Project Denver

[Cerb]

Tegra 3 has the whole Android tablet market, is one of the reference design of Windows RT and has a few smartphone wins.

And a year later, then what?

That the competition is better after a year is normal in this business.

So, it's a year later, now what do you have? Qualcomm has a whole line-up, with mixes of old and new CPU and GPU designs, FI. Most people I know with smart phones either have a Qualcomm-powered one, or an iPhone, and not the higher-end Qualcomms, either. The MSM7227, FI, has quite a showing.

I'm not dissing nVidia; just pointing out that they need to find a way to compete, over the course of years, and it will be in their best interests to not directly follow what their competition is doing, nor directly follow ARM. They need to have something--or things--that can't just be easily one-upped by the next new chip coming out in several months. Right now, their GPU tech is most of that. Once they add more to their chips, they'll still want something more than just a lego set SoC, so that when the performance gets one-upped, they can fall back on more than an ability to reduce prices (that will just lead to a race to the bottom, which isn't a very JHH thing to do). I think customized CPUs, to improve performance/efficiency and GPU integration, could be just such a feature set in the coming years, and it would fit well with nVidia's existing technological and marketing strengths.

 

[joshhedge]

What? First products with Tegra 2 were out in August 2010. At the same time you got only Samsungs Single-Core A8 SoC with SGX540.
Tegra 3 has the whole Android tablet market, is one of the reference design of Windows RT and has a few smartphone wins. That the competition is better after a year is normal in this business.



Tegra 2 was not only the first Dual-Core ARM design it used as the first SoC the A9 processor. Tegra 3 is not only a Quad-Core ARM design, it has a low-power A9 processor and the best video processor in the market.

Next year they will release a DualCore A9 Design with an integrated radio modem, which put them on par with Qualcomm. They can beef up their SoC dies size in a huge way to increase the GPU performance.

I would more worry about their time to market than their SoCs.

"and the best video processor in the market" You're wrong. Very wrong. I cannot be bothered to link graphs but the A6X/A6 Apple chip has the best GPU in the mobile market.

 

[sontin]

And a year later, then what?
So, it's a year later, now what do you have? Qualcomm has a whole line-up, with mixes of old and new CPU and GPU designs, FI. Most people I know with smart phones either have a Qualcomm-powered one, or an iPhone, and not the higher-end Qualcomms, either. The MSM7227, FI, has quite a showing.

A year later nVidia forced TI nearly out of the market. Why do you compare nVidia with Qualcomm? They have a huge advantage because of their radio modem IP.

I'm not dissing nVidia; just pointing out that they need to find a way to compete, over the course of years, and it will be in their best interests to not directly follow what their competition is doing, nor directly follow ARM.

nVidia was not following. They were leading the whole market with Tegra 2 and Tegra 3:
Tegra 2 was the first DualCore ARM design and the first SoC with A9.
Tegra 3 was with Sony the first QuadCore ARM design and Tegra 3 was the first little.BIG approach which you can buy.
The video processing unit (video decoder) is able to decode 1080p@60FPS with 40MBits/s and High Profile. That means you can play Blu-Ray rips of your own disks without encoding them.

We will see what they will do with their first 28nm SoC but until them nVidia is not the company which is following others.

 

[nismotigerwvu]

It's also that there was no SoC Tegra wireless or 3G/4G solution until very recently. Otherwise we'd have seen many more mobile devices with Tegra 2 out there. Now they have it and Tegra 3 is rapidly affecting it's adoption.

Ahh very true, 1 chip solutions will bring better margins and battery life.

 

[Nemesis 1]

Browsermark:

Quadrant CPU-Benchmark:

And GLBenchmark 2.0:

http://www.anandtech.com/show/4144/...gra-2-review-the-first-dual-core-smartphone/8

Tegra 2 was a beast in 2010 and early 2011. The problem was that the first buyable product was LG's 2X smartphone. And then they got beaten by Apple, Samsung and TI with their DualCore products in the second half of 2011.

Ok I see what your saying . Do you have link to how many sales that phone generated. I mean tegra 2 was consider a fail. It was a learning curve for NV . Just as intels medfield are overpowering for a 1 core 2 thread cpu . Its sales won't be great but it will outsale the tegra II by along ways. But not the tegra III. THe next phone from Intel 22nm will change the market as to who has the real performance and efficiency, When AMD arms chips appear with the mighty AMD gpu all others can just leave the market' LOL

I think NV is going to do fine in this market .. AMD with arm not so much. TO late to little .. Even intel is ahead of AMD in graphics in this areana

 

[Nemesis 1]

--------------------------------------------------------------------------------

Quote:
Originally Posted by sontin
Tegra 3 has the whole Android tablet market, is one of the reference design of Windows RT and has a few smartphone wins.


If your going to put something like this in a post finish the post . Intel clover trail will be up against these chips . and let me tell ya . Intel slaps the tegraIII around big time.
AMDS jag will be interesting but a power hog that won't effect the market at all..

 

[Siberian]

A full ARM cpu will be the biggest threat to x86 ever.

 

[Blitzvogel]

A full ARM cpu will be the biggest threat to x86 ever.

Packed into a fully featured desktop computing environment, perhaps. That has yet to happen, and servers are a big step in moving that direction much like with the x86 ecosystem where servers and consumer processors share technology.