Intel Enters into Strategic Agreement with Rockchip

[kimmel]

http://finance.yahoo.com/news/intel-enters-strategic-agreement-rockchip-160000727.html

Intel Corporation today announced it has entered into a strategic agreement with Rockchip to expand the breadth of and accelerate the rate at which it brings its Intel architecture and communications-based solutions to market for a range of entry-level Android* tablets worldwide.
Under the terms of the agreement, the two companies will deliver an Intel-branded mobile SoC platform. The quad-core platform will be based on an Intel Atom core integrated with Intel’s 3G modem technology.

Well that's interesting. Not really sure what it really means though.

 

[Vesku]

Sounds like Rockchip will be providing the distribution and support for an Intel SoC. That or Intel has expanded their Atom brand to cover a Rockchip ARM SoC.

 

[witeken]

Intel won't be using ARM.

 

[Homeles]

More information here:
http://www.anandtech.com/show/8061/...gic-agreement-with-chinese-soc-maker-rockchip

 

[witeken]

Seems quite interesting, certainly the quadcore Silvermont SoFIA solution, but I don't really see the value of this. Why would this be helpful for Intel? As I see it, Intel's using Rockship to get its x86 IA faster into mobile, which is of course nice, but I'm not sure if there any other reasons.

 

[Roland00Address]

So is this sofia but remarketed another another name?

If it is not sofia than how is it different than sofia?

If this is just using rockchip so you can get time to market faster than you making the modem yourself why then did you spend 1.4 billion to get Infineon wireless. Sure this is not major money for a company like intel but that is still billion with a b.

I want to understand but I do not, so someone with more information please explain at an intermediate level.

 

[IntelUser2000]

If this is just using rockchip so you can get time to market faster than you making the modem yourself why then did you spend 1.4 billion to get Infineon wireless. Sure this is not major money for a company like intel but that is still billion with a b.

In the future, there will be a higher variant with their own LTE as well. The SoFIA chips are very low cost parts.

 

[witeken]

If this is just using rockchip so you can get time to market faster than you making the modem yourself why then did you spend 1.4 billion to get Infineon wireless. Sure this is not major money for a company like intel but that is still billion with a b.

The Infineon acquisition isn't a waste of money. Without it, the claims of certain people that Intel will never get anywhere in mobile would be valid.

 

[cbn]

If this is just using rockchip so you can get time to market faster than you making the modem yourself why then did you spend 1.4 billion to get Infineon wireless. Sure this is not major money for a company like intel but that is still billion with a b.

In the future, there will be a higher variant with their own LTE as well. The SoFIA chips are very low cost parts.

Yep, the Intel SoFIA for 2015 will be a quad core part with LTE. Rochchip's SoFIA for 2015 will be a quad core part with 3G.

This will allow Rochchip to handle a lower part of the market that Intel might otherwise miss.

 

[IntelUser2000]

cbn: You are right, but I am not sure of their LTE plans, so I was talking about the Broxton and the successors. It doesn't look like Broxton has integrated LTE, but the one after that might. And Broxton still needs discrete LTE.

 

[Roland00Address]

So if I understand everything this is Intel throwing everything at the wall, all the wet spaghetti hopping something sticks. Sure you may be wasting money, sure you won't get the best profit maximization but if nothing sticks you are a dead company walking? Is that what everyone is saying?

If so I can understand that, but the reason why I ask is that I don't see how this adds even 3 months to time to market based off the info on roadmaps I seen. Maybe I am wrong, but I just don't see how one piece of wet spaghetti will stay up if the other piece falls down.

 

[bullzz]

looks like anand has updated his article

"In return, Rockchip becomes the first company to be able to more or less license Intel x86 CPU IP for use in this combined, Intel-branded SoC. The low-cost ARM based SoC market is crowded and it's tough to differentiate when all of your competitors have access to the same ARM CPU IP. In this case, Rockchip gets access to Silvermont which it may be able to use to set itself apart from the competition. There's obviously a tradeoff in shipping x86 into the Android space, but if Rockchip can help Intel get its numbers up the x86 problem could reduce over time."

so intel is licensing its CPU IP. interesting

 

[theeedude]

Keywords being "Intel-branded".
This is not Intel following ARM model of licensing its CPU core to other companies to do as they see fit.
It's more of Intel outsourcing their SOC design and distribution to Rockchip while retaining control, so that Rockchip doesn't use this chip to compete with Intel for Windows sockets.
This is why I don't think this model is going to be successful against ARM. Same reason central planning isn't successful relative to free markets. It's too restrictive, and lacks the incentives and competition that ARM ecosystem has.

 

[Homeles]

looks like anand has updated his article

"In return, Rockchip becomes the first company to be able to more or less license Intel x86 CPU IP for use in this combined, Intel-branded SoC. The low-cost ARM based SoC market is crowded and it's tough to differentiate when all of your competitors have access to the same ARM CPU IP. In this case, Rockchip gets access to Silvermont which it may be able to use to set itself apart from the competition. There's obviously a tradeoff in shipping x86 into the Android space, but if Rockchip can help Intel get its numbers up the x86 problem could reduce over time."

so intel is licensing its CPU IP. interesting

Intel's been trying to "license" Atom for a while now.

It's too restrictive, and lacks the incentives and competition that ARM ecosystem has.

Regardless of long term viability, if Intel's got the performance advantage, you might as well take it. If Intel does end up leaving ARM's performance in the dust with Airmont, Rockchip would benefit greatly.

 

[IntelUser2000]

so that Rockchip doesn't use this chip to compete with Intel for Windows sockets.

They can. ALL Atom chips in the future will have Windows and Android variants, helped by recent Microsoft announcement that Windows+Office under 8-inch or smaller devices will be essentially free. Besides, the cost will be in a category of its own. Cherry Trail will take the mid and high-end.

Windows has $99 devices in sight, and obviously low cost variants like Rockchip is going to help Intel will fit in there.

Of course, they are not trying to kill ARM, just as they are not trying to AMD. But you need to break the barrier to establish a foothold.

 

[TrulyUncouth]

I wonder if this isn't just a clever way to keep their ASP number higher on paper while continuing what is in essence contra-revenue. Let Rockchip look like they are selling really cheap android chips and keep it off intel's books. Maybe slip Rockchip's payments to intel in a simple line item so they look like they are still selling chips at a high price?

I am massively under-qualified to say if any of that is accurate- just was my first thought on seeing the story and was hoping someone could debunk or give their opinion.

 

[Khato]

Quite an interesting move on Intel's part as it sounds like a direct competitive offering to ARM's core designs... with the added bonus of having communications IP to throw in as well. I can easily see how it makes sense for Intel in the near-term - customers that want an x86 chip can either buy the latest design and process technology directly from Intel or they can get the budget chip from Rockchip. The advantage for Intel being that they can license the IP for a trivial cost (or maybe even free?) and increase x86 presence in the budget market without impacting their margins. And then I'd guess that Rockchip is going along with it due to pricing - I doubt Intel cares much about revenue from this venture since all the components they're offering are re-use anyway.

 

[Idontcare]

I wonder if this isn't just a clever way to keep their ASP number higher on paper while continuing what is in essence contra-revenue. Let Rockchip look like they are selling really cheap android chips and keep it off intel's books. Maybe slip Rockchip's payments to intel in a simple line item so they look like they are still selling chips at a high price?

I am massively under-qualified to say if any of that is accurate- just was my first thought on seeing the story and was hoping someone could debunk or give their opinion.

Doubtful. This is more about Intel needing a success story for their marketing team to then go and bang on the other holdout OEMs.

All Intel needs is for Rockchip to start taking up marketshare in China and the other OEMs will start getting worried they are about to miss the boat, will jump on board the Intel bus for fear of losing out on the 10nm generation of products. (as they probably should be)

This is all business strategy with a 2-3yr horizon, nothing to do with fluffing up the 2014 or 2015 financials IMO.

 

[cbn]

I am pretty sure this tablet is powered by Rockchip:

http://www.walmart.com/ip/HKC-8-Tabl...-Core/23709346

1.5GHz Dual-Core Mobile processor
1GB DDR2 of system memory and 8GB on-board storage memory
8.0" touchscreen, 1024 x 768 resolution
Built-in 802.11b/g/n WiFi
Webcam
Mini-USB 2.0 port
microSD card slot
Android 4.1

Not a bad spec at for something that costs $69.99 FS at Walmart.

So I wonder at least two things:

1. At what point does Rockchip have the incentive to start making the budget dual core x86 SOCs.

2. Do we actually start seeing $70+ Windows Tablets now that MS is giving away Windows on tablets and phones with 9" and smaller screens? (Keep in mind Windows minimum requirement is 16 GB eMMMC.....so that would be a small cost adder compared to the specs of the tablet above)

I would so like to have two things:

1. A very inexpensive Windows x86 tablet from Walmart.

2. A Windows x86 trac fone (yes, I wrote Trac fone) that I could dock as a desktop if I wanted to

 

[shady28]

This is what I found most interesting about the deal, from AT :

"Intel wouldn't go into specifics on how the arrangement works, other than to say that Intel would work with Rockchip to do the SoC integration and Rockchip will bring its own IP to the table as well. Intel gave the example of Rockchip bringing 3rd party graphics IP to the SoC."

http://www.anandtech.com/show/8061/...gic-agreement-with-chinese-soc-maker-rockchip


So.. it appears Rockchip is able to customize the design at least to some degree in the same way they can add on or take away from a core ARM design. How much, we don't know.

Now assuming the above statement is true, then this IMO is a huge change for Intel.

The major advantage of ARM has been the ability of ARM licensees to customize the design, adding special "parts" to the SoC to differentiate themselves in one way or another. I am fairly certain that this is one of the reasons why ARM chips seem so well suited to mobile - they're all optimized for that environment, nevermind the general use computing model b/c who cares about that anymore.

It looks to me like Intel is taking a very tepid step in that direction in order to grow market share.

This one will really be worth keeping an eye on. I think its success depends a lot on how much RockChip trusts Intel, and how much freedom Intel is willing to give this licensee. Rockchip is pretty large in China, if they can successfully one-up competitors like MediaTek with a Silvermont / Airmont core + heterogenous compute items tailored for Android, it could truly be a game changer.

 

[cbn]

if they can successfully one-up competitors like MediaTek with a Silvermont / Airmont core + heterogenous compute items tailored for Android, it could truly be a game changer.

From what I understand, it does make a lot of sense for Intel to get the lowest common denominator on Android. (Apparently programming in the native machine language helps performance in some cases.)

However, with that mentioned ARM does have even smaller cores (Cortex A5, Cortex A7, etc) for that. Still Rockchip is known for making very affordable (actually I think it is safe to say they are the most affordable) SOC.....so maybe this doesn't matter that much for all practical purposes.

P.S. Here is something to think about though:

http://www.anandtech.com/show/7993/...hones-in-2018-20-smartphones-coming-this-year

ARM believes the floor for an entry level smartphone running Android (today with a single-core Cortex A5 based SoC) is $20, and that we'll see the first devices on sale at that price point in the next few months. Manufacturing limits will likely prevent cost scaling below $20. Keep in mind that a single Cortex A5 is faster than the ARM11 in the original iPhone, which retailed for $599 in 2007.

Currently the lowest priced Android Trac fone with a single Cortex A5 processor I know of is $49.99:

http://www.bestbuy.com/site/tracfone...1-000019adfa50

 

[Exophase]

so intel is licensing its CPU IP. interesting

Here's what I think is happening: Intel is licensing their baseband processor IP, but you only get it as a package deal with their CPU.

I don't think anyone else is licensing cellular modems, this could be Rockchip's only opportunity. I don't think they've ever had any market share in phones. They used to sell a lot of PMPs and for a while had a big market share in Chinese tablets, but that share was heavily eroded by other Chinese SoC makers like Allwinner, as well as MediaTek, and Qualcomm at the higher end. 3G will help their position in phones and to some extent tablets. AFAIK SoFIA is also supposed to have integrated wifi, BT, and GPS, so there's a good chance this chip will too. In the tablet space that's actually a pretty big deal and could take a lot back from Allwinner.

Will be curious to see what GPU IP Rockchip uses here. They've used ARM Mali pretty much exclusively for a while now. Would be pretty funny to see an ARM GPU paired with an Intel CPU...

Currently the lowest priced Android Trac fone with a single Cortex A5 processor I know of is $49.99

Those single core Cortex-A5 SoCs are an abomination (and some of them are even only 800MHz). They've set back the minimum specs on Android to a level from seven years ago. All to save a dollar vs an SoC that's a few times faster.

 

[Shivansps]

this is the bolt part for me

Intel wouldn't go into specifics on how the arrangement works, other than to say that Intel would work with Rockchip to do the SoC integration and Rockchip will bring its own IP to the table as well. Intel gave the example of Rockchip bringing 3rd party graphics IP to the SoC. Rockchip's existing products use ARM and Vivante GPU cores, so we may end up seeing an SoC that uses Intel x86 cores with Mali graphics.

Rockchips = Silvermont + Malis-760?

 

[cbn]

Those single core Cortex-A5 SoCs are an abomination (and some of them are even only 800MHz). They've set back the minimum specs on Android to a level from seven years ago. All to save a dollar vs an SoC that's a few times faster.

According to the ZTE website the Valet uses a 1 Ghz Qualcomm MSM7625A (single core cortex A5) with 512 MB RAM and 4GB eMMC:

http://www.zteusa.com/phones/zte-valet.html

So I guess this will be my benchmark to compare Intel and Rockchip's efforts against at the lowest end.

Intel/Rockchip, can you give me a better extreme budget smart phone than Qualcomm?

 

[Arachnotronic]

My bet is that ImgTec gets the GPU IP slot for all SoFIA products.