Intel’s upcoming Atom Z3770 processor is a low power, quad-core chip that will be part of the Bay Trail family. It uses the company’s new Silvermont technology, which is expected to power upcoming Atom, Celeron, and Pentium chips, although the Z3770 is a lower-power chip that’s likely aimed at Windows or Android tablets.

Benchmark results for the new chip showed up recently at the Geekbench website, which provides us with an idea of what kind of performance to expect — and gives us a way to see how the new Atom Z3770 chip compares with other processors.

Geekbench Bay Trail T Atom Z3770

While the Intel Atom Z3770 chip is expected to run at speeds up to 2.4 GHz, the benchmark results are for a quad-core chip with a clock speed of 1.46 GHz. It was tested on a device running Windows 8.1 Pro Preview 32-bit.

Geekbench gives the chip a score of 2093.

That means it outperforms a system with a 2 GHz quad-core Intel Atom Clover Trail+ chip (Geekbench Score: 1473), or a 1.8 GHz dual-core Clover Trail chip (Geekbench Score: 1425).

In other words, the new processor is faster than any Intel Atom chip released to date, at least according to Geekbench. If faster versions of the chip really do run at higher speeds, the gap could be even wider.

But while Bay Trail chips are faster than any Atom processors that came before (and more energy efficient, for that matter), these are still low-power chips that won’t offer the same kind of performance you’d get from an Intel Core processor — or even a mid-range AMD processor, if Geekbench’s comparisons are to be believed.

For instance, a computer with AMD’s A6-1450 “Temash” chip scored 2769 in the same test. And a model with an AMD A4-5000 “Kabini” processor did even better, scoring a 2893.

Both of those chips are expected to use more power than the Intel Z3770 processor. AMD’s quad-core Temash chip has a TDP of 8W, and the quad-core Kabini is a 15W processor.

Intel hasn’t provided a TDP figure for the Z3770 chip yet, but it’s expected to have an SDP of 2W. That’s Intel’s figure for average power use under typical conditions, and it tends to be a bit lower than the Thermal Design Power or TDP which chip makers have historically used to measure a processor’s power consumption.

But I’d be surprised of the Z3770 uses as much power as the AMD A6-1450, which means that you’ll likely find this Intel processor in tablets or notebooks which provide long battery life and moderate performance.

Bay Trail

Bear in mind, the Z3770 is one of the least powerful processors in the upcoming Bay Trail family. Other chips based on Intel’s Silvermont architecture are expected to consume more power, but offer the kind of performance you’d expect from a laptop or desktop chip, and some will be sold under the Celeron or Pentium names to help differentiate them from the Atom processors, which have earned a reputation as relatively slow chips.

Also keep in mind — these Geekbench results appear to be for an unannounced, pre-release chip. While it’s likely that someone at Intel or one of the company’s partners is running some tests on a prototype, there’s no guarantee that the processors which hit the streets later this year will score the same on benchmarks as the pre-release models.

via @luka, @DresdenBoy, and Investors Hub

Support Liliputing

Liliputing's primary sources of revenue are advertising and affiliate links (if you click the "Shop" button at the top of the page and buy something on Amazon, for example, we'll get a small commission).

But there are several ways you can support the site directly even if you're using an ad blocker* and hate online shopping.

Contribute to our Patreon campaign

or...

Contribute via PayPal

* If you are using an ad blocker like uBlock Origin and seeing a pop-up message at the bottom of the screen, we have a guide that may help you disable it.

Subscribe to Liliputing via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 9,544 other subscribers

17 replies on “Intel Atom Z3770 Bay Trail chip specs, performance leaked”

  1. Isn’t that geekbench score only half of a Tegra 4 (4148) ?. Sounds dissappointing.
    I wonder what the Antutu will be now that Intel were reportedly caught cheating with Antutu scores for clovertrail.

    1. Geekbench is suppose to be OS agnostic but different OS can produce different results as how memory is managed effects the test and Geekbench includes testing the system memory performance…

      The Tegra 4 scores may also be inflated because they often test versions of the SoC clocked to the max but going into actual tablets and other mobile devices often means lowering max clocks to keep TDP low enough for fan-less designs.

      The Qualcomm Snapdragon 800 for example can have a load of 4W for either the CPU cores or the GPU, which means a max load of up to 8W but ARM SoCs typically keep average operational usage within 5W or less…

      So there’s a difference between real world usage and reported benchmarks by companies like Nvidia.

      Mind, Silvermont Architecture can scale higher as well and if this report is accurate then it’s not even running at the max clock for the Z3770, which is reported as 2.4GHz, and it is instead operating at 1.46GHz for the benchmark result reported…

      Mind, it’s also rather dubious that the Tegra 4 could score that much higher than the Kabini, which is not targeting the tablet market like Temash is, and is a much higher 15W APU!

      While another factor could be that since AMD hasn’t added support for Android yet that the tests were performed on Windows 8 and Geekbench may not be properly tuned for that OS yet…

      As to AnTuTu… the issue was the use of ICC compiler instead of the GCC compiler… Since ICC could be relevant for say Windows 8 but not for Android and the SoCs they were testing were all only going to run Android…

      So AnTuTu corrected the issue and the results went down about 20%, which falls better in line with what other benchmark results were giving…

      Benchmarks in general though don’t really reflect real world usage… So what really matters for the average user is to see reviews showing real world usage for proper comparisons and none of the potential fake one oneupmanship that tends to plague benchmark results…

        1. No, you said it means nothing!

          Even now that it’s revealed to be 20-30% off the actual score still shows how much the performance has improved.

          While that result was from a different test system that was clocked even slower than the results from this Geekbench result.

          The Z3770 is a quad core that according to this test was being tested at 1.46GHz but it’s a 2.4GHz rated part.

          While the unknown model that was tested in the other Article was only suggested to be a 2.1GHz part and could have been a dual or even single core as Intel is still giving a range from single to quad cores.

          How these Geekbench compare to the quad core Temash and Kabini at near equivalent clock speeds shows how close performance is now. Since the older ATOM had a much wider gap in performance.

          While memory performance could have been a factor as well because the Bay Trail T can only use LP-DDR3 RAM but both the Temash and Kabini are using faster DDR3L RAM.

          While the Kabini is a 15W part and the quad core Temash is also using near 15W when in its max performance mode.

          Compared to Bay Trail T Z3770 that should be less than 5W, with just a 2W SDP rating for it’s 2.4GHz state.

  2. It would be interesting to see a comparison between z3770 and A4-1200 or 1250 as all these chips will be used in fanless tablets.

    1. Well, the A4-1200, with 3.9W TDP, will be but I’m not so sure about the A4-1250 as that increases the TDP to 8W and fan-less designs typically needs to keep around 5W or less usually… at least for average operational usage.

      Especially for tablets because of the thin designs make it harder to rely on passive cooling to effectively dissipate heat and the small designs don’t usually give that much surface area for the heatsink to work with…

      As for performance, there hasn’t been much reported about AMD Temash SoCs, except for the A6-1450 quad core… but a reference tablet running the A4-1200 was shown…

      https://www.youtube.com/watch?v=tLGj3_246LA

      At 1:04 with the detailed Windows Experience Index score…

      3.1 – CPU
      4.5 – RAM
      3.9 – GPU
      5.6 – 3D Graphics
      7.5 – Primary Drive (SSD)

      In comparison, a Dell Latitude 10 running on the Clover Trail 1.8GHz Z2760 gets…

      3.4 – CPU
      4.7 – RAM
      3.8 – GPU
      3.3 – 3D Graphics
      5.5 – Primary Drive (eMMC)

      Of course the WEI is only a general estimate but I wouldn’t expect too much more performance than you would have gotten out of a Hondo Z-60, they are clocked the same and Jaguar only has up to a 20% performance advantage over Bobcat cores, but at about 0.6W lower TDP, as the Z-60 is rated as 4.5W for its max TDP…

      So you can look at longer battery life than you would have gotten with a Hondo but it seems that’s about it for that target range APU…

      Though the A4-1250 should get a good bump in graphical performance because it clocks the iGPU 75MHz (300MHz) faster than the A4-1200 (225MHz) and is apparently the main reason for the higher TDP rating…

    2. Thanks for all those details CyberGusa. I am trying to get tablet as an HTPC and something I can bring alone on my way to work. Now I’m not sure either Temash or Silvermont would do me any good. I was expecting nothing less than a big leap from the current Hondo and Clover Trail. Sigh..

      1. Well, mind that most of these results are for the very low end models.

        The A4-1200/A4-1250 is only a dual core APU and most of the promised performance increase is for the quad core models.

        Like the Geekbench results show above, the 15W TDP quad core A4-5000 Kabini got the highest score. While the next best was the quad core A6-1450 Temash… Only problem with AMD solutions is the still high TDP for performance, since the quad core Temash goes from 8W for its tablet optimized mode to about the same 15W for its Turbo Mode.

        While for Bay Trail… The Z3770 was reportedly running at 1.46GHz but it’s a 2.4GHz rated part… So the Geekbench result may not be its maximum that it can achieve… But that model is also optimized for tablets and will supposedly fall into a lower TDP range than Bay Trail M (Celeron) and Bay Trail D (Celeron/Pentium), which according to the model numbers revealed for those will go from 5W to 10W range models.

        So you could still see a higher performance Bay Trail that may suite your needs.

        Similarly, as long as you don’t need to be too small and run cool… The Kabini also scales up a bit with the A6-5200 clocked at 2GHz, versus the 1.5GHz for the A4-5000 model…

        The difference between the quad core Temash and the A4-5000 Kabini in the Geekbench result is because the A6-1450 Temash maxes out at 1.4GHz.

        So you should see a bigger difference in performance between the two versions of Kabini with the 500MHz clock speed difference.

        While Bay Trail D, will also scale up to Octo (8) Cores… So for a desktop HTPC, the max performance Bay Trail could offer may even exceed what Kabini will offer.

        Though, I know you specified tablet for HTPC usage but if they can squeeze a Core i5 into a 10.6″ MS Surface Pro then they can definitely put either AMD Temash/Kabini or Intel Bay Trail’s higher range solutions into semi-Pro tablets.

        So, I’d suggest keeping an eye out for those type of models… Performance won’t be better than a Core processor based system but they should be more affordable and more likely to be in a thin and light design that’ll make them easier to carry…

        Though, if money isn’t a issue… the upcoming Haswell Y series chips should definitely be a good solution that should also provide good battery life…

  3. I really hope Linux support for Bay Trail is decent when devices start coming out. I’m hoping to get a small Bay Trail device (10″ or smaller screen) and install Arch Linux or some other distro on it.

    I’m concerned about those leaked slides from some time back ( https://www.abload.de/img/45nl9q.png ) where Linux isn’t a target OS for Bay Trail T. It only supports 32-bit Windows. There could be some things in Bay Trail T that Intel doesn’t support with their Linux drivers.

    I wonder if Bay Trail M chips would make it into 10″ or smaller screened devices. Since the slide says Bay Trail T has a TDP of <= 3 W then maybe the lower end 4 W TDP Bay Trail M might make it into these mini devices.

    1. Oh wait, Bay Trail M isn’t targetted for Linux either but it seems Intel is specifically trying to support Chrome OS though…

      1. Linux supports booting for Secure Boot enabled devices as long as the system is 64bit… Silvermont is a 64bit architecture and it just depends if they allow that on Bay Trail T… M and D will definitely support it though…

        It’s up to the OEMs if they allow Secure Boot to be disabled, but part of the lack of support right now for the present ATOMs is because the GMA lacks support for Linux but Silvermont uses a GMA based on Ivy Bridge GMA HD4000… So driver support should be there.

        Meaning it’s very likely you will be able to run any Linux distro you want, but it may depend on the OEM who makes the system!

    2. I’m hoping to get a Bay Trail based UMPC running Linux. It’d be nice if a 5″-7″ screened device with maybe a slide out thumb keyboard (split design for 7″) with a mouse comes out.

      If need be, I’d settle for a Linux virtual machine since Bay Trail supports hardware virtualization acceleration. However, from that slide, Bay Trail T may be limited to 32-bit. Less than 4 GB of RAM available for VMs plus overhead may not be enough. Also, I use VMware Workstation and the Windows version requires a 64-bit CPU even when running a 32-bit Windows host.

      1. I’m hoping for a 7″ screened device or a 5″ screened device with slideout thumb keyboard. Both should have a thumb mouse integrated (much more useful than a Wacom stylus for non-artists and inkers). It should be chargeable through microUSB at at least 10 W (lower Watt charges and PC USB ports should still be able to trickle charge it without damaging the source as well via some sort of sense circuits and negotiation when connected to a PC).

        I’d definitely want to run a Linux destkop distro on it preferbly not as a virtual machine (probably not Ubuntu Touch until it’s been out for a few years). I have several Bash scripts and console C++ code I’d like to run for automating tasks.

        I might run Windows 8 in a virtual machine just in case. Not sure what for right now but it’s good that Bay Trail supports virtualization acceleration. Hopefully, Intel doesn’t disable it on Bay Trail T chips which they sometimes do.

  4. Seems like the Intel chips are better for mobile and AMD is better for budget desktop. Sounds familiar. 😉

    1. Possibly, at least for Kabini but remember that Intel is also making higher end Bay Trail chips that will be marketed under Celeron and Pentium brand names.

      Bay Trail D, specifically, can scale up to Octo (8) cores… though, that’s not counting custom designs from either…

      AMD should be able to push better graphics though, but the gap between them is much smaller than the old netbook days when AMD had 5x to 9x performance lead on graphics… More like the difference between Intel HD4000 and AMD Trinity performance differences in around 40% and no longer multiples!

Comments are closed.