Intel's Medfield SoC: Blazing!

[Belegost]

What I want to see is the dual core tablet version running x86 Windows 8, all the sleek metro goodness and with an x86 chip it should be compatible with legacy applications from Win7. That is a beautiful future to think about.

 

[smartpatrol]

Why is that? One of the big selling points of Java is that it's architecture-independent. There's lots of Java already running on standard Windows-based desktops. Why would there be architecture concerns about app compatibility? And why would Intel's translation layer have a lot of overhead when there's already a pretty good Java bytecode interpreter for x86?

When you look at something like this:
http://en.wikipedia.org/wiki/Java_virtual_machine

and then look at the diagram on the right, you can see that the flows are pretty similar.

And ARM is not like x86 where everything is backwards compatible and new instructions are supersets of prior ones. Various cores have various instructions and some are subsets of previous versions.
http://en.wikipedia.org/wiki/ARM_architecture
You can't run an ARMv6 binary on an ARMv7 core. So new instruction set architectures running Android are not unusual.

Yes but not all Android apps are Java. Some of them are native code. Most graphically-intensive 3D games/game engines are not written in Java.

 

[jpiniero]

Yes but not all Android apps are Java. Some of them are native code. Most graphically-intensive 3D games/game engines are not written in Java.

This isn't the iPhone, there aren't that many games on Android. It shouldn't really be a problem.

 

[MrX8503]

What I want to see is the dual core tablet version running x86 Windows 8, all the sleek metro goodness and with an x86 chip it should be compatible with legacy applications from Win7. That is a beautiful future to think about.

I would bet that's what MS is shooting for. With this new SOC from Intel, that might just what happens.

 

[pm]

Yes but not all Android apps are Java. Some of them are native code. Most graphically-intensive 3D games/game engines are not written in Java.

Yeah, good point. I didn't think of that. But I'd think most of the apps are in Java... But yeah, you are right, games likely aren't.

I've written a few Android "apps" - with the quotes around them because they are mostly stupid things that were intended to teach me the basics. I made one of a photo of a hamster that made various sounds as you swiped your finger across it and would say various things. Nothing that I'd even think about putting on the Android market and nothing that I would say gives me the ability to claim that it's a real app. But I got the gist of things. I do think that anything that uses the Android SDK and Eclipse for development probably would work on any architecture with no changes at all as long as the Android versions matched.

 

[Fire&Blood]

Intel needs partners now. The biggest players in the smartphone/tablet market are linked to SoC's, Intel needs to seek out their own. Apple delegates it's own to others, HTC goes to Qualcomm first and Samsung makes it's own.

Motorola, LG, Acer and Asus, they all don't have a go to chip maker in the mobile market. Doesn't sound like much but in theory that's enough support to help Intel gain market share. Maybe Intel could hook up with Amazon for phones and tablets.

 

[MrX8503]

I think Intel already has support from Motorola. If Intel can deliver, the support will come.

 

[CTho9305]

I work for Intel and there's a pretty good chance that I'd go pretty far out of my way to get my hands on an Intel-powered phone. If it's not ludicrously expensive, and it works on AT&T and the reviews say it's a decent phone, I'll get one.

I have a coworker that feels the same way.

I'm pretty impressed with the Austin team. I don't think people realize how many teams are involved to win the smartphone/tablet/ultrabook market segments. It's kind of crazy when you count it up.

Did the Austin team do the SOC or just the core? From what I've heard, the biggest challenge when it comes to building a chip for phones is all the stuff around the core. And this SOC is clearly very different from anything we've seen from Intel before!

It will be interesting to see how it compares to 28nm competitors which will probably be available in similar timeframes.

 

[kaerflog]

Heres the reference Intel phone.
http://www.bgr.com/2012/01/11/hands-on-with-intels-medfield-android-phone/
I like this one much better than the Lenovo.
If this one was a little more rounded on the corners, it would look alot like the IP4.

 

[lothar]

While the results are impressive for Intel's first shot out of the gate, by the time it hits the market, it will (probably) be already behind. If intel's recent track record with desktop/laptop CPU/chipset in any indication, they will have a hard time to keep pace int eh mobile phone market.

 

[lothar]

Intel needs partners now. The biggest players in the smartphone/tablet market are linked to SoC's, Intel needs to seek out their own. Apple delegates it's own to others, HTC goes to Qualcomm first and Samsung makes it's own.

Motorola, LG, Acer and Asus, they all don't have a go to chip maker in the mobile market. Doesn't sound like much but in theory that's enough support to help Intel gain market share. Maybe Intel could hook up with Amazon for phones and tablets.

Asus is an exclusive Tegra partner.
They get 1st dibs on anything Tegra(See: Transformer Prime), just like HTC does with Qualcomm processors.

 

[Fire&Blood]

Asus is an exclusive Tegra partner.
They get 1st dibs on anything Tegra(See: Transformer Prime), just like HTC does with Qualcomm processors.

Well the ASUS Padfone will feature Qualcomm's S4 over Tegra3 but it's not as relevant, Intel should be able to rally enough support to penetrate the market with or without ASUS. Still, as promising as Intel chip appears today, it may turn out to be meh at launch but should pave the way for future editions.

 

[Mopetar]

Engadget has an article where they say that the Padfone will now be using the Tegra 3 so they may have changed things since they first announced it.

 

[Fire&Blood]

Verge posted a Padfone hands on article 3 days ago, stating they found the S4 in the settings: http://www.theverge.com/2012/1/9/2694298/asus-padfone-hands-on

Not that it really matters for the Medfield anyway. ASUS alone isn't critical for the chip and Intel should muster enough support with or without them. Motorola and LG alone could be enough for Intel to get the foot in door.

 

[Mopetar]

Not that it really matters for the Medfield anyway. ASUS alone isn't critical for the chip and Intel should muster enough support with or without them. Motorola and LG alone could be enough for Intel to get the foot in door.

Honestly, if those are the partners that Intel has to depend on, they're in trouble. LG has never really released a device that made people want it over anything else at the time, and Motorola really hasn't had one since the original Droid or possibly the Droid X.

If Intel wants a foothold they need to get their chip in a model that sells several millions of units. Otherwise they're just going to be in a bunch of phones that don't sell well and eventually get left by the wayside when it comes to support and updates.

Intel really needs something like an EVO or a Galaxy S where they can get an appreciable number of sales and have a phone that people still want months down the road.

 

[cbn]

Heres the reference Intel phone.
http://www.bgr.com/2012/01/11/hands-on-with-intels-medfield-android-phone/
I like this one much better than the Lenovo.
If this one was a little more rounded on the corners, it would look alot like the IP4.

It would be so interesting to see a teardown of that device!

 

[lothar]

Honestly, if those are the partners that Intel has to depend on, they're in trouble. LG has never really released a device that made people want it over anything else at the time, and Motorola really hasn't had one since the original Droid or possibly the Droid X.

If Intel wants a foothold they need to get their chip in a model that sells several millions of units. Otherwise they're just going to be in a bunch of phones that don't sell well and eventually get left by the wayside when it comes to support and updates.

Intel really needs something like an EVO or a Galaxy S where they can get an appreciable number of sales and have a phone that people still want months down the road.

My sentiments exactly.

 

[MrX8503]

Honestly, if those are the partners that Intel has to depend on, they're in trouble. LG has never really released a device that made people want it over anything else at the time, and Motorola really hasn't had one since the original Droid or possibly the Droid X.

If Intel wants a foothold they need to get their chip in a model that sells several millions of units. Otherwise they're just going to be in a bunch of phones that don't sell well and eventually get left by the wayside when it comes to support and updates.

Intel really needs something like an EVO or a Galaxy S where they can get an appreciable number of sales and have a phone that people still want months down the road.

You forget that this is Intel, Chipzilla. If anyone is going to take down ARM, its going to be Intel. Medfield is their first foray into the smartphone biz so I'm not sure why you expect them to win over vendors left and right. If the benchmarks are true, Medfield already beats ARM and is in a good position for Intel to roll out their future SOCs. In two years, ARM could become the next AMD, holding onto dear life.

If Intel can deliver, the support will come.

 

[theeedude]

ARM will have a higher market share on Windows than x86 will have on Android.

 

[IntelUser2000]

Intel really needs something like an EVO or a Galaxy S where they can get an appreciable number of sales and have a phone that people still want months down the road.

You know its amusing that Intel announced partnerships with only Motorola and Lenovo when all the rumors said LG, and even Samsung. Hell, the former even had an executive saying that! But they are taking it differently this time. Rather than announcing a chip without partners, they are waiting partners to annouce them instead. Perhaps we'll see LG and/or Samsung at MWC.

 

[alent1234]

You forget that this is Intel, Chipzilla. If anyone is going to take down ARM, its going to be Intel. Medfield is their first foray into the smartphone biz so I'm not sure why you expect them to win over vendors left and right. If the benchmarks are true, Medfield already beats ARM and is in a good position for Intel to roll out their future SOCs. In two years, ARM could become the next AMD, holding onto dear life.

If Intel can deliver, the support will come.

2012 will have new ARM CPU's that will wipe the floor with medfield

intel should be making medfield on 22nm so to be the fastest and most power efficient. but they are making a classic mistake others have made in the past. when you control most of a market and have good margins and a lower end competitor comes along they are making a half assed attempt to compete in an effort to protect the margins of their main products.

i bet apple makes a smaller profit on each iphone than they do on a $2000 MBP but they released the products anyway because as Steve Jobs said, unless you do it someone else will. and he didn't make it up

 

[smartpatrol]

You forget that this is Intel, Chipzilla. If anyone is going to take down ARM, its going to be Intel. Medfield is their first foray into the smartphone biz so I'm not sure why you expect them to win over vendors left and right. If the benchmarks are true, Medfield already beats ARM and is in a good position for Intel to roll out their future SOCs. In two years, ARM could become the next AMD, holding onto dear life.

If Intel can deliver, the support will come.

Qualcomm S4 will easily beat Medfield in benchmarks. Anand estimated that it will get ~30% better performance/clock compared to Cortex A9, and it will run at 1.5-1.7GHz. We should be seeing S4 mid-year.

Plus ARM A15-based SOCs will launch later this year/early next, which offer double the performance/clock of A9. TI demoed their OMAP5 reference smartphone at CES.

I'm excited for Medfield, but I don't expect Intel to hold the performance crown for long, if at all, this year. I'm more interested in seeing what their next-gen SOC will be like.

 

[MrX8503]

2012 will have new ARM CPU's that will wipe the floor with medfield

intel should be making medfield on 22nm so to be the fastest and most power efficient.

Medfield is already more powerful and more efficient than an ARM dual core...so what are you getting at? New ARM CPUs?, there will be new Medfield SOCs too. Medfield is only a single core, I'm sure it'll be able to hold its own as a dual core if Intel can deliver. Right now, Intel is on the right track.

Qualcomm S4 will easily beat Medfield in benchmarks. Anand estimated that it will get ~30% better performance/clock compared to Cortex A9, and it will run at 1.5-1.7GHz. We should be seeing S4 mid-year.

Plus ARM A15-based SOCs will launch later this year/early next, which offer double the performance/clock of A9. TI demoed their OMAP5 reference smartphone at CES.

I'm excited for Medfield, but I don't expect Intel to hold the performance crown for long, if at all, this year. I'm more interested in seeing what their next-gen SOC will be like.

Next gen Medfield would take care of ARM A15, the point is, Intel has something faster than ARM, which opens doors for Intel in the future to take over.

 

[alent1234]

apple is not going to switch to medfield
samsung is not going to switch

intel is fighting for a small part of the market here. only way intel can win is if they partner with MS again and enable manufacturers to sell cookie cutter phones at CHEAP prices with good performance. like dell and gateway used to do in the 1990's. make a reference platform with a ready OS where a manufacturer can go to market a lot quicker than the 2 year current design cycle

if i can buy a nice android phone for say $200 or so with no contract i will dump my iphone. Intel needs to do the same to the phone market like they did to the computer market. there really is no reason to spend months writing custom code just to sell a cell phone. you should slap some hardware together, license an OS and send it to best buy

 

[Red Storm]

The only thing Intel needs is for their first chip to be competitive. It doesn't have to smoke the competition. Having as much money as they do as well as your own fabs is an immense advantage that very few other companies can match. As long as the first chip isn't a failure, they'll be alright.