant that it barely registers as existing.