POST A COMMENT

78 Comments

Back to Article

  • jhoff80 - Thursday, September 27, 2012 - link

    I hadn't realized that Windows 8 x86 is the only version to support Connected Standby. It doesn't quite make sense to me why they would put the feature in one and not the other. Is it just because they figure they have until Haswell comes out before they really need it in x64 anyway? Reply
  • dagamer34 - Thursday, September 27, 2012 - link

    Considering most of these devices only have 2GB of RAM and CPU performance isn't the blazing performer we expect from a Core iX chip, 64-bit support isn't much of a priority.

    Besides, when you look at Intel's roadmap and see the stuff they are offering for 2013, you can that those chips are far more clearly designed with mobile in mind. Atom gets it's first CPU rearch in 5 years since it comes out, plus a die shrink, plus Intel HD graphics. Haswell becomes available in SoC form, available at 10W (and possibly even lower), and has far more integration with other chips on the board.

    While Windows 8 is coming out this year with some really good hardware, the amazing stuff is all for next year.
    Reply
  • Shadowmaster625 - Friday, September 28, 2012 - link

    Really good hardware?? 1.8GHz atom is still 1.8GHz atom. It is just too slow. A penryn CULV from 5 years ago is faster and it is only 10 watts. The fact that intel cannot shovel out any better turds than this in 5 years is just insulting. Reply
  • MrSpadge - Friday, September 28, 2012 - link

    Sure they could, they just choose not to. Not that this makes the situation any better for us.. Reply
  • BSMonitor - Friday, September 28, 2012 - link

    And to think it outperforms all the ARMs you have been using in any of your smart phones for the past 4 years...

    Guess 5 y/o Intel turds are better than start of the art ARM turds...

    In your comparison.
    Reply
  • relic1138 - Tuesday, October 16, 2012 - link

    Well yea but this CPU is not for RT but the regular version. Microsoft is using ARM for RT. The Atom really is a horrible CPU, very, very slow. Literally for 10 dollars more per chip you can have a i3. I bought a Chromebook last year that I had an Atom, I wanted to throw it out of the window every time I used it. My new Chromebook is 100 times way better now that Samsung dropped the Atom. Hate that chip, hate it, hate it and will never own another one. I even tried Android on my Chromebook for kicks and it was molasses slow compared to my Asus Slider SL101 Tablet. Reply
  • Zink - Friday, September 28, 2012 - link

    If it's as fast or faster than the competition I don't see what the problem is. Windows 8 RT will have A15 in the same form factor which is a much newer architecture so get one of those if it will hurt your feelings less. Reply
  • maroon1 - Wednesday, November 07, 2012 - link

    penryn CULV doesn't have built-in GPU, and memory controller like modern CPU's Reply
  • RU482 - Friday, September 28, 2012 - link

    Whoa....have any references to that 10W and below haswell claim? All the intel docs I've read state 15W TDP, even for the GT3 SoC version Reply
  • vdx660 - Friday, September 28, 2012 - link

    Haswell will have a 10W -TDP
    http://www.anandtech.com/show/6248/haswell-at-idf-...
    Reply
  • vdx660 - Friday, September 28, 2012 - link

    Please read this

    http://msdn.microsoft.com/en-us/library/windows/ha...

    and

    http://msdn.microsoft.com/en-us/library/windows/ha...

    regarding connected standby. It seems that all Windows 8 versions including x64, x86 and RT support Connected Standby
    Reply
  • jhoff80 - Friday, September 28, 2012 - link

    Well, that's what I had thought, but the article says:

    "Connected standby is only currently supported by 32-bit Windows 8, so although Clovertrail supports x86-64 the platform will launch as 32-bit only. There's no support for alternate OSes at this point."
    Reply
  • ssiu - Thursday, September 27, 2012 - link

    How does the Z2760's GPU compare to the ones in iPad3/iPhone5? e.g. how many "cores"? Reply
  • ltcommanderdata - Thursday, September 27, 2012 - link

    The SGX545 is only available in a single core configuration and is based on the older Series5 design. Apple uses the SGX543MP which uses the newer Series5 XT design, in a 3 core configuration for the iPhone 5 and a 4 core configuration for the iPad 3. Reply
  • frostyfiredude - Thursday, September 27, 2012 - link

    533Mhz MP2, last I checked. The A5X is MP4 at 200Mhz and the A6 is MP3 at 266Mhz, so Clover is a solid 33% faster. As far as mobile is concerned, quite impressive. Reply
  • frostyfiredude - Thursday, September 27, 2012 - link

    Or not, Looks like it changed... Reply
  • lowlymarine - Thursday, September 27, 2012 - link

    It's a single "core" since it's a Series 5, not Series 5 XT design, as someone pointed out below. However, the SGX545 is still capable of higher performance than one SGX543 "core" at the same clock speed per the Wiki article (http://en.wikipedia.org/wiki/PowerVR#Series_5), and is running at twice the clock speed of the SGX543 MP3 in the A6. So it should broadly be a wash, especially at a low resolution like 1366x768.

    The real question is, as always, if Intel will be able to release working drivers.
    Reply
  • ltcommanderdata - Thursday, September 27, 2012 - link

    You mention the SGX545 is only supporting DX9.0L3 at launch. However, the hardware is capable of DX10.1/OpenGL 3.x. Any word on plans to support this? Reply
  • toyotabedzrock - Thursday, September 27, 2012 - link

    This website mentions that as well.
    http://www.notebookcheck.net/Intel-Graphics-Media-...
    Maybe they just claimed it is dx10 compatable. I also noticed that the sgx544 is only dx9.
    Reply
  • powerarmour - Thursday, September 27, 2012 - link

    Unlikely to change, Intel has also hobbled the previous PowerVR GPU'd Atom's in the past this way, and I can't see any reason to think different. Reply
  • toyotabedzrock - Thursday, September 27, 2012 - link

    Intel really is fumbling with Atom, they need a real core refresh and they should have pushed to get the low power 22nm node ready.

    The gpu is way behind Apple's and I bet the display will not get more than 10 seconds of attention. And without DX11.1 there is no acceleration for the interface or the browser. I'm sure they will not use a high resolution display. And i bet one of the vendors will try and sell a tn based tablet

    I wonder if all the MS devs wear glasses or something.
    Reply
  • ssj4Gogeta - Thursday, September 27, 2012 - link

    I'm not sure about the GPU being much slower than Iphone's. It's running at much higher clocks. Reply
  • Wilco1 - Thursday, September 27, 2012 - link

    But the Z2760 only has a single GPU while the iPhone5 has 3 GPU cores at 266MHz, and the iPad has 4 at 200MHz. Assuming the SGX543 and 545 are identical in performance, Apple's GPUs are 50% faster overall. Note that using several low clocked GPUs results in lower power consumption than a single highly clocked GPU (but requires more area of course). Reply
  • Ryan Smith - Thursday, September 27, 2012 - link

    Interestingly 545 and 543 are not identical in performance. Despite the model numbers there are some pretty significant differences between them. Reply
  • B3an - Friday, September 28, 2012 - link

    From reading the wiki page a single core 545 is faster than a single core 543 would be at the same clocks. So with the 545 clocked at 533MHz it should be quite competitive to the 3 core 543 @ 266MHz in the A6 or the 200Mhz 4 core 543 in the latest iPad... Reply
  • dragonsqrrl - Thursday, September 27, 2012 - link

    They aren't identical in performance. The SGX545 offers more performance per core then the SGX543. The config in the iPhone 5 will probably still be faster, but I'm not expecting a huge difference in performance. Definitely not 50%. Reply
  • lilmoe - Friday, September 28, 2012 - link

    No. The difference isn't 50%.

    PowerVR 545 (single core at 200mhz) is capable of 7.2 GFLOPS in apposed to 6.4 for the SGX 543. On paper, the 545 is 12.5% faster than the SGX 543 (single core at 200mhz). If you calculate the number of cores vs clock speeds you'll get this out of the GPUs:

    - Apple A5X: 25.6 GFLOPS
    - Apple A6: 25.54 GFLOPS
    - Clover Trail: 19.2 GFLOPS

    Assuming that the Z2760 comes in single core configuration, it should be 25% slower on paper. What we aren't taking into account is the efficiency of the bus and memory controller and the platform as a whole, it's well known that the speed of the GPU core itself isn't the only bottleneck. As stated by the article, the GPU is "fed" much more efficiently in Clover Trail. The number might be smaller or larger in the end product. We'll have to wait and see.
    Reply
  • r3loaded - Thursday, September 27, 2012 - link

    A real core refresh is coming next year with Silvermont, with a big leap in the core architecture and a GPU based on Intel's own HD graphics IP instead of being licensed from PowerVR. Clover Trail is nothing more than stopgap to fend off ARM's advance. Reply
  • r3loaded - Thursday, September 27, 2012 - link

    Also, if you do find a vendor who makes a tablet with a TN screen, let me know so I can throttle them. >:( Reply
  • DanNeely - Thursday, September 27, 2012 - link

    Look at the $100ish android 2.x crap TigerDirect runs a sale on every few weeks; nothing but TN crap there. Reply
  • B3an - Friday, September 28, 2012 - link

    "And without DX11.1 there is no acceleration for the interface or the browser"

    Totally not true. On Win 8 the UI is completely GPU accelerated with DX10, 10.1, 11, and 11.1.

    You dont seem to understand how Win 8 or DirectX works. Even if a GPU does not support DX11.1 the 11.1 pipeline is still used, the hardware will just not support all the features of it, which is totally not needed anyway for UI acceleration. The UI will be COMPLETELY accelerated regardless if the GPU is DX10 or DX11.1.
    Reply
  • relic1138 - Tuesday, October 16, 2012 - link

    Have you seen the prices for these new tablets running this horrible CPU, the Samsung starts at 700, ridiculous. Battery life, 4 - 6 hours, nah, hook me up with an ARM version. Reply
  • ssj4Gogeta - Thursday, September 27, 2012 - link

    Anand, can you confirm that the GPU is indeed PowerVR SGX 545? Wikipedia lists it as PowerVR SGX 544MP2. Reply
  • Kidster3001 - Thursday, September 27, 2012 - link

    That would probably be the Cloverview flavor intended to support Android. You don't think they would completely ignore phones do you? Reply
  • DanNeely - Thursday, September 27, 2012 - link

    Cloverview is the CPU in the Clovertrail platform... Reply
  • lilmoe - Friday, September 28, 2012 - link

    You forgot to mention that Saltware is the core inside the CPU... Intel is getting on my nerves with their codemanes, it's really confusing every time. Reply
  • ssj4Gogeta - Friday, September 28, 2012 - link

    *Saltwell
    ... which is based on the Bonnell microarchitecture. :P
    Reply
  • lilmoe - Saturday, September 29, 2012 - link

    ughhhhh Reply
  • relic1138 - Tuesday, October 16, 2012 - link

    Double ughhhhh! Reply
  • Kidster3001 - Tuesday, October 30, 2012 - link

    Meaning it is completely not possible to have another variation of the Clovertrail name? Reply
  • MonkeyPaw - Thursday, September 27, 2012 - link

    Where I can really see a big deal with x86 tablet hardware is the potential to build cheaper inventory setups. So long as it has the power to scale GIS programs like ArcView from desktop to tablet (and smartphone), then this could be a rather valuable processor to a particular industry looking for more powerful, compatible, and cheaper field hardware. Reply
  • PubFiction - Thursday, September 27, 2012 - link

    Netbooks started at $200 tons in the $300-400 range and when you subtract a keyboard it should get cheaper, atom should be starting closer to ARM chips while perhaps not on as nice of hardware they need to hit $200 IMO. Reply
  • ssj4Gogeta - Friday, September 28, 2012 - link

    You forgot to add a touch digitizer and an IPS screen which is necessary for the wider viewing angle the tablet use case requires. Reply
  • seapeople - Sunday, September 30, 2012 - link

    If it wasn't for Apple, you'd probably be getting your wish with tablets competing in the $150 range. Of course, they'd have 600x400 TN panels with two hours of battery life, but they'd be a lot cheaper. Reply
  • powerarmour - Thursday, September 27, 2012 - link

    Well seeing how the current PowerVR 545 drivers (GMA 3600/3650) fare under Windows 8 (i.e. they don't without Metro applications crashing, and the performance is diabolical) I'm intrigued by how it'll actually perform.

    Is this suddenly going to be a driver miracle, where a 5 series PowerVR GPU actually works in a DirectX environment?!

    I'm not holding my breath too long...
    Reply
  • ajp_anton - Thursday, September 27, 2012 - link

    Dual-core Cortex A9 is faster than Snapdragon S4?
    Or clock for clock, S4 is only 5% faster than A9?
    Reply
  • tuxRoller - Friday, September 28, 2012 - link

    Heh, I should've read all the posts first, I guess. Reply
  • lilmoe - Friday, September 28, 2012 - link

    That's because the A9s are clocked at 1.8 GHz, and Krait is clocked at 1.5.... So yea, at the same clock speed, Krait is faster. Reply
  • Krysto - Friday, September 28, 2012 - link

    What's funny is that Clover Trail is barely competitive with S4 that was launched in the beginning of the year. And Clover Trail is not even out yet.

    And it STILL isn't competitive from a power consumption point of view. If it needs a 30 Whr battery, which is twice as large as other tablets, then it's definitely not competitive in power consumption:

    http://www.theverge.com/2012/9/27/3418260/intel-sa...
    Reply
  • tuxRoller - Friday, September 28, 2012 - link

    So, normalized to dual core krait (which is the only arm15 class chip in the mix) a dual core a9 clocked 20% higher does exactly 14% better. Assuming linear scaling, upclocking Krait to 1.8 GHz gives it only a 5% advantage over the two year old a9.
    Is krait particular weak in this benchmark or is something else going on?
    What happened to the 3.1 dmips/Hz vs. 2.5dmips/hz advantage?As far as I have been able to determine those figures were never verfied.
    Reply
  • B3an - Friday, September 28, 2012 - link

    I'm sure it's just that particular benchmark. Maybe its why Intel decided to use that one, so it made them look much better (i'd still expect Clover Trail to be faster in other tests anyway, just not by as much). We all know that Krait is certainly more than 14% faster than A9 in most benchmarks, clock for clock. Reply
  • Wilco1 - Friday, September 28, 2012 - link

    Why would you expect Clover Trail to be faster? Like Medfield it has a turbo but likely runs at 1.3GHz by default (possibly even lower if both cores are active). Therefore I wouldn't expect it to be much faster than Medfield, except for the slightly increased turbo speed or on multi-threaded tests.

    Note also that the real competition will be with quad-core Cortex-A15 (such as 1.8GHz Tegra4 due early next year) - not the older 40nm SoCs selected for benchmarking by Intel. Once the A15 is out, I wouldn't expect Clover Trail to win any benchmarks, not even SunSpider.
    Reply
  • B3an - Friday, September 28, 2012 - link

    From what i've read Clover Trail's cores and platform seem to be a slight improvement over Medfield. And with a Turbo of 1.8GHz. i can imagine it being faster than competing ARM based Win RT tablets on release.

    I agree that when A15 is out it will be faster, but on the release of Win 8 i think Clover Trail has a good chance of being the fastest SoC in this power consumption category (for a short time).

    And a bit off topic but unless ARM can make *vastly* better SoC's i dont see any point in Win RT any more. A year ago i would argue that Win RT made perfect sense as Intel had nothing that came close to ARM for power consumption and battery life. But now that Intel can match ARM (or even just get close) and also getting in to form factors just as thin and light as ARM can while at the same price points, i dont see any point in Win RT at all now.

    Clover Trail Win 8 tablets will now offer literally everything that ARM Win RT tablets offer, but with the big advantage of running all desktop software. While A15 will be faster, it wont be massive jump, in REAL world usage i doubt anyone would notice a difference. It's not worth having another OS for ARM when it has no big advantage and will just confuse consumers and be a marketing nightmare.
    Reply
  • Wilco1 - Friday, September 28, 2012 - link

    I agree Clover Trail is an improvement, but even at release it will be up against newer 1.7GHz quad-core Krait, 1.7GHz Tegra 3's and (if ready at launch) 2.0GHz OMAP5. With just dual cores and a far lower frequency (the turbo is only for short bursts), I do not see how Clover Trail could beat any of these.

    I can't see why you'd think Win RT is useless. Win RT devices will have much faster CPU/GPUs and better battery life, smaller form factors and lower prices. When Medfield came out we saw similar marketing from Intel how it was faster and lower power, but actual Medfield phones turned out to be mediocre in performance and battery life compared with phones based 40nm ARM SoCs - despite Medfield being 32nm (note Sunspider was the only benchmark that made Medfield look good, but that is mostly due to software optimization, not raw CPU performance).

    Yes Win RT can't run existing x86 applications directly. But is that really a big issue? Which x86 applications are you actually planning to run on your Win8 tablet? Will they work well on a low res touchscreen interface? I doubt it. Will they work well with power saving states? I doubt it. Important applications will be ported and adapted to tablets, just like Office and other basic Windows apps. At that point they should be available for both Win RT and Win8.

    However in the end I guess a lot of people will be persuaded by pricing, not the technical details.
    Reply
  • Penti - Friday, September 28, 2012 - link

    Those new chips aren't out, and aren't powering Windows RT (WinRT is short for Windows Runtime which exists on both for confusion) next gen GPU's is still 6-12 months away for most users/devices.

    How about Office, Outlook? Business/Accounting software? Office 2013 RT won't be out the same time as these ARM-tablets. How about the fact that Windows RT is less capable then Android and alternatives? Price isn't everything here.

    Look the Surface RT will have a keyboard, it will have a multi-touchpad, it will have a touchscreen without pen-support (no dual-digitizer). x86 variants are just so much more attractive if you don't go with Android/iOS instead. It's not like it's easier to port apps to Windows Runtime then to Android NDK, iOS, and the competitors. SDK isn't even out yet in full for the platforms. When A15-based SoC's will be prevalent Intel will move on to Valleyview/Silvermont with none-Imgtec graphics thank god. Different devices and categories to begin with. Windows RT and Windows Phone 8 is quite limited in their platforms and supporting C#, C++, etc isn't winning over developers when every other platform is a lot more robust and still support all the technologies they would expect. Sure you will be able to do some DX stuff but Tegra 3 performs pretty badly. OpenGL ES reaches more users and developers have no reason to not target those if the engine of their choosing has support. Next to those devices (Windows RT) you would be able to pick up IvyBridge/Haswell tablets and they are easy enough to target for game dev.
    Reply
  • Wilco1 - Sunday, September 30, 2012 - link

    The 1.7GHz Tegra3 is already out, for example HTC One X+ and Transformer Infinity use it. Faster Kraits are on the way, and Cortex-A15 will be released in the next few weeks. So it's not like Clover Trail will be competing with 6-12 months old SoCs, it will have to compete with the latest ARM cores from the very start.

    Office 2013 will be included in Windows RT from the start (it's called a preview as it's not the final version but 95% of functionality is there). IIRC Outlook won't be supported as RT already has an integrated mail program. Note that big business apps often support access via a browser, so in many cases you wouldn't need to install any software at all.

    You're right that x86 Win8 may be more capable, but you will be paying more as well (a lot more if the rumours are correct).
    Reply
  • Penti - Monday, October 01, 2012 - link

    Tegra3 doesn't have an A15 or A15-class cpu neither does it have a next generation GPU so what are you on about? Tegra 3 is old. A clock increase doesn't change that. Kraits aren't using Adreno 320 yet. S4 Pro aren't available yet. LG Optimus G will become available with it pretty soon though. RT's Integrated mail application is worse than that built in into Android, iOS, Blackberry, Nokia etc. It's a lot worse then Android third party mail clients like TouchDown or iOS integrated. No Office 2013 RT aren't finished by October 26 it's coming next year for many. "Customers can expect to get these updates starting in early November through January depending on their language." - Office blog. Your right that a preview will be available when shipping though.

    Many large enterprise applications are accessible mainly through the web, but that doesn't cover all the large enterprises that for example uses Microsoft's technologies and business systems which runs as stand-alone app or all the small business apps you preferably run stand-alone. So it leaves out pretty much any business that doesn't afford a multi-million dollar ERP/CRM/Payroll system. It doesn't take into account that you can also run Citrix receiver to run those Windows stand-alone apps on an Android tablet just the same. Streamed apps probably works better on none-windows platforms here.

    On the other hand the preview of Office 2013 is available on Windows 8, 7, today, but not RT. Yet. Get some sense. Your off the target.

    Yes x86-tablets costs just the same as today. It's tablet pc's flying spaghetti damn it, about 800-1000 dollars. But that is for 600-800 gram devices, not second devices for none-existing entertainment. Later the slates will be able to go down to 600-700 dollar though i.e. same as higher-end Windows RT devices. Office 2013 RT home & student editions aren't something you your meant to run in business either for that matter. There is no reason the hardware in an CloverTrail Z2760 device should cost more then in an ARM equivalent either. For business it's all about the software, management and capabilities. While Office 2013 RT seems very powerful it's not a full replacement so alternatives still bode well. Of course 8 Pro devices has already licensed Office 2010 and productive environments to go from the start. In hybrid devices it shouldn't matter that 2013 preview has touch control-enhances. It's not like it's easier to port third party software to C++/CX/WRL then to any other platform. There is no definite advantage and it has to compete with much more powerful x86-devices that is only 150-200 dollar more or something of that character.
    Reply
  • Wilco1 - Tuesday, October 02, 2012 - link

    The original point I responded to was: "And with a Turbo of 1.8GHz. i can imagine it [Clover Trail] being faster than competing ARM based Win RT tablets on release."

    My point is that it won't be faster as current ARM SoCs have much higher frequencies than the ones mentioned in the article.

    For example the latest Tegra3 has a 30% faster CPU and 25% faster GPU than the one benchmarked by Intel in the article. That means that even the old Tegra3 will outperform the Z2760 at launch.
    Reply
  • B3an - Saturday, September 29, 2012 - link

    I guess we'll have to wait until Clover Trail tablets are available but the main advantage ARM/Win RT had was smaller and lighter form factors, and longer battery life. Being as Clover Trail will be in tablets that are just 8.5mm thick (shown off recently) it already matches ARM variants in that category. Then you have weight, which again it pretty much matches. Thats two main ARM/Win RT advantages already gone... Then you have battery life, and thats looking good but i'll wait for benchmarks. But lets say ARM tablets last an hour or so longer, even then Clover Trail will get "close enough". Say it gets 10 hours battery life and ARM gets 11.... just not worth going for ARM/Win RT with such a small difference.

    If theres a quad-core Krait out on release though, and has the Adreno 320 GPU, then that could actually make a noticeable difference in real world usage which is the only that that matters (99.9% of tablet users wont care about benchmarks). But that would be the on ONLY advantage Win RT tablets would have, and for how long? Before Clover Trail the choice was really simple: Intel + Win 8 for performance but bigger heavier designs with lower battery life and ARM + Win RT for slower but lighter and smaller designs with much longer battery life. Now that gap has become very blurred an it makes RT look quite pointless.
    Reply
  • relic1138 - Tuesday, October 16, 2012 - link

    Right, why wouldn't a person just buy a full powered i5 - i7 laptop if he or she needs it. I never understood why people try to force their devices to work past the scope of it's design/purpose. Even running the none RT version on a i3 -i7 tablet sounds silly to me, Photoshop, have you seen those wonderful 24" Wacom monitors, silly.

    I do however like the detachable keyboard idea, I own a Asus Slider SL101 and I love using it to type out memo's, email's, doc's,ect.
    Reply
  • somata - Friday, September 28, 2012 - link

    SPECint typically does a decent job of gauging performance on general-purpose code, and the numbers presented seem believable to me. I too was surprised that Krait didn't perform better, but Geekbench also shows Krait performing worse than expected (vs the A9), even if you just focus on the single-threaded tests. Both SPEC and GB are suites and not just single benchmarks, so this makes me think perhaps Krait isn't the A15 competitor it's made out to be. Nevertheless, at least in GB, Krait tends to just edge out Atom's per-clock performance in the single-threaded tests, on average.

    Apple's A6 cores OTOH are surprisingly powerful. In fact, I predict they'll turn out to be more than a match for Cortex A15, both in IPC and perhaps even in absolute performance once they get clocked higher in the next iPad. Likewise, given that Intel's in-order Atom is still competitive with the latest out-of-order ARM cores, and how easily Intel's current out-of-order designs annihilate even the A6, I think Intel will have little trouble overtaking ARM from a performance standpoint once the next-gen Atom is released.
    Reply
  • Wilco1 - Sunday, September 30, 2012 - link

    Spec is well known as the most gamed benchmark as it depends on the compiler and options used. If you think Intel did this test fairly then think again... If you look at less biased benchmarks then you'll see a Cortex-A9 beating Atom by a big margin. For example in Geekbench a 1.4GHz Galaxy S3 easily beats a 1.6GHz Medfield on either single or multithreaded benchmarks.

    Eg. http://browser.primatelabs.com/geekbench2/1090947 (Z2460 score 882)
    and http://browser.primatelabs.com/geekbench2/1107936 (Exynos4 score 1773)

    So your conclusion that Intel's in-order core is competitive with ARM out-of-order cores is based on Intel's marketing claims rather than reality.

    Krait is a little faster than an A9 clock for clock. However given that it turns out the A6 actually runs at 1.2-1.3GHz, it will come close but I don't expect that it will match the A15 (which is about 50% faster than an A9 clock for clock).
    Reply
  • somata - Monday, October 01, 2012 - link

    Fair point about SPEC, but like I said, in my experience Geekbench results lead to a similar conclusion. The results you provide paint Atom in an especially bad light since you're comparing a 1C Atom to a 4C Cortex A9. That may be a fair comparison as far as overall platforms go (for now), but I'm only referring to the strength of the core; going multi-core is trivial.

    I never said that Atom necessarily offers better performance than competing ARM cores, only that it's competitive, which it certainly is. Notice that Krait somehow loses most tests to slower-clocked A9:
    http://browser.primatelabs.com/geekbench2/compare/...
    Against Atom it's basically a wash (again focusing on single-threaded performance). Also, Android on x86 is AFAIK still pretty immature, so I'm not sure how much the results you cite can be trusted. In Linux the same Atom would score a bit over 1000 overall and win several tests against the A9.
    Reply
  • Wilco1 - Monday, October 01, 2012 - link

    Comparing single-core Z2460 with a quad-core Exynos is reasonable as those CPUs are used in the latest phones. The Exynos is 33% faster on the integer benchmarks and 68% on the floating point (single-threaded only, at the same frequency). I do not call this competitive. Dual cores and a higher turbo won't be enough to make it competitive against current Exynos4/Tegra3, let alone the A15-based next generation.

    Yes, it seems current Krait's are hobbled by a slow memory system or a bad memory controller, the memory scores look really bad. QC have to fix this and increase frequency to ~2GHz if they want to stay competitive.

    Android is based on Linux, so I'm not sure why you think Linux would improve scores. It's true that netbook Atoms have better scores, but these have a much faster memory system and use far more power. I've seen Phoronix results comparing OMAP and Exynos development boards with netbooks on Linux, and the results are the same: an in-order Atom is only competitive if it is clocked much higher. For mobiles and tablets this isn't feasible because of power.
    Reply
  • somata - Tuesday, October 02, 2012 - link

    Indeed I was using an Atom N455 as a stand-in for the Z2460 since I believe both share nearly the exact same microarchitecture. I thought maybe I overlooked differences in the memory interface, but both have 6.4GB/s of available bandwidth. With that in mind, I can't see why the N455 should have ~20% better IPC than the Z2460 in Geekbench other than the OS difference...

    Then again, I'll grant you there may be lower-level differences in the memory/cache interface that negatively affect performance. If I concede that the Z2460 is accurately portrayed in the handful of available GB results, then its per-clock performance is in fact closer to that of a Cortex A8. I'd still argue its overall performance is high enough to be "competitive" with an A9 though, in the sense that they will provide very similar user experiences. Exynos may offer much greater peak performance with its four cores, but I'd bet typical core utilization is even lower than on the desktop, further narrowing the effective gap.

    Don't get me wrong, more performance is always nice, and it's unfortunate that Intel isn't introducing a new architecture until 2013, but given the relatively simplistic usage scenarios of current tablets, I have no doubt Intel will be ready in time for the real "post-PC" revolution. ;-)
    Reply
  • Wilco1 - Tuesday, October 02, 2012 - link

    Yes the micro architecture is nearly identical. However LPDDR2 is slower than DDR3 and I'm not sure whether all Z2460 based phones utilise the fastest memory speed. Also the Z2460 can only turbo up to 1.6GHz for a short time, unlike the N455 which can continuously run at 1.66GHz (but uses 6.5W). So a Z2460 is never going to be as fast as the N455.

    Note Z2760 results have now also appeared on Geekbench: http://browser.primatelabs.com/geekbench2/compare/...

    I agree performance seems good enough, but it just isn't better performance like Intel's marketing claims (which is how this thread started).
    Reply
  • tuxRoller - Tuesday, October 02, 2012 - link

    I don't what specint tests and GB doesn't release their code, so I don't know how they test.
    Assuming these numbers are accurate, I would have to agree that Krait is not a next gen arch. A 5% architectural advantage is more a bump than anything else.
    Assuming the cortex-A15's are all they are supposed to be, they will blow all the rest of these chips out of the water.
    The main advantage the Apple cores seem to have is that they are using lpddr-1066 memory (looking at the GB breakdowns, it is their massive memory bandwidth and streaming perf that gives them their advantage). That seems to be what accounts for the difference, but perhaps when AT does the review we'll here more.
    Reply
  • tuxRoller - Tuesday, October 02, 2012 - link

    Actually, the difference isn't 14% but only 5% at the same clock, for that benchmark.
    I wasn't aware of any benchmark where krait performed that poorly compared to cortex-A9.
    Sure, Intel may have picked this benchmark b/c it makes them look particularly good, but I'm more concerned than with the A9 vs krait numbers.
    A 5% architectural advantage seems absurd.
    Reply
  • Krysto - Friday, September 28, 2012 - link

    Clover Trail is still not competitive with ARM. It needs a 30 Whr battery, which is more than twice as large as most 10" tablets with a similar resolution (not counting iPad which has a much larger resolution. Clover Trail won't even support such resolutions, because they know battery will die too quickly).

    So even if they achieve the 10 hour mark, which I still doubt it. They're doing with a double battery. Doesn't exactly make for a "competitive" chip with ARM, does it?

    They're being misleading about a ton of other things, too. I suggest you give this a read:

    http://semiaccurate.com/2012/09/27/intels-clover-t...

    Also, it's a chip that's launching at the end of 2012. Where is the OpenGL ES 3.0 support?
    Reply
  • Death666Angel - Friday, September 28, 2012 - link

    No idea what you are talking about battery wise. Most 10.1 tablets I know of use ~25 Wh batteries ( Asus, Samsung, Acer, Chinese tablets). Reply
  • Penti - Friday, September 28, 2012 - link

    The New iPad (iPad 3-gen) uses a 42Wh battery which is actually a larger battery then MacBook Air 11.6. Android tablets tend to use less power consumings SoC's and displays so they do fine on 20-30Wh. Down to about 15Wh. As does x86 Atom stuff when Ivy Bridge does fine on 35Wh (MBAir11). Current atom-powered Tablets PC's like HP Slate 2 or Dell Latitude ST uses a 30Wh battery and they are way more power consuming then these chips. Granted though that those are only good for 6 hours but are more power hungry. They don't need a 60Wh battery here to achieve 10 hours, not 50Wh either. iPad 2 uses a 25 Wh battery. With it's 1024x768 resolution.

    Z2760 is a 1.7W chip rather then a 3.75W platform though. Max TDP estimates of course. If you save just 2W every hour that is still 10 Wh for a 5 hour usage period. It's still mostly WLAN, Display, WWAN etc that uses power in these devices.

    Semiaccurate article is ridiculous for several reasons first there is no Office 2013 RT yet, there is no Outlook for WOA/Windows RT. There won't ever be any Outlook or Microsoft powered Enterprise email client for Windows RT. They are devices and are managed way worse then even iOS and Android tablets are. Then again their competition aren't Android and iOS tablets, and Atom computers is an full feature computer. It's managed by IT as a computer, it's not competing against Windows RT devices and it can run Office today where there is none available for Windows RT at release. Plenty of ARM SoC's use more silicon then the Intel Clovertrail in mobile variants. That it is designed for Microsoft is moot when ARM-powered Windows RT devices will be just as locked down if not more. It's not like you can virtualize another OS on the ARM-devices either. Drives for PowerVR on Windows RT won't exactly be any better either it's still WDDM drivers. What this chip competes against is Brazos 2.0 and Hondo APU's and coming Temash/Kabini APU's from AMD. Temash will have GCN-graphics but Intel's alternative will be on their own Intel graphics by around then. Besides Atom Z2580 is capable of running Android. If that is your gripe.

    Z2760 is already pretty much out in products today, reviews of pre-production stuff is already around. Availability is 26 October. I.e. Windows 8 release. They are fully competitive against early Windows RT ARM-tablets. Even battery wise. They do run Office 2010/2013 beta, where you have nothing for Windows RT today. If you look at Surface RT vs Pro battery, RT will have a 31.5Wh battery and Pro will have a 42Wh battery, but that is a Ivy Bridge powered device! A Samsung Galaxy Note 10.1 uses a 25.9Wh battery. The Galaxy Note 10.1 battery weights 136 grams. It's still much smaller then the 42.5Wh battery of iPad 3-gen. No real issue here. How is that more then twice as large? iPad is a 9.7-inch device. Note 10.1 is of course a 10.1-inch. It's roughly twice as large as a 7/7.7-inch device but much of that comes form the display! It's note twice as large as Note 10.1 or Tab 2 10.1 with the same 25.9 Wh battery for Tab 2. Add 25-30 g and you fit the slightly larger battery in there.

    PowerVR Series 6/Rogue aren't out in ARM-SoC's yet either, but Intel's own GPU technology does support Windows/DX/OGL fairly well compared to PowerVR-Windows drives. When you get Valleyview next year with Intel graphics, it should work pretty well as a computing platform. After Saltwell based stuff you also get Silvermont (like Valleyview). It's not like Z2760 are suppose to compete against the competition in a years time, late 2013. It's not a new generation chip it's still 2008 Atom-architecture/pipeline. Z2760 is only aimed at tablets where 1.7W TDP is actually okay. Android stuff can always use Z2460, Z2560 when available or use whatever else Hondo aren't meant to run it but you can always make it run, should compete well with Windows-based devices though.

    Intel's reference platform here on Z2760 uses a 30 Wh battery just read the damn press release, it's video playback with WiFi activated, but local playback. Over 10 hours for that scenario is totally reasonable. The only real difference on this platform against full featured notebook chips is that it will run 32-bit Windows rather then 64-bit. The platform will use about the same size batteries as ARM-based Windows RT devices. But for summary i.e. 25Wh iPad 2 is not significantly smaller then a 30Wh battery Z2760 device.
    Reply
  • relic1138 - Tuesday, October 16, 2012 - link

    I will bet anything that this CPU will get you a max of 6 hours of battery life when it's first introduced. I have never seen a 10 plus hour Atom system, real world not paper specs. My Samsung 7.7 Tab, now that get's 12 hours consistently. Reply
  • Ailuros - Saturday, September 29, 2012 - link

    SGX545 belongs to the Series5 family as ltcommanderdata already noted and therefore has four Vec2 ALUs, a quite strong triangle setup unit and 16 z/stencil units. Series5 is not configurable in multiple cores, it's just one core in Cedartrail/Clovertrail clocked at 533MHz.

    Apple's GPUs since the iPad2 have Series5XT GPUs which can scale multiple cores and the SGX543MP3 in A6 should have a frequency of 300MHz at least.

    In terms of triangle throughput a SGX545@533MHz vs. a SGX543MP3@300MHz should be on comparable levels, the MP3 has almost 60% more texel fillrate, about 70% higher z/stencil/pixel fillrates and >120% more ALU throughput.

    Frequency by itself doesn't define final performance. As for Clovertrail Intel stated DX9.1 there (whatever that means...probably DX9.0L3) and OpenGL3.0. OGL3.0 is far and beyond above DX9L3 capabilities so it's probably just an issue with IMG getting WHQL certification for DX10.1 for Cedartrail initially and now probably also for Clovertrail.

    Other than that Apple's SGX543 GPU cores are just DX9.0, while SGX544 is DX9.0L3, but under iOS anything above DX9 seems to be redundant for the time being. A windows platform is an entirely different chapter.
    Reply
  • fteoath64 - Monday, October 01, 2012 - link

    Good estimation on the the gpu performances. We shall see soon when the benchmarks are released on an equal (as close as possible) footing. The good news for Intel is that the power consumption of this chip is respectable but clock for clock, it is almost on par with Tegra3 and considering Tegra 3 has only a single channel memory and a 12 pipeline gpu. The next iteration of Tegra3 will be significant. Even if Nvidia choose to release a slightly tweaked Tegra 3.5 (eg dual ch LPDDR3 1333Mhz, 16pipe Geforce update (400mhz --> 600mhz turbo core) and cpu cranked to 1.8Ghz. This will step the Tegra another 30% ahead of this atom!. Also Intel with their cheapo 32nm process where it could very well do 22nm runs since this is going to be small volumes due to very few customers ....
    Almost there but the competition just went another step or two further.

    Does it look like Intel is intending to "lock" ARM out of the low-end of the Win8 tablet market ?. If they later slash prices aggressively, it might just work for them. But the issue is that Android is relevant and will keep up its momentum in the tablet space for a long time to come. Win8 RT might see a limited run due to high prices!. That would be a shame.
    Reply
  • Wilco1 - Monday, October 01, 2012 - link

    There are indications that Win8 tablets will cost much more than WinRT ones despite WinRT having faster CPU/GPUs, lower power consumption and likely better resolution screens. Win8 makes sense if you absolutely need x86 compatibility and are willing to pay more. That would make Win8 popular with businesses and WinRT with the general public.

    So on the consumer side I predict the battle will be solely between Android, iOS and WinRT tablets. Businesses are likely to go with Win8 by default.
    Reply
  • relic1138 - Tuesday, October 16, 2012 - link

    No indications they will cost more, Samsung and Asus in Europe have both posted pre-order pages for their tablets running this CPU. They start at 600 without keyboard and 700 with, this is for the bottom configs. Reply
  • thebeastie - Tuesday, October 09, 2012 - link

    I just cant help but think about the fact that if I am in a Windows ARM RT environment any anti-virus scanner I have wont need to look for like 1 million+ viruses since they cant run x86.. Just makes me feel a lot better.

    I think the only fair benchmarks you can run is an Clover Trail Atom Win 8 tablet WITH active AV going in the background VS a Windows RT tablet.
    Cant help but think that the Clover Trail will need all the extra grunt it can muster to hold onto some performance, and it better be doing that AV protection with out sucking the battery, yeah good luck with that.

    Anyway will see in the benchmarks, oh yeah its not going to be benchmarked to satisfy the average user setup with active AV installed is it?
    Reply
  • Wolfpup - Friday, October 26, 2012 - link

    I'm impressed that Intel really seems to have gotten a just plain better product here...looks like they're beating performance, matching or beating power, and of course most importantly...this can run real Windows, your real programs, etc.

    Very cool, but I have to wonder, what about AMD's c and e series CPUs?

    Are they still manufactured at 40nm? Do they not support this new polling stuff? If not, then I guess this Atom would make more sense...but I'd still feel better about having an actually-kinda-modern-ish dual core 1.6GHz (or whatever) e chip from AMD...or of course an A series chip would be ideal, since you can actually plausibly use an A8 or A10 as your main PC, even for games...
    Reply

Log in

Don't have an account? Sign up now