Is there really any point in having different clocks for docked and undocked in the CPU department anyway? I would have imagined that most devs would want their game logic to be identical between the two or just wouldn't be bothered to enhance it (ai for instance). Unless the intention is for devs to look at lower poly on handheld vs docked and I am not sure they'd do that.
-
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.
DF: Nintendo NX Powered By Nvidia Tegra! Initial Spec Analysis
- Thread starter DeathSparks
- Start date
- Status
Is there really any point in having different clocks for docked and undocked in the CPU department anyway? I would have imagined that most devs would want their game logic to be identical between the two or just wouldn't be bothered to enhance it (ai for instance). Unless the intention is for devs to look at lower poly on handheld vs docked and I am not sure they'd do that.
Oh, but I did. The A72s would clock to 620MHz at the same power draw as the A53s at 1.5GHz (387mW for a quad-core cluster, FYI), so you were off by about 25% on clock speed, and the A72s in this case would provide 22% lower performance (by the same Geekbench floating point metric I've been using). In any case my point was more about the lack of data in the linked slide rather than the validity of your results.
It basically all depends on the kind of workloads Nintendo would be using to test the performance of each option, which is why it's so hard to judge what course they would take. If they're using sample code from Wii U games, then even a 1.4GHz A53 should comfortably outperform Espresso in single-thread workloads, and with an octo-core 1.4GHz A53 configuration coming it at 2/3 of a Watt it would massively outperform it in heavily threaded tasks. Otherwise they could use an asymmetric little.little configuration (say a 1.6GHz A53 cluster and a 1.1GHz A53 cluster, which would also come in at 0.66 W) if they want to squeeze a bit more single-thread performance at the expense of a little multi-threaded performance. I also don't think that Nintendo is quite as cut off from the rest of the world in this regard as we usually assume. They switched to a tri-core setup for Wii U after XB360 had used the same configuration, and they may well have decided that, following on from PS4 and XBO, developers would be comfortable with an octo-core config.
This is actually my original point of contention with blu's claim that Nintendo would use a quad-core A72 plus quad-core A53 setup, yet leave the A72s disabled while in handheld mode. Unless you expect to have games that only run in docked mode, there's not really much reason to expect a game which runs smoothly on a 1.5GHz A53 cluster to make proper use of a 2.5GHz set of A72s. You can certainly do some things to use up the extra processing power if it's there (e.g. increase the complexity of physics simulations, etc.), but none of it strikes me as worth dedicating the die space necessary for a quad-core A72 cluster. Certainly if docked mode gives you the thermal/power draw headroom you might as well clock the CPU up a bit, but I don't see the point of including an entire redundant CPU cluster on there just for that scenario.
GhostTrick
Banned
Question to you guys:
Where do you find these CPU power consumption ?
From anandtech, A72 doesn't seem to consum that much
http://www.anandtech.com/show/9878/the-huawei-mate-8-review/3
Where do you find these CPU power consumption ?
From anandtech, A72 doesn't seem to consum that much
http://www.anandtech.com/show/9878/the-huawei-mate-8-review/3
RoadHazard
Gold Member
Because many of us want a Nintendo home console with some decent power and don't really care for the portable aspect which will gimp what it can do.
Shin Johnpv
Member
I don't know if this has been brought up in the thread, it's a long thread and hard to keep up with it all. People keep talking about docking vs undocked speeds and what they could do. Would it be possible to have a different chipset in the dock? Like a higher clocked SoC in the dock than in the handheld?
It would be an expensive release, but is that something they could do? Basically really selling you 2 systems as "One"? So docked it runs on the dock chips, and when on the go it runs the hh's internal chips?
It would be an expensive release, but is that something they could do? Basically really selling you 2 systems as "One"? So docked it runs on the dock chips, and when on the go it runs the hh's internal chips?
Relaxed Muscle
Member
You must have had a rough time since the Wii release, then.
You must be struggle for decades now. I feel sorry for you.
The only thing that's technically feasible is a GPU in the dock. That won't happen due to cost and Nintendo, though. Again, if there's no GPU in the dock I'll maintain that this has no right to call itself a hybrid.
Gamecube has it's 15th birthday soon.
It would cost way too much for gains that Nintendo don't care about, and more importantly, would easily push this in the 350$+ realm and it's suicide to sell an handheld at those prices.
Dreamwriter
Member
Huh? Rumors state that it's a single device meant to be both a handheld game system and a home game console, that is quite a bit more powerful than the Wii U. One device for two purposes. How is that not a hybrid? I would say your requirement would actually make it NOT a hybrid, because putting a different GPU in the dock would make the dock a game console.
That's just my opinion. It's like giving a phone an HDMI out and calling it an HTPC. And a console is obviously more than a GPU.
Dreamwriter
Member
A lot of people considered the OUYA to be an HTPC (one of its few good uses), and the OUYA was basically a phone with HDMI out and no screen,
And I should care what other people think... why? It's at least designed as a console/HTPC though. NX is a handheld shoehorned into replacing a console based on what's known. I'd let it slide of it weren't possible to make it a true current-gen console by just supplying one accessory for people who might want it, even at a premium, but if we're not getting that it doesn't do anything that a handheld can't do.
Unless they were actually able to let you use your smartphone as the handheld screen. It would be sort of messy and you might need adapters for different phones, but not having to include a screen would definitely save a lot of money. I'd feel it was worth it if it meant they could put a Tegra X2 in the handheld and another X2 in the dock for $299 or so.
Or... the opposite. Since that quantity just doesn't bother with what has been put out, because it's not what they want.
If your favorite action moviestar starts making romantic comedies you don't like, would you still be watching those? Or would you just look elsewhere for your action movies?
This is my theory.
Nintendo will most likely have its own custom version of the Tegra rather than an X1 or X2.
From a pure handheld perspective they probably will devote most of the Watt budget to the GPU rather than the CPU so I would expect 4xA53.
However I also think a big important requirement for the NX is that it can emulate earlier Nintendo consoles.
A53 emulating the Wii CPU is most likely not possible, does anyone know of any Dolphin tests runnning on A53s?
Because of that I think they will add at least one higher performance CPU core.
nVidias Denver CPU is extra interesting here because it is actually not a native ARM CPU.
It translates ARM instructions to its native instruction set and has hardware to optimize this by storing the translated instrcutions in a rather large cache.
http://www.anandtech.com/show/8701/the-google-nexus-9-review/4
Nintendo could potentially work with nVidia to make the Denver CPU execute PowerPC code very efficiently.
On the other hand an A72 or A73 would be fast enough to emulate the Wii CPU the traditional way so it might be unnecessarily complex to write the Denver translator.
If Nintendo want to be able to emulate the Wii U also they probably need at least 3 fast cores and 3 A72/A73 running fast enough to emulate the Wii U would draw a lot of power.
In the end my theory is 4xA53 + 1 Denver, or 4xA53 + 4 Denver depending on if they want to emulate Wii U or not.
The Denver cores would run at low clock speed in handheld mode but could run either ARM or PowerPC code.
In docked mode they could run at full speed.
What do you guys think?
Nintendo will most likely have its own custom version of the Tegra rather than an X1 or X2.
From a pure handheld perspective they probably will devote most of the Watt budget to the GPU rather than the CPU so I would expect 4xA53.
However I also think a big important requirement for the NX is that it can emulate earlier Nintendo consoles.
A53 emulating the Wii CPU is most likely not possible, does anyone know of any Dolphin tests runnning on A53s?
Because of that I think they will add at least one higher performance CPU core.
nVidias Denver CPU is extra interesting here because it is actually not a native ARM CPU.
It translates ARM instructions to its native instruction set and has hardware to optimize this by storing the translated instrcutions in a rather large cache.
http://www.anandtech.com/show/8701/the-google-nexus-9-review/4
Nintendo could potentially work with nVidia to make the Denver CPU execute PowerPC code very efficiently.
On the other hand an A72 or A73 would be fast enough to emulate the Wii CPU the traditional way so it might be unnecessarily complex to write the Denver translator.
If Nintendo want to be able to emulate the Wii U also they probably need at least 3 fast cores and 3 A72/A73 running fast enough to emulate the Wii U would draw a lot of power.
In the end my theory is 4xA53 + 1 Denver, or 4xA53 + 4 Denver depending on if they want to emulate Wii U or not.
The Denver cores would run at low clock speed in handheld mode but could run either ARM or PowerPC code.
In docked mode they could run at full speed.
What do you guys think?
Hell yeah you guys showed him! He should have known not to have a different preference to you!
It was totally probable for the NX to reach XB1 or PS4 levels if it were a traditional console considering it will be a mid gen console and tech around that level is very affordable. Its not like the PS4/Xbox1 are straight beasts here, guys. If he isn't in to handhelds, he has a right to be let down if this is all the console offers. This was supposed to be Nintendo's next home console after all.
Yes but you wouldn't do it as a CPU. So a common scenario that looks like that is PC CPUs. Intel bakes integrated graphics into all of them now. However if you pair it with an external GPU you use that instead and several laptop manufacturers are exploring using USB/ePCI types of connectors such that you can plug-in external cards to laptops. This is certainly feasible. Note that this would essentially continue to use the existing CPU but disable the integrated GPU, maybe they could come up with a way to use both but that's tricky and probably not worth the effort.
TLZ
Banned
The graph I posted above is based on that article, but it's worth keeping in mind that I'm quoting figures for quad-core clusters of A72s and A53s (not individual cores), and I'm assuming constant clock speeds (unlike phone SoCs where clocks jump up and down as required).
I don't really take screen size into account, as generally a larger screen means space for a larger battery, which cancels out the increased power draw from the larger screen. Regarding power draw, yes, I would be assuming a total of about 2W for the SoC, with that heavily weighted towards the GPU (so the GPU consuming perhaps twice the power of the CPU). It's conceivable that Nintendo could use a slightly higher TDP, though, and/or a different allocation of power between CPU and GPU, so there's definitely some wiggle room there.
In theory Denver is well-suited for emulation (that's effectively what it's doing when it runs ARM code), but there would be a few drawbacks when using it in a console:
- While running ARM code, it performs well in synthetic benchmarks, but poorly in less predictable real-world scenarios. If the same is true while running PPC code, then there's the potential for stuttering or erratic performance of VC games, which would be difficult to fix (unlike traditional emulation where there's a software emulation layer you can tweak on a per-game basis if necessary).
- The A72 is much smaller, much more power efficient, and significantly outperforms it in both general purpose and game-related scenarios. It doesn't really make sense to give up that much performance just for the sake of slightly better VC.
- The A72 should be able to handle GC and probably Wii games adequately anyway. The TX1 (which uses the older A57 cores at 1.9GHz) can run Dolphin very smoothly, and that's on top of Android with a reverse-engineered emulator. Nintendo can run their emulation as close to the metal as they like, and obviously have a perfectly detailed knowledge of how the original hardware operates, so you would expect them to achieve better performance with their own emulator. In addition, the A72 has higher performance per clock than A57, and could potentially clock higher even in handheld mode (see discussion below), so there wouldn't really even be a need for Denver.
This does follow on neatly to a thought that I had about possible CPU configurations for NX, though, in relation to discussions with blu above about single thread versus multi-thread performance. It occurs to me that a more esoteric core configuration might be worthwhile, and that it's worth considering a 2:4:2 config consisting of 2x A72, 4x A53 and 2x A35 (with the A35s reserved for the OS, with crypto clocks but without NEON). Clocked down to ~800MHz the A35s could consume as little as 50mW while handling background OS duties (which would also keep standby power down), and the remaining six cores would provide a good mix between single thread and multi thread performance in the power envelope we're looking at.
Given that I have the data in front of me in a spreadsheet, I quickly threw together this graph showing the tradeoff between single and multi threaded performance within a 660mW power budget as you alter power draw between the A72s and A53s in this scenario:
As the graph moves to the right more of the 660mW is allocated to the A72s, while on the left more is allocated to the A53s. The blue line shows total multi-threaded performance at that particular power allocation, and the red line shows the peak single thread performance (ie performance of the most powerful individual core). The two lines are scaled relative to just using 4x A53 in the same power envelope. The relative performance between A53 and A72 at the same clock is based on the same Geekbench floating point figure I've been using in my previous comparisons.
Looking at the graph it would appear that dedicating a power budget of around 70% to the A72s would give a good balance between single and multi-thread performance, giving a moderate boost to each over a quad-A53 setup. (This would put the A72s at about 1150MHz and the A53s at about 1000MHz, for reference)
Getting back to VC, though, this kind of setup would allow them to achieve substantially better emulation performance by changing the power distribution between the different cores and the GPU. In this case, where single-threaded CPU performance is a priority, they could clock the A72s up to about 2GHz, clock the A53s down considerably (or even disable them altogether) and clock the GPU down quite a bit as well, all while staying in a ~2W TDP for the full chip. Given Dolphin performance on TX1, that should allow Nintendo to achieve good quality GC/Wii emulation even in handheld mode, which wouldn't really be possible at all on A53s alone.
Furthermore, they could also adopt phone-style burst clocks when browsing the OS, and using non-gaming apps like the web browser. There wouldn't be anything stopping them from clocking the A72s up as high as 2.5GHz for short bursts, which would result in a very snappy user experience.
A 2:4:2 configuration like that seems reasonable. It would probaly rule out Wii U emulation though.
For their own games they could potentially recompile/port them to the NX. But it would obviously be much more work even if they port over all the support libraries and APIs.
A73 is supposed to be even more efficient than A72.
I wonder if its too new for the NX. ARM claims it will be in consumer devices end of this year but mass produced in a Nintendo consol by March 2017 feels like a stretch.
For their own games they could potentially recompile/port them to the NX. But it would obviously be much more work even if they port over all the support libraries and APIs.
A73 is supposed to be even more efficient than A72.
I wonder if its too new for the NX. ARM claims it will be in consumer devices end of this year but mass produced in a Nintendo consol by March 2017 feels like a stretch.
Yeah, I wouldn't expect Wii U emulation at all, even if we were talking about a traditional stationary console with active cooling and high power draw. My guess is they'll just port over any games they feel would have a market on NX.
I hadn't realised ARM claimed that we'd see A73s in devices by the end of the year (their clock/power draw claims were all based on 10nm, so I assumed they would be a while off). They claim the A73 offers a ~20% perf/W gain even on the same process as A72, so it would be nice if Nintendo managed to squeeze them in there, but I wouldn't really count on it.
Yeah, this. We already know about Breath of the Wild, I wouldn't be surprised to see ports of older Wii U games that aren't day-and-date. Not to mention that given the comparatively small jump in power this time, we'll likely see some 'new instalments' built on the foundation of enhanced ports. For example, Mario Kart 9 will probably look a fair bit like MK8 but in 1080p with a bunch of extra features, content, etc.
Toadthemushroom
Member
Thanks, that would be good. It would be a shame if Nintendo decides to cut back on some aspects of the device only for short term benefit.
It's important they get the first generation NX hardware right so it is a good baseline for developers to target when making forwards-compatible games for newer generations of NX devices with more powerful hardware.
(It's also why I was disappointed that Apple went with the A8 SoC for the Apple TV, since that limits the scope of larger games).
Not a chance. The source of that rumor is a known Nvidia hater and will say anything to make Nvidia look bad. The only truth in that article was that he heard that Tagra is powering NX and that NX is a handheld.
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
It's not about losing face, it's about the survival of the Tegra lineup. And that part is quite believable - NV do need a solid customer for the Tegras, there's not way to spin that, "hater" or otherwise.
Plinko
Wildcard berths that can't beat teams without a winning record should have homefield advantage
Unless you're an insider and have knowledge, there is absolutely no way you can definitively make this statement.
dr_rus
Member
Tegra lineup is fine, it's growing in revenue all the time and is selling fine to the auto industry. People seem to have that stupid assumption that mobile SoCs are needed only in smartphones while in fact that's the market where they aren't needed anymore as main players all have their own h/w now and the rest are just fighting for pennies.
Still, i'm sure more mobile phones and tablets are being sold worldwide, per family over a period of 5 years, than cars.
dr_rus
Member
Mobile market is on decline, it's mostly about how cheap you are there now. NV's strong graphics side doesn't interest mobile market much and NV's factual unwilling to sell anything with no profit means that they've essentially left mobile market completely - they've said as much on one of recent investor calls.
I'm pretty sure that Tegra deal for NX will be profitable for NV. They gain nothing by sponsoring their presence in console space. Contrary to a popular believe they aren't "salty" about it at all.
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
Tegra is the epitome of 'not doing fine'. It's selling only to the auto industry, where the margins are worse than mobile (since your brought up revenue and pennies). NV need a good margin market to flourish - NV are living off PC (enthusiast) and HPC. Unfortunately PC is on a decline itself and IGPs/APUs are dominating there, whereas HPC is pretty boutique. So NV need desperately to move up their mobile parts to at least mobile levels of profit, if that division hopes to survive.
Bigrx1
Banned
This is kind of like bragging you have the biggest and baddest canoe. I mean, ok, that's cool I guess, but it's a canoe....... I would rather have a new house or speed boat.
Why if it IS a portable device?
Do you have the biggest and baddest house and speed boat?
LegendofLex
Member
I'm sure if you asked Apple customers what they wanted in 2006, they wouldn't have said they wanted an iPhone.
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
It's really ironic when people go for absolutes, only to realise down the line their absolute is somebody else's peanuts.
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
There's some distance in between 'not doing fine' and 'flourishing', you know.
Having a few high-profile customers like google and nintendo puts you in a much better position than having zero mobile deals and clinging to automotive, even though you might not be exactly flourishing yet.
dr_rus
Member
Tegra is a as much a byproduct of gaming h/w as is HPC Tesla and professional Quadro. They consciously moved out of mobile market which is slowing down and has become a lot less interesting for 3rd party SoC providers. Whatever they've lost by this move is compensated by automotive / deep learning pushes which might be even more interesting long term than mobile.
NV does not "need desperately" anything for Tegra as it's basically still a side business for them which is doing fine on whatever markets they have it on right now. Again: NV certainly doesn't "desperately need" neither mobile nor console presence.
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
I never said they took a hit. I've said that the benefits from this deal for NV are most likely larger as a foot-in-the-door than as margins.
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
Long-term automotive and deep learning are hanging off a cliff, soon to follow the way of the bitcoin mining, and NV are well aware of that. Google have long announced they're moving to in-house ASICs. And so will do automotive - again, they're way more margin-sensitive than mobiles - it makes zero sense for them to stick to GPGPU in the long run.
It's their Tegra division that desperately needs to catch traction. Unless you believe it's a free hobby of sorts.
ed: oops.
I think people are missing the obvious. Nvidia doesn't need to get high margins in 2017 when there will be years of revisions, especially if Nintendo moves to an iterative model. There's plenty of time to get those margins, in the mean time they have steady work and a great ad for the Tegra line.
- Status