Intel's Medfield & Atom Z2460 Arrive for Smartphones: It's Finally Here
by Anand Lal Shimpi on January 10, 2012 8:00 PM ESTThe GPU
The PowerVR SGX 540 in Medfield is no different from what you'd get in an OMAP 4460, with the exception that it's clocked a bit higher at 400MHz.
The SGX 540 here is a remnant of Intel's earlier strategy to have Medfield out far sooner than it actually is going to show up on the market. Thankfully Intel has plans to introduce a PowerVR SGX 543MP2 based Medfield successor also before the end of the year.
Video Decode/Encode Support, Silicon Hive ISP
Intel relies on two more IP blocks from Imagination Technologies: the VDX385 and VDE285 for 1080p video decode and encode. Intel claims support for hardware accelerated 1080p30 decode, High Profile. Maximum supported bitrate is apparently up to 50Mbps, although Intel only demonstrated a 20Mbps High Profile stream:
Intel also claims support for 1080p30 video encode.
Medfield's ISP is provided by Intel owned Silicon Hive. The ISP supports cameras ranging from 5MP to 16MP (primary sensor), with the reference design standardizing on an 8MP sensor. Medfield supports burst capture at up to 15 fps (8MP).
The Process
Intel bifurcated its process technology a few years ago, offering both low power and high performance versions of each of its process nodes. Today those process nodes are staggered (45nm LP after high perf 32nm, 32nm LP debuts after high performance 22nm, etc...) however Intel plans on bringing both in lockstep.
Medfield debuts on Intel's 32nm LP process. The only details we have from Intel are that leakage is 10x lower than the lowest on 45nm. Compared to Moorestown, Medfield boasts 43% lower dynamic power or 37% higher frequency at the same power level.
The bigger and more valid comparison is to TSMC's 28nm process, which is what companies like Qualcomm will be using for their next-generation SoCs. It's unclear (and very difficult) to compare different architectures on different processes, but it's likely that Intel's 32nm LP process is more comparable to TSMC's 28nm LP process than it would be to any 4x-nm node.
It is important to note that Intel seems very willing to sacrifice transistor density in order to achieve lower power consumption where possible. I don't believe Intel will have the absolute smallest die sizes in the market, but I also don't believe it's clear what the sweet spot is for mobile SoCs at this point. It's quite likely that Apple's ~120mm^2 target is likely where everyone will eventually end up in the near term.
The Roadmap
Although Medfield is already posting competitive performance numbers, its current competition is roughly a year old. Within the next two quarters we'll see smartphones and tablets shipping based on Qualcomm's Krait. The next-generation Snapdragon platform should be Cortex A15-like in its performance level
Today we have Medfield, a single core Atom paired with a PowerVR SGX 540 built on Intel's 32nm LP process. Before the end of the year we'll see a dual-core Atom based Medfield with some form of a GPU upgrade. I wouldn't be too surprised to see something like a PowerVR SGX 543MP2 at that point either. In tandem Intel will eventually release an entry level SoC designed to go after the more value market. Finally we'll see an Intel Atom based SoC with integrated Intel baseband from its Infineon acquisition - my guess is that'll happen sometime in 2013.
Facebook
Twitter
RSS
164 Comments
View All Comments
mrtanner70 - Wednesday, January 11, 2012 - link
I am a long time fan of the site and still think your actual product testing is the best. However over the last year or so there has been a noticeable pattern of really being a bit too enthusiastic toward Intel PR and spin. I would like to see a little bit more intelligent skepticism when dealing with Intel reference designs and slide decks.Lucian Armasu - Thursday, January 12, 2012 - link
I have to agree, about Intel spinning stuff more and more lately, too. I've noticed it since they announced the "3D" chips. Everyone reacted as if the 3D are 10x better than SNB or whatever, when in fact that's not true. The 3D chips have only +37% performance at the same power level, OR -50% power consumption at the same performance level. That's barely as news worthy as they made it sound. They put 1 extra year ahead of AMD at most on top of the 6 months difference they already have now.So I don't know how everyone got from that that now that they have that they will be able to compete with ARM, which ARM at the same performance has much better power efficiency and also costs several times less.
Hector2 - Thursday, January 12, 2012 - link
What you said is about right for "active" power. A couple of things not mentioned is that the 22nm chip will be quite a bit smaller than 32nm (meaning even lower cost) and that the FinFET transistors drop the "standby" power by a factor of 10X-20X lower than even the 32nm version.As was said somewhere, Medfield gets Intel's foot in the door with a competitive product --- and I think most here will agree that it's at least competitive. Certainly, Intel's new partners must think so. The coming of the 22nm upgrade in 2013 is where it really gets interesting.
It's taken Intel awhile to carve out a viable single chip design that's product-worthy. Now that they have one, I expect them to march with the same tick-tock cadence that they've driven the PC market with -- and they already have working 3D FinFET 14nm chips in the lab.
Finally - Thursday, January 12, 2012 - link
Mark my words:The day that AnandTech welcomes a revolutionary AMD product (think e.g. Brazos II) with the words "It's finally here" will be the very day hell freezes over.
Hector2 - Thursday, January 19, 2012 - link
Hmmm. When do you think AMD might be entering the smartphone market ?thunng8 - Wednesday, January 11, 2012 - link
That intel graph for graphics performance is way off and misleading.If one of the smartphones tested is the iphone4S, then that phone should have >2x performance advantage in graphics compared to Medfield reference platform.
solipsism - Wednesday, January 11, 2012 - link
If Medfield is as good as it looks today and they can keep moving along as Cortex-A15 comes along tablets and smartphones could be Intel-based and Apple could be serious trouble by supporting ARM. It could be the whole the PPC issue all over again where Apple will linger for way too long before switching architectures. That's a worse scenario perfect storm, of course, and a long ways off, but it's still something that can't be good for companies with vertical integration with heavy ARM support.markit - Wednesday, January 11, 2012 - link
as infact 2 year old cortex a-9 (even if single cored) outperform it / mhze.g. if i scale down medfields results linearly
(which is imho quite accurate for this benches)
i get:
@1ghz
2130ms and 72766 points
and compare against my old tegra 2 based toshiba ac100 (running ubuntu and chrome 13)
with one core running @ 1 Ghz, and second core turned off i get:
sunspider 1962ms ( per mhz 9% faster)
browsermark 111872 ( per mhz 54% faster as medfield)
so regarding browsermark infact the ac100 nearly matched the 1.6ghz results of this medfield reference plattform while running on only one 1 ghz core,..
btw. both cores turned on and @ 1Ghz:
sunspider 1941 ms
browsermark 114523
so hmm, i think i can guess why intel didn`t choose dual-core friendly benchmarks,..
and remeber the ac100 did hit the stores June 2010!
FunBunny2 - Wednesday, January 11, 2012 - link
I still am puzzled by this. It's been at least a decade since an Intel chip executed X86 code in hardware; the chip is a RISC design fronted by an emulator. Or has Intel gone back to X86 instructions in hardware for these SoC?dealcorn - Thursday, January 12, 2012 - link
Was I deceived? I read about fin-fet and 3d transistors and was impressed because it is credible technology that will permit Intel to do good things before others. That colored my perspective on mobile Atom leading to statements such as "Show time starts at 22 nm." I felt that absent the 3d benefits, Intel would find it tough to go head to head against ARM. While Intel focused attention on 3d, the implication was they are dead in the water until 22nm arrives.That is wrong. Mobile Atom is never going to be built using 22nm transistors. It will be built using 22nm LP transistors. Furthermore, it appears that 32nm LP transistor get the job done just fine (even if 32 nm is worthless in this context). Is 32nm LP the secret sauce that lets it all work? Should the headline have been "Intel's process advantage muscle works almost as good at 32 nm LP as it was supposed to work at 22nm."?