[Ars] Tegra 4i benchmarks are in!

[Homeles]

Most of that cost is the software to enable the encryption & security that are the selling points of that phone.

Yeah.

 

[NTMBK]

The Blackphone was going to go with a different SoC, until NVidia threw a ton of free developer time at them in return for using a Tegra 4i... That kind of desperation is pretty telling.

 

[Homeles]

The Blackphone was going to go with a different SoC, until NVidia threw a ton of free developer time at them in return for using a Tegra 4i... That kind of desperation is pretty telling.

I'm sure Nvidia's devs were thrilled.

 

[PPB]

How does it compare in battery life tests against the s400?

 

[blackened23]

Didn't JHH basically say MediaTek basically beat them on the OEM deals?

They re-positioned their market strategy for tegra, in a nutshell they don't want to be a low margin commodity player. Which is essentially what is happening with tons of cheap smartphones and 100$ mediatek android tablets everywhere. They tried and were successful up to tegra 3 with volume sales, but tegra 4 was just late to market and did not have integrated LTE. Lack of LTE killed nearly everyone except qualcomm and mediatek in China.....and by the flip side of the coin, with the commodity market people just don't give a crap about how great the graphics performance is in the sub 200$ market. People with mediatek devices don't give a F about their performance, it's a drive to the bottom in terms of price and that's all that matters. And the high end is dominated by qualcomm because of their modem IP. (LTE)

From the last tidbit I read, Nvidia want to create high end graphics/CPU performance with their SOCs, with the higher margins associated with it. So that chip (Denver and K1) will be in gaming devices, high end automobiles, scientific applications, developer kits, and things along those lines. It will not be a huge volume product, but a low volume/high margin/high profit/high end one. Similar to their graphics cards which are aligned towards the high end, much like the 780ti is aligned toward the high end of the market with a price to match. The Tegra K1 from what I've read will be marketed for the high end, and the benchmarks show it to have extremely high performance along with being the first true 64 bit ARM SOC. It won't be positioned at all for smartphones or sub 200$ devices.

So I wouldn't expect to see tons of K1 devices. Just isn't going to happen. But I would expect NV to make decent profit from it though, since they're positioning it towards professional and scientific applications with the high performance and price to match.

 

[mrmt]

From the last tidbit I read, Nvidia want to create high end graphics/CPU performance with their SOCs, with the higher margins associated with it. So that chip (Denver and K1) will be in gaming devices, high end automobiles, scientific applications, developer kits, and things along those lines.

(...)

So I wouldn't expect to see tons of K1 devices. Just isn't going to happen. But I would expect NV to make decent profit from it though, since they're positioning it towards professional and scientific applications with the high performance and price to match.

As AMD taught us gaming devices aren't high margin markets, and not even in Nvidia's dreams an auto-maker will pay high margins for what is essentially a commodity product.

If anything the focus on these markets means that Nvidia will retreat to small niches, but not high margin/high profits ones.

 

[Rakehellion]

So it scored worse than every other phone.

Most of that cost is the software to enable the encryption & security that are the selling points of that phone.

Isn't most of that free software anyway? I mean, every phone comes with an operating system, so it makes no sense to charge a premium for that.

 

[DeathReborn]

So it scored worse than every other phone.



Isn't most of that free software anyway? I mean, every phone comes with an operating system, so it makes no sense to charge a premium for that.

Services and subscriptions included in the phone (most if not all are 24 month) costs over $700 to us consumers, probably less for ODM/OEM's but that's a hell of a lot of software included that bumps up the price. Add to that the niche market that secure phones are in and you soon see how that's a bad device to point out Tegra 4i device pricing.

Yes there's not much choice but that's like only using the Vertu Signature Touch $11k phone as a example of Snapdragon S800 pricing.

 

[blackened23]

As AMD taught us gaming devices aren't high margin markets, and not even in Nvidia's dreams an auto-maker will pay high margins for what is essentially a commodity product.

If anything the focus on these markets means that Nvidia will retreat to small niches, but not high margin/high profits ones.

AMD gaming products? Shield didn't turn out too well, but i'd attribute that to the state of android gaming thus far. I do think K1 is faster in paper specs than prior generation consoles. Besides which, these companies are not alike. AMD doesn't know how to make money, nvidia know how to make money. AMD only knows marketing, while NV doesn't seem nearly as good at marketing, but marketing apparently doesn't make AMD money. If NV are positioning Tegra for a high profit area, i'd say they have a good shot at doing just that.

Granted, they did make a mistake by attempting to make their Tegra chips commodity type chips. They wanted to make it all things for all people, but the truth of the matter is that there are so many low end android devices where the market doesn't care good graphics performance. OTOH, the high end is dominated by qualcomm simply because of their modem IP, and Apple as well but Apple isn't a direct competitor since they ONLY run iOS (and not android). They learned from that mistake. The K1 will be aimed at completely different, profitable segments and is not limited to just gaming: Tegra K1 is already in quite a few very high end automobiles which is presumably a very high profit margin for NV. NV isn't bottom feeding for consumer level cars, they have designs in high end premium automobiles. Aside from that, they're making devices for professionals, developers, and scientific applications, which is uncharted territory thus far for ARM Type SOCs. How that pans out, I guess we'll see, but so far they have tons of automobile wins.

Since NV was successful with Tesla and Quadro, it'll be interesting to see if they can create a high end niche for ARM SOCs. Whereas if they continued on the path that they were on with T4, failure was a guarantee. It's practically impossible to compete with Qualcomm's modem IP, which is why everything (high end) uses Qualcomm nowadays. And the low end garbage bargain bin generally used mediatek chips. Who wants to compete with those companies? Not an awful lot to gain from that per JHH, so we'll see how they do.

But, as far as the AMD comparison goes, AMD is a not for profit organization, while NV actually delivers for their shareholders. So I don't get the comparison with AMD. NVidia's gaming products (Geforce) actually delivered very high profits in their last quarter, despite the retracting PC market. I don't think nvidia has a bad bet with android gaming. It isn't where it needs to be YET, but it is getting closer; simply look at the graphics performance jumps that happen with every iteration of new ARM SOCs. It really is unbelievable how much graphics performance has increased, I think android gaming could be a thing 4-5 years from now even for the living room - as of now though, android gaming is mostly limited to simple smartphone games. NV may not do amazingly well with the K1, but it's not because it's a poor product. T4i is a bust, sure, but the K1 is actually a very impressive SOC in terms of performance. Whether NV translates that into a meaningful money maker remains to be seen.

 

[waffleironhead]

Nvidia had a bunch of chips sitting around that no one was interested in, so they sold these off cheap with dev support thrown in to sweeten the deal.

 

[Idontcare]

The Blackphone was going to go with a different SoC, until NVidia threw a ton of free developer time at them in return for using a Tegra 4i... That kind of desperation is pretty telling.

Better to die trying than to die not trying at all. The market needs more competition, not less. If Nvidia bows out of the race we will have less competition.

I will say this though, all that hyperbole from Jensen about opening a can of whoop-ass on Intel sure fell flat in the biggest ways possible.

 

[Homeles]

Better to die trying than to die not trying at all. The market needs more competition, not less. If Nvidia bows out of the race we will have less competition.

I will say this though, all that hyperbole from Jensen about opening a can of whoop-ass on Intel sure fell flat in the biggest ways possible.

That's JHH for ya.

 

[xpea]

So it scored worse than every other phone.

comparing a mid range product aimed at sub $200 phones to the top SoC found in $500 phones and above is not very fair
now compare T4i with S400 and mediatek crap that are in same segment, then it becomes much better...

 

[ams23]

The Blackphone was going to go with a different SoC, until NVidia threw a ton of free developer time at them in return for using a Tegra 4i... That kind of desperation is pretty telling.

That is not what happened at all. Blackphone was going to go with an alternate SoC vendor (Qualcomm?) but that SoC vendor decided to not support Blackphone's smartphone project (probably due to limited quanties and specialized software tuning required), so NVIDIA was called upon to provide an SoC solution and they did to the great satisfaction of Blackphone.

 

[ams23]

As AMD taught us gaming devices aren't high margin markets, and not even in Nvidia's dreams an auto-maker will pay high margins for what is essentially a commodity product.

If anything the focus on these markets means that Nvidia will retreat to small niches, but not high margin/high profits ones.

Tegra processors are and will be used in a wide variety of different products (including tablets, high end smartphones, micro gaming consoles, chromebooks, high res monitors, high res smart TV's, automotive infotainment/navi/ADAS systems, robotics, military drones, etc) and have very small die size and transistor counts, relatively speaking, compared to something like an Xbox One APU, so that is a very poor point of comparison.

As for automotive, gross margins for Tegra automotive are higher than that for Tegra consumer business. Contrary to popular belief, low power and high throughput advanced visual computing capability in a car is not so easily commoditizeable, and that is one reason why NVIDIA has $2+ billion of Tegra automotive sales contracts signed with higher gross margins for the automotive VCM (visual computing module) compared to any Tegra consumer offering.

 

[Rakehellion]

comparing a mid range product aimed at sub $200 phones to the top SoC found in $500 phones and above is not very fair
now compare T4i with S400 and mediatek crap that are in same segment, then it becomes much better...

Well, it sure isn't a sub-$200 phone.

 

[ams23]

Well, it sure isn't a sub-$200 phone.

The Blackphone will cost >$600 USD regardless of SoC inside. The T4i-powered LG G2 Mini and Wiko Wax are much lower priced phones, relatively speaking. T4i is a very low cost SoC design with a very small SoC die size (~ 60 mm^2) with LTE modem built in on the die, so it has similar production cost compared to something like S400, but with higher performance.

 

[xpea]

Well, it sure isn't a sub-$200 phone.

Wiko Wax is at this price range:
http://blogs.nvidia.com/blog/2014/02/24/lte-europe/

The blackphone is a very very special device ; his price doesn't come from SoC cost but from custom software and delivered service

 

[Grooveriding]

As AMD taught us gaming devices aren't high margin markets, and not even in Nvidia's dreams an auto-maker will pay high margins for what is essentially a commodity product.

If anything the focus on these markets means that Nvidia will retreat to small niches, but not high margin/high profits ones.

Basically this. What nvidia said is they are giving up on the huge market smartphones and going for 'high-performance' niche tablets and phones. The translation was they are limping off after getting their butt kicked repeatedly.

I shook my head at the 'high-performance phone and tablet market' There is no such thing. Mobile users do not care or even know what is inside their tablets/phones for the most part. Those products aren't even marketed on their internals and performance figures. A 'high-performance' phone or tablet is an iPhone 5, HTC M8, iPad, Galaxy Tab 10.1 etc. And these products are all wrapped up by the big players already and are what nvidia couldn't get into.

The gaming device thing also rings hollow. Shield has been an abysmal failure with three price cuts so far. They keep dropping the price and it still doesn't sell. Not very promising for 'high-margin' gaming devices in the future. If Nintendo and Sony can't sell expensive hand-held gaming devices, nvidia sure can't.

 

[mrmt]

As for automotive, gross margins for Tegra automotive are higher than that for Tegra consumer business. Contrary to popular belief, low power and high throughput advanced visual computing capability in a car is not so easily commoditizeable, and that is one reason why NVIDIA has $2+ billion of Tegra automotive sales contracts signed with higher gross margins for the automotive VCM (visual computing module) compared to any Tegra consumer offering.

What Nvidia is telling you is that automotive margins will be higher than their margins, which is a true statement. What they didn't tell you is that their *current* margins on consumer Tegra is far below what they used to get with Tegra 2 and even sometimes Tegra 3 (when Tegra margins were ABOVE corporate average, as stated in a Q&A). In other words, margins on their consumer business are so low that an embedded business suddenly became more profitable than sell bleeding edge devices on the consumer market. This is the elephant in Nvidia's room.

And why do you think the kind of visual computing the manufacturers are putting in a car aren't commodities? What does Tegra brings to the table that Intel, Mediatek, Qualcomm and others won't be able to match in terms of hardware capabilities if they decide to exploit those same markets? A car is a much less constrained environment than, let's say, a tablet, and car software isn't above what an android tablet or windows convertible are going to offer you, and volumes are fine as long as you do not have Globalfoundries as your main foundry partner.... what are your constraints on this one? So yes, it is commoditizable.

 

[NTMBK]

There is some hope from Android TV. Google themselves are finally backing the microconsole concept, which could see a worthwhile niche for Tegra.

 

[NTMBK]

What Nvidia is telling you is that automotive margins will be higher than their margins, which is a true statement. What they didn't tell you is that their *current* margins on consumer Tegra is far below what they used to get with Tegra 2 and even sometimes Tegra 3 (when Tegra margins were ABOVE corporate average, as stated in a Q&A). In other words, margins on their consumer business are so low that an embedded business suddenly became more profitable than sell bleeding edge devices on the consumer market. This is the elephant in Nvidia's room.

And why do you think the kind of visual computing the manufacturers are putting in a car aren't commodities? What does Tegra brings to the table that Intel, Mediatek, Qualcomm and others won't be able to match in terms of hardware capabilities if they decide to exploit those same markets? A car is a much less constrained environment than, let's say, a tablet, and car software isn't above what an android tablet or windows convertible are going to offer you, and volumes are fine as long as you do not have Globalfoundries as your main foundry partner.... what are your constraints on this one? So yes, it is commoditizable.

Both Apple and Google are pushing thin client car systems powered by the driver's phone- given the mismatch between car and phone replacement rates, this makes a lot more sense if you don't want your in car system to look like outdated junk in 5 years. I just don't see that big a market for Tegra there.

 

[mrmt]

But, as far as the AMD comparison goes, AMD is a not for profit organization, while NV actually delivers for their shareholders. So I don't get the comparison with AMD. NVidia's gaming products (Geforce) actually delivered very high profits in their last quarter, despite the retracting PC market. I don't think nvidia has a bad bet with android gaming. It isn't where it needs to be YET, but it is getting closer; simply look at the graphics performance jumps that happen with every iteration of new ARM SOCs. It really is unbelievable how much graphics performance has increased, I think android gaming could be a thing 4-5 years from now even for the living room - as of now though, android gaming is mostly limited to simple smartphone games. NV may not do amazingly well with the K1, but it's not because it's a poor product. T4i is a bust, sure, but the K1 is actually a very impressive SOC in terms of performance. Whether NV translates that into a meaningful money maker remains to be seen.

Nvidia might be good at strategy but they can't really go against the market on this one, as they do not have the leverage for anything. The embedded business is low margin, or at least lower margin than the consumer business, albeit you can make up with that on volumes and longer product time frame. Had the embedded car market be more profitable than the consumer market, we would be seeing IHVs fighting to death in order to achieve deals with BMW, Ford and others, and we would see IHVs debuting their latest and greatest on luxury cars. And since Nvidia wanted a high margin business, we would see them starting their mobile venture on cars, not on tablets. And yet what we see in reality is far from what I'm describing here. Nvidia is going after cars just after FAILING on the consumer mobile market.

That said, I think that also answers your question on whether Nvidia will make money with K1. They will, but it won't be as much as they would if the bulk of K1's sales came from the consumer market.

As for gaming... Android gaming will take off once Google starts to rein in the Android wild west. The cannot take the Apple's route and go with something like Metal, but they can take the Microsoft route and build an Android DirectX. Before that, we won't have a descent Android gaming ecosystem. This process is far beyond Nvidia's abilities as far as the leadership of it is concerned, and others like Qualcomm, ARM and Mediatek will have much more influence to influence.

 

[mrmt]

Both Apple and Google are pushing thin client car systems powered by the driver's phone- given the mismatch between car and phone replacement rates, this makes a lot more sense if you don't want your in car system to look like outdated junk in 5 years. I just don't see that big a market for Tegra there.

Your car is an outdated junk after 5 years from an automaker perspective. They will probably want to sell you something else before 5 years.

IIRC that was Charlie Demerjian argument in some of his Nvidia's bashings... and I think he completely misread the market trend. What the automakers are trying to build is a bridge between the car and Apple/Android devices, in order to easily plug some of their capabilities on the car. But... What if you forgot your phone on the office or someone stole it, you are suddenly back to the 90's without GPS, without music, without electronic controls of anything of your car? Not likely. The automakers still need *some* CPU power native to the car.

The kind of integration might in fact hinder the development of the mobile embedded market (why would I put some advanced capabilities when someone can just plug their phones and have it?), but there will be need for CPU power in your car. 0 CPU power won't cut it, especially because the embedded chips are cheap and the infrastructure will have to be in place for phone integration to work.

 

[NTMBK]

Your car is an outdated junk after 5 years from an automaker perspective. They will probably want to sell you something else before 5 years.

IIRC that was Charlie Demerjian argument in some of his Nvidia's bashings... and I think he completely misread the market trend. What the automakers are trying to build is a bridge between the car and Apple/Android devices, in order to easily plug some of their capabilities on the car. But... What if you forgot your phone on the office or someone stole it, you are suddenly back to the 90's without GPS, without music, without electronic controls of anything of your car? Not likely. The automakers still need *some* CPU power native to the car.

The kind of integration might in fact hinder the development of the mobile embedded market (why would I put some advanced capabilities when someone can just plug their phones and have it?), but there will be need for CPU power in your car. 0 CPU power won't cut it, especially because the embedded chips are cheap and the infrastructure will have to be in place for phone integration to work.

Oh it will need an embedded SoC to drive the display and decode, yes, but what do you think the margins are on an A53 core and a h.264 decoder?