Intel Announces Atom x3, x5 and x7, First SOCs With Integrated 3G and LTE Modems
MojoKid writes Intel is unleashing a new family of Atom processors today, taking a cue from its highly successful Core series with model branding. Similar to the Good, Better, Best strategy with the Core i3, i5 and i7, Intel is renaming its Atom family with x3, x5, and x7 designations. The biggest news comes from the low-end Atom x3, which will be available in three distinct variants; all of which will come with integrated modems — a first for the Atom family. All three variants are 64-bit capable cores. The Atom x3-C3130 tops out at 1GHz, incorporates a Mali 400 MP2 GPU, and includes an integrated 3G (HSPA+) modem. The Atom x3-C3230RK bumps the max clock speed to 1.2GHz, throws in a Mali 450 MP4 GPU, and the same 3G modem. Finally, the Atom x3-C3440 clocks in at 1.4GHz, features a Mali T720 MP2 graphics core, incorporates a Category 6 LTE modem, and can optionally support NFC. Using handpicked benchmarks, Intel claims that the Atom x3-C3230RK can offer up to 1.8x the media editing performance of competing SoCs from Qualcomm and MediaTek. Then there's Intel's Cherry Trail-based Atom x5 and x7. These are the first 64-bit Atom SoCs to be built using a 14nm manufacturing and they incorporate eighth generation Intel graphics. While the Atom x5 and x7 don't feature integrated modems like the Atom x3, they do support Intel's next generation XMM 726x and 7360 LTE modems. Intel claims that the Atom x7 offers two times the graphics performance of the existing high-end Atom Z3795 in the GFXBench 2.7 T-Rex HD benchmark and 50 percent greater performance on the 3DMark Ice Storm Unlimited benchmark.
As the subject says, this is one way to get into the mobile market or cement a position in it, assuming the modem is high performance, good quality, good support, like their other networking products.
That or it's a way for three-letter government agencies to spy on users of detachable laptop computers, using a built-in hardwired SIM card even when the user thinks the cellular radio is off.
Probably premature to say this, but it would be funny if Intel does to mobile processing and ARM what it did to Mac computers and RISC. For a long time the Mac-heads were constantly harping about how superior PowerPC was to anything in the Wintel world. And then suddenly everything was x86 again. It seems x86 is the technology that can't be killed.
The X3 line is very weak, and will be competing against $5 to $10 SoCs from MediaTek, AllWinner, etc. This market is very price sensitive, and battery life is also important.
The X5 and X7 look more capable, it will be interesting to see how they compare against the competitor SoCs using A57 cores. The 14nm process will also help with the battery life significantly.
BMW has anything to say about that
Sweet, so now we can have portable space heaters in our pockets?
MediaTek and AllWinner don't have integrated modems. Intel is aiming squarely at Qualcomm and Samsung with these chips.
“Common sense is not so common.” — Voltaire
Weak?
These chips don't even compete in the same markets.
The intel offerings easily offer 5-10x computational power over the budget-minded SoCs from the likes of AllWinner. The /previous/ gen baytrail based Atom SoCs can run a full-fat windows 8.1 installation in a tablet formfactor. A real tablet. USB charging, thin, touch screen fanless device with no BS tablet battery life. I've got two. (Dell Venu 8 pro and Asus transformerbook T100)
I'm not saying that Allwinner and company don't sell a lot of chips, but those devices mostly go in to ultra-budget android devices, "stick" computers, and media set top boxes. They're also really really really poorly supported and are generally useless for anything other than the build of Android that comes shipped with them.
The Intel offerings, however, are standard architecture and can run whatever OS you want. Android, Windows, Linux, or even iOS or MacOS if Apple cared. (Don't fool yourself. There's a skunkworks x64 build of iOS somewhere int the depths of Cupertino, waiting in the event that Intel throws its literally 2 generations superior chip fabrication technology full on in to the mobile chip market. These new Atoms are EXACTLY the sort of thing that would tempt Apple)
Until the price and the modem support and performances are verified, it's too early to pretend that integrated modem is an advantage. On the SoC market Intel have for years making big press release of chips that vanished into insignificant niche market compared to the SoC leading chips that massively use ARM cores.
Don't fool yourself. There's a skunkworks x64 build of iOS somewhere int the depths of Cupertino, waiting in the event that Intel throws its literally 2 generations superior chip fabrication technology full on in to the mobile chip market. These new Atoms are EXACTLY the sort of thing that would tempt Apple
Except your claim doesn't mesh with all their hiring and purchases related to custom chip design. They aren't going to throw that away for a power-hungry SoC with a shittier GPU.
Well I work for an actual company that offers integrated modems. And the silicon price is basically the same. The additional cost we add to the chip is to pay for the R&D investment.
A big company like Intel can soak a lot of R&D costs initially if they wish to make a long term play into the market.
I have no doubt that the Atom X3 is going to make it cheaper to put an x86 into a LTE capable tablet/phone. And Intel gets to get paid for the modem instead of a third party, so it's a big advantage for them.
The only barrier I see at this point is if their modem's performance is good enough to compete with Qualcomm. I'm familiar with other vendors that failed to take over the mobile market with wireless integration.
“Common sense is not so common.” — Voltaire
Incorporating Mali GPUs is bound to piss off the OSS crowd - they tried that before with PowerVR before, and those chips were the bane of any nettop user. They should have tried to slim down their own GT chips.
I would love to be able to install my x86 desktop aps on my mobile devices, even on my phone! This is all I have ever really wanted from mobile devices.
Sure.. a lot of them will suck cramped into a small screen and a touch interface they were never designed for. But.. they would not be totally unusable when I really need them. If they weren't then nobody would bother with vnc or remote deskop on their phones!
If such an environment was common then new versions of applications would be developed to scale well to both small and large screens and to work well with both keyboard/mouse and touch interfaces. Proving it is possible I have seen some Android devices that do this really well on a lapdock when they switch between their phone and tablet modes. In the tablet mode they don't just blow up the same view, they have different layouts with additional controls enabling more desktop-like features.
This is how mobile computing should have played out in the first place. Blackberry and Symbian plus the later iOS and Android have been only a poor imitation of what could have been and have slowed down mobile development as all the wheels have had to be re-invented.
Over the weekend I had an opportunity to play with a friend's Intel Windows 8 tablet. It was nice using real, full featured applications for once. An x86 Windows phone would definitely tempt me away from Android although I would much prefer a Linux (as in real Linux desktop) phone. That did almost exist once. There was an alternative desktop named GPE which ran on old Zaurus and HP PDAs back in the day. It used GTK on top of X. Qt was available as a separate package. So.. pretty much any Linux software, if you had the source it could be built! It even had a phone dialer although I don't know of any hardware that the dialer could be functional on. Unfortunately the cross-compiler was nearly required a PHD to get working and there weren't enough tools included to build packages on the device itself. It would have been a mini-desktop if it weren't for that.
Maybe someday....
Oh.. also on my wishlist.. these desktop-software running mobile devices should have HDMI out and USB host. When it does everything a desktop does there is no reason we shouldn't be able to plug it in and use it like one when we are not on the go.
Additionally... there is no reason it shouldn't work with something like the Motorola Lapdock. Except.. there should be some upgrades on that device which Motorola never offered. How about a touch screen? Better speakers? And.. why not make that keyboard removable. Then it becomes just a bigger screen for your phone.. your phone is a tablet.. is a laptop. Of course you could still have the HD docking station so it's a desktop too! Unless you want an always on-server then your phone is the only computer you need!
This is how it should be. What we have now sucks in comparison.
First x86 SOC perhaps..
Is it common to make new chips that only support 3G now? Being HSPA+ also means they won't work for Verizon/Sprint or anyone else who zagged when the world zigged. I'm guessing those are intended for light use appliances? E-Readers level of downloading maybe? It just feels weird that they wouldn't squeeze in the LTE modem while they're at it. I guess maybe Intel built these to budget and simply ran out of silicon? Are they going to be like the early generation GMA adapters that were just total crap and were only there because Intel had a little extra silicon available on each chip?
I read the internet for the articles.
I've briefly checked the specs and I must say that they are hilarious :) :( ! For the god's sake, these procs are supposed to allow Intel to (finally) target the mobile market! IMHO they didn't made a big show out of it because well, they didn't had what to show...
- They are announcing socs with Mali 400 and mali 450 this year ?!?!
- I was keen to know more about x5 and x7 CPU/GPU performance and consumption, but, incredibly, Intel had no live video from MWC
Apple plays the long game.
Their entire mobile product line relies on ARM chips they develop (Remember who's shipped the first 64 bit arm cpu in a production product when no one else was even sampling one?) Of course they're going to continue to develop Arm in the near term.
Keeping a quiet team working on an intel port of iOS would cost next to nothing in comparison and would reap HUGE benefits if Intel manages to upset the market. They would, frankly, be stupid not to.
They did it once before, with OSX. Apple saw the writing on the wall years in advance and started working on their port long before Intel released the Core2Duo, the chip line that put Intel in to the dominant position they remain today.
Your Apple speculative theory is interesting but completely unproved at this stage. Or did you have some information to share ?
I have no doubt that the Atom X3 is going to make it cheaper to put an x86 into a LTE capable tablet/phone. And Intel gets to get paid for the modem instead of a third party, so it's a big advantage for them.
Not really, the X3s are all made with third party GPU and modem functionality at TSMC. It's a bought design where they add a CPU and a brand to pretend they're competing in a market they're really not. The X5/X7s are Intel's homegrown solution with their own graphics and LTE modem and aimed only at the premium segment. You will not get Intel tech for cheap.
Live today, because you never know what tomorrow brings
The Neo900 project is creating a cell phone where they deliberately chose a CPU that does not have a modem built in. They keep the modem on a separate chip, and have a watchdog chip that monitors the modem for activity when there should not be (such as when the modem is supposed to be disabled). The monitor chip can instantly reset the modem when it is mis-behaving and alert the user.
Why? These days the modem has access to the same memory as the application processor. You no longer need backdoor coorporation with the OS, the modem can do whatever it wants.
I wonder why Intel have to buy an external design to bring the X3 chips on the market. What prevent them to cut down the X5 into a X3 ?
AllWinner and MediaTek are made with 3rd party everything. The only thing they do is the glue.
NVIDIA makes GPU and sometimes CPU, but most of the Tegra's license the CPU from ARM.
Tegras aren't cheap either.
Intel can charge pretty much anything they want to because they don't need to make a profit up front. That is the case with the current mobile Atom processors like the Z3740. Everyone in the industry is certain Intel is charging below cost, maybe only charging the manufacturing cost (assuming very high yields) and writing off all the operation costs.
tl;dr - Intel isn't out of the game yet.
“Common sense is not so common.” — Voltaire
It just takes a dev or two building and testing the code on x86 just to make sure its cross platform.
Such an effort can pay for itself simply through the bugs it will find. What is a very subtle and sometimes currently unnoticed bug on one platform may become a highly visible bug on another platform. Bugs sometimes manifest differently on different platforms. I've been on several teams over time where we moved good "working" code to a new platform and watched it crash spectacularly over and over again. Each time the original devs who thought their code was in great shape were shaking their heads wonder how it ever ran on the original platform. Lucky values in uninitialized variables and such.
So yes, someone is probably building and testing iOS on x86 but it has little to nothing to do with any plans regarding using x86 on any devices. Sort of similar to the Microsoft's efforts to internally build and test Windows on a non-x86 platform after it gave up on shipping MIPS, PowerPC and Alpha binaries. Its more about testing and future proofing a core asset than any short term plans.
And if I were running Intel I would be supplying the engineers to do so if Apple was not doing it on their own initiative.
Apple plays the long game.
Yeah, which is why they are all-in for custom chip design not some shitty Intel SoC.
Everyone in the industry is certain Intel is charging below cost, maybe only charging the manufacturing cost (assuming very high yields) and writing off all the operation costs.
Intel is overcharging there products, just look at the Intel annual benefice. If it's true that the Z3740 is charged below cost, it's not only a bad sign of weakness but probably an illegal way to gain market.
Intel is overcharging there products, just look at the Intel annual benefice.
Could you please explain what this means, I am not able to parse it.
“Common sense is not so common.” — Voltaire
It is well known that the baseband radios for 3g and 4g phones have direct memory access and easily own those devices. Putting this "feature" on a CPU die is effectively creating an out of band backdoor on every chip.......
So yes, someone is probably building and testing iOS on x86
Yeah, Darwin already exists for x86 and x86_64.
I'm referring to much more than Darwin. Possibly iOS in its entirety. At a minimum various iOS specific frameworks in a regression test environment. Including some hardware specific frameworks where inputs are simulated.
Wait, speaking of "simulated", we have the Mac OS X based iOS device simulators used during development. Doh!
Apple plays the long game.
Yes, the "long game" being that they want to use their own custom chips.
Their entire mobile product line relies on ARM chips they develop (Remember who's shipped the first 64 bit arm cpu in a production product when no one else was even sampling one?) Of course they're going to continue to develop Arm in the near term.
Why would they be spending hundreds of millions on acquistions and billions on R&D for their custom chips just to throw all that away for some Intel chip that has yet to appear? Apple has already long since proved that they don't need Intel in either the iPhone or iPad.
Keeping a quiet team working on an intel port of iOS would cost next to nothing in comparison and would reap HUGE benefits if Intel manages to upset the market. They would, frankly, be stupid not to.
iOS already runs on x86(_64). It's called the "iOS Simulator" (previously the iPhone Simulator). There isn't any need for a "quiet team" since it's all out in the public and has been part of the iOS SDK for going on 6 years.
I'm referring to much more than Darwin. Possibly iOS in its entirety. At a minimum various iOS specific frameworks in a regression test environment. Including some hardware specific frameworks where inputs are simulated.
Wait, speaking of "simulated", we have the Mac OS X based iOS device simulators used during development. Doh!
{ Oops, forgot to log in, didn't mean to post AC }
The long game is that they are not going to bet their whole existence on their being able to keep ahead of the performance curve.
They will be testing the Atoms and if the Atoms happen to produce a better power and performance package, you can bet that they will flip over to Intel.
One of the advantages they have is that they already differentiate the binaries by device, so it's not a stretch for them to recompile all submitted code right away and have it working on an x86 iPhone if they need it to. Seamlessly too as far as customers are concerned.
Intel was forced to "invest and partner' in China, or face same sanctions Qualcomm did. They decided to throw China a bone in form of $1B and license for lowest performance Atom CPU cores.
X3 will be made 100% by chinese 'partners', if at all - previously Intel dropped ATOMs $40 sticker price down to $4 Allwinter 4core level and there still were almost no takers.
Who logs in to gdm? Not I, said the duck.
The long game is that they are not going to bet their whole existence on their being able to keep ahead of the performance curve.
Says you. Their actions say otherwise.
They will be testing the Atoms and if the Atoms happen to produce a better power and performance package, you can bet that they will flip over to Intel.
People claimed that regarding the iPad for years and years and years. And yet Apple continues to not care about Intel's mobile chips. While in the universe of all possible things is clearly a possibility, it is highly improbable based on their actions.
Pad bound chips don't benefit from expensive high density processes.
Intel is overcharging there products, just look at the Intel annual benefice.
Could you please explain what this means, I am not able to parse it.
Haven't you heard about the Intel annual benefice? IAB as it's known to employees is considered the most valuable part of employee compensation. It adds an addition 0.5" every year. Older employees are swinging some impressive schlong. This also explains their difficulty hiring and retaining women.
Possible that this do not make a difference, but I still don't see why this is an advantage to buy an external design over cut down an internal one.
Allwinner is suspected of being a GPL violator, meaning if you want to build a product you'll be stuck with binary blobs.
We have hear of the hidden operating system inside cellphone that has capabilities to modify the SD that (e.g.) Android is running on. Perfect way for the NSA to tap into any phone when it wants to.
So, does this mean that laptops should now be as easy for them to hack you as your phone is?
There is a big market for low-end SoCs for phones and tablets. Look at all the recent phone models featuring Snapdragon 400 series chips. Atom x3 can sell to that market if the price is right; the integrated cellular modem will help keep down the total cost of the phone.
Exactly right. When you are as big as Apple, you can easily afford to hedge your bets and you would be stupid not to. Switching to Atom is a low probability event in my opinion, but if there is even a 1% chance that it would make sense, paying those developers is cheap insurance.
Low-end SoC market is already full of competitors. A new ugly chip (external design with Intel label, and no previous base) is unlikely to change anything. The bad Intel records into the SoC market don't help either.
They already pay developers to make x86 iOS. It's the very people who create the iOS SDK. The simulator runs iOS as native x86 code.
Huge different between shipping a product with Atom, and having an internal product that can be brought to manufacture quickly.