NVIDIA Tegra Architecture

[liolio]

Android is perfectly fine for what it's called for to do and especially under the perspective that it has to serve such a multitude of different hw and sw configurations. As for Lollipop there is always a price for early adopters and it's not like Apple's new OS versions are always troublefree and not in the very least in windows land either. No idea where you're trying to get to, but if you should blame even indirectly Android for the state Denver is in or how it behaves, all you have to do is compare it in real time against Tegra K1/32bit either in Android Lollipop or Kitkat and then it'll be easier to find the real scapegoat.

Not what I'm getting to at all, more pointing out that neither the hardware or the software was great with that round of nexus devices.
Look at the Shield tablet battery life while running lollipop in Anandtech's review.
I think you are reading way too much in one post.

How can they give up Android swing to windows10 and re-address the smartphone SoC market at the same time? It's not impossible at all, but I'd say that NV would have to have quite suicidal tendencies to go for windows phone right now.

Well they go for no phones at all. Better (try to) be a big fish in a tinier pond and take a bet on Window future, as an aside they have software ready.

Apple has its own OS and R&D can go hand in hand both for hardware and software and NONE of them out there has that kind of luxury. At the very least Cyclone is quite more efficient at its frequency, doesn't burn a hole in their devices while overheating and it doesn't cause any occassional stutters either. As for their GPUs yes Apple likes them big and with as high as possible sustainable performance. For the prices they're asking they SHOULD deliver whatever best is possible for any given timeframe.

Reading issues? I just stated that Apple does not sell specs to its costumers, there are a lot more thing to a product than the specs of its SoC. It actually also applies to the screen, the camera, etc.

Finally for the point that NVIDIA should have had mainstream SoCs for smartphones, that's something I've been saying for a very long time (and have been attacked for....). It should have happened years ago though; now I'm afraid there's only one Mediatek the market can have especially for chinese white box deals.

That I definitely agree with.

 

[RecessionCone]

Finally for the point that NVIDIA should have had mainstream SoCs for smartphones, that's something I've been saying for a very long time (and have been attacked for....). It should have happened years ago though; now I'm afraid there's only one Mediatek the market can have especially for chinese white box deals.

Nvidia structurally cannot compete with Mediatek. Nvidia needs high margins and differentiation to fuel its R&D costs. Nvidia cannot survive in its current form while fighting Mediatek, Allwinner, Rockchip, or any other designer whose R&D expenses consist of floorplanning modules together from 3rd party vendors.

 

[liolio]

Nvidia structurally cannot compete with Mediatek. Nvidia needs high margins and differentiation to fuel its R&D costs. Nvidia cannot survive in its current form while fighting Mediatek, Allwinner, Rockchip, or any other designer whose R&D expenses consist of floorplanning modules together from 3rd party vendors.

They cannot compete with qualcomm, Intel or Apple either... In any case it is not like any company in this world has the luxury to choose its competitors, even Apple can't, whether you win or lose is another matter.

The real issue is whether or not Nvidia has an IP worth selling on that market. Somehow if they can't sell their GPU IP to any of the aforementioned actors it is possible that the answer to the prior question is no.
Now there are other thing to sell than raw hardware, Apple does not sell raw hardware to its costumers. It (could) makes sense for Nvidia to produce SoC if they foremost produce them for themselves.
EDIT
I wonder if the later is even true. To some extend I wonder if the best motive for Nvidia to be on that market is to avoid facing (misslead, mania loving) investors if they were not...

 

[A1xLLcqAgt0qc2RyMz0y]

They cannot compete with qualcomm, Intel or Apple either.

Hyperbole.

Nvidia did win the Nexus 9 and seem to be doing well in auto.

 

[Ailuros]

Not what I'm getting to at all, more pointing out that neither the hardware or the software was great with that round of nexus devices.
Look at the Shield tablet battery life while running lollipop in Anandtech's review.
I think you are reading way too much in one post.

Then I apologize. It's clearer now.

Well they go for no phones at all. Better (try to) be a big fish in a tinier pond and take a bet on Window future, as an aside they have software ready.

To go for smartphones they'd need something like the Tegra4i and a solution for CDMA for the US market

Nvidia structurally cannot compete with Mediatek. Nvidia needs high margins and differentiation to fuel its R&D costs. Nvidia cannot survive in its current form while fighting Mediatek, Allwinner, Rockchip, or any other designer whose R&D expenses consist of floorplanning modules together from 3rd party vendors.

They also announced quite some time now that they are selling from Kepler and onwards GPU IP. NVIDIA structurally cannot compete against ARM or Imagination either for GPU IP and compared to the above it's far more absurd.

For the record I didn't say they should do it NOW, I said they should have gone for it years ago before Mediatek came to hundreds of millions smartphone SoCs sold on an annual basis.

Finally when it comes to survival, is Tegra actually "surviving" over all those years? With the expection of the Tegra3 and this year where they managed to cash in some minor profit from the department is there something I've missed. Holding your head barely over the water isn't a solution either.

Hyperbole.

Nvidia did win the Nexus 9 and seem to be doing well in auto.

Then you must have an awefully weird definition of "competing". Let's try marketshare persentages instead and then we'll see how a single digit persentage exactly is a "hyberbole".

 

[A1xLLcqAgt0qc2RyMz0y]

Then you must have an awefully weird definition of "competing".

No it is you who has an issue in understanding the definition of "competing".

by saying "They cannot compete with qualcomm, Intel or Apple either" implies that they have NO CHANCE ever of winning any designs.

Since they did win the Nexus 9 and are doing well in auto they can compete.

 

[Ailuros]

No it is you who has an issue in understanding the definition of "competing".

by saying "They cannot compete with qualcomm, Intel or Apple either" implies that they have NO CHANCE ever of winning any designs.

Since they did win the Nexus 9 and are doing well in auto they can compete.

They can have 1000 design wins as a matter of speach and sell only 10 devices from each. If that's still a "success" in your book then all hail to your blindfolds.

 

[Deleted member 13524]

As long as nVidia is making money with the Tegra division (which AFAIK they are), isn't it a "healthy" business, regardless of how much more money the competitors are making?
I'm not sure if the Nexus 9 is going to sell a lot because Google decided to sell it at a premium price.. but between the Nexus 9, the Mii Pad and the 13" Chromebook, I'd say it's possible that nVidia is making money out of TK1.


I'm more worried about the Tegra line and their advantage in gaming performance getting more irrelevant each year.
No games in Android seem to be pushing the boundaries in 3D performance. Gameloft are probably the ones making the most demanding games and they all run fine @720p in a 3-year-old Snapdragon S4 with an Adreno 225 or the cheap S400 with the Adreno 305.
Console ports are still limited to the 6th generation (Playstation 2) like GTA 3 and such. There's Half Life 2 and Portal but nVidia is actively killing the demand for these games with its own greed, by blocking them from playing in other SoCs.
Moreover, gamepad support in Android keeps being terrible so making a "serious" gaming machine out of our smartphones/tablets/TV-boxes is still not a realistic option.

Very high 3dmark and GFXBench scores can only take you so far, if there are no games to show it off.
x86 windows with legacy support has a huge and established user base for high-end videogames, but Android doesn't.
Who will want to buy nVidia their SoCs with very powerful GPUs, when everyone figures out there's no practical difference in android's gaming performance/visuals between a Tegra X2 and a Mediatek MT6895 (made up names, of course)?

 

[Picao84]

nVIDIA has a press event on March 3rd. GsmArena speculates it is about the new Shield Tablet... However, the invite talks about "5 years in the making". We all know that the Shield line was created around 2 years ago. They could be making a reference to Tegra X1 with that, but stilll... And I doubt a new Shield Tablet would be enough to redefine the future of gaming...

http://www.gsmarena.com/nvidia_shield_tablet_successor_might_debut_on_march_3-news-11146.php

 

[liolio]

The tegra 2 got available 5 year ago going by wikipedia data.

 

[Gubbi]

The tegra 2 got available 5 year ago going by wikipedia data.

One of the first products to launch with Tegra 2 was the LG Optimus 2x. It launched in january 2011 in Korea and the rest of the world after march.

Cheers

 

[Picao84]

The tegra 2 got available 5 year ago going by wikipedia data.

True, but Shield only showed up with Tegra 4 which was what? 2 years ago? In any case, Tegra X1 was already announced, so at this point is nothing new.. Only a new Shield would be. But unless they are talking about a radical new form factor for Shield (e.g. Set top console), I cannot see what would justify the use of the expression "redefine gaming"...

 

[liolio]

Along with with Tegra 2 Tegra Zone and matching "optimized" games were released too (iirc /fuzzy timeline) may be it relates to that?
I don't know just searching

 

[tangey]

As long as nVidia is making money with the Tegra division (which AFAIK they are), isn't it a "healthy" business, regardless of how much more money the competitors are making?

If the following figures are correct, they suggest that Tegra is a loss making division.

Years figures to end of Jan 2014
http://csimarket.com/stocks/segments_anual.php?code=NVDA
Tegra lost $268M, on revenue of $398M

And the following are figures for the quarter oct '2014
http://csimarket.com/stocks/segments.php?code=NVDA
Lost $53M on revenue of $168M

 

[Alexko]

If memory serves, the Tegra division has never made a cent of profit.

 

[Ailuros]

Along with with Tegra 2 Tegra Zone and matching "optimized" games were released too (iirc /fuzzy timeline) may be it relates to that?
I don't know just searching

Someone at the 3DC forum just suggested that it might be some sort of standalone (Android based?) VR HMD. That could revolutionize gaming and it truly could be a long time under development?

 

[Ailuros]

If memory serves, the Tegra division has never made a cent of profit.

They made a small profit for the Tegra3 year.

 

[Erinyes]

Someone at the 3DC forum just suggested that it might be some sort of standalone (Android based?) VR HMD. That could revolutionize gaming and it truly could be a long time under development?

That seems to fit the description. Either that or an android based gaming console. Another shield wouldn't exactly be revolutionary.

They made a small profit for the Tegra3 year.

Quite possible..maybe even for the Tegra 2 year. But do we know this for sure? I dont think they've ever given a per division profit/loss breakup.

 

[Ailuros]

Quite possible..maybe even for the Tegra 2 year. But do we know this for sure? I dont think they've ever given a per division profit/loss breakup.

If you dig deep way into the past you might find a statement from Jensen claiming that the department's break even point is at $1b. Apparently it must have shrunk down to ~700Mio considering that the T3 year revenue was a tad over that

 

[xpea]

first leaked Tegra X1 benchmark:
http://blog.gsmarena.com/nvidia-tegra-x1-flies-past-70000-antutu/
as usual it's Antutu (a Chinese love affair) but looks promising with a very high GPU score...