MediaTek’s Helio X30 processor is a 10nm chip with 10 CPU cores and PowerVR Series 7XT graphics.

The Taiwanese chip maker says the processor will show up phones with 2X optical zoom cameras, 4K HDR video support, and virtual reality features including support for Google’s Daydream platform.

But the MediaTek Helio X30 isn’t necessarily reserved for high-end phones: it’s expected to be used for handsets that sell for between $300 and $500. You’ll probably start to see some of those phones in the second quarter of 2017.

MediaTek unveiled the X30 last year, but now the company says mass production of the chip is underway.

The deca-core processor has different CPU cores for different purposes:

  • 2 high-performance ARM Cortex-A73 CPU cores clocked at 2.5 GHz
  • 4 lower-power 2.2 GHz Cortex-A53 CPU
  • 4 energy-efficient 1.9 GHz ARM Cortex-A35 CPU cores

The idea is that the Cortex-A73 cores kick in when you need a lot of power, but they can shut off and let the other CPU cores handle the work when they’re not needed, which helps balance performance and energy efficiency.

MediaTek says its new 10-core, tric-cluster chip offers 35 percent better performance and 50 percent more power savings than “MediaTek’s previous generation chipset,” which probably means the Helio X20/X25 chips which also use a 10-core setup, but which are 20nm chips that rely on different ARM technologies for some of their components.

Among other things, MediaTek says the new chip supports up to 8GB of LPDDR4X RAM, up to two 16MP cameras with 4K video recording capabilities, 4G LTE Cat 10, and 802.11ac WiFi.

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 9,545 other subscribers

8 replies on “MediaTek Helio X30 deca-core chip coming to mid-range phones in Q2, 2017”

  1. There’s three things wrong with this:
    – Availability very low
    – Triple cluster cores are IN-efficient
    – MediaTek gives no sources, drivers, documentation, and support

    What’s great?
    – 10nm competition, yay!
    – More ARM Cortex A73 chips
    – Finally!!!, the Cortex A35 chips hit smartphones (honestly they’re awesome)
    – Top of the line RAM, big bandwidth, no bottlenecks!
    – PowerVR’s 7XT GPU is very powerful.

    I would’ve preferred they:
    – Be open with drivers, sources, documentations, and support
    – Used the new developer-friendly GPU from ARM Mali G71MP16 (900MHz)
    – Stick to simple design:
    Single-Cluster Design: 4x ARM Cortex A73 (0.6GHz – 2.8GHz)
    Dual-Cluster Design: 2x Cortex A73 (1.5GHz – 2.8GHz) + 2x Cortex A35 (0.3GHz – 1.6GHz)

    Such designs would provide almost equivalent performance at a greater efficiency.
    Why?
    Because having too many cores can mean you’re just burning through power without any performance gain, especially on Android with developers targeting the minimum requirements barely and not optimising anything for those overheads. We’re talking a difference of having a phone last 3 days battery life, over 2 days battery life…. I’ll take the more efficient extra battery life option anyday.

    1. Not fully agree :
      – availabily : as Qualcomm limit usage with unfair costs on patents (see chineese or Korean jugements or US and EU investigations), There are not enough customers for Mediatek
      – Tri clusters proove their efficiency on X20. Despite 20nm and Emmc5.1, CPU usage is as fast as 14/16nm to open apps or website. It’s right that a few apps are badly code (for exemple Chrome isn’t able to use Cortex A72 on X20…All others Browser can include Chrome Beta and Chrome Dev). You really should try X20.
      – sources become more and more available (not fully and not fast enough that’s right)

      Single cluster is suicide for saving energy.
      Two more A73 drain more power than a full cluster of 4A53
      Too heavy GPU often shows hard throttling when sustained full power is needed (Exynos 7420, 8890 – Snapdragon 805, 810, 820) : on some games sustained FPS are better on Stable Kirin 950 with MaliT880mp4 than on Exynos with mp12.
      The main goal of X30 is sustained performances as Mediatek explain it in videos.

      I regret that they don’t push frequency on A73…we will probablement see very soon X35 with 2A73 2.8ghz + 4A53 2.5ghz + 4A35 2.0ghz…in Solid Slate Circuits Conference, Mediatek shows a slide with theese frequencies

      With 10nm and UFS2.1, X30 may be one of the fastest soc this year…
      But I’m not sure they will sell many…unfortunately

      1. Agree to disagree.
        ….But MediaTek doesn’t pay royalties to Qualcomm, and doesn’t care about the western market. They’re focussed on China, India, and Brazil first and foremost. They aren’t limited in this aspect.

        And Qualcomm is using Samsung’s fabrications this time, so all the 10nm chips from TSMC is free to go to MediaTek.

        There’s nothing to stop MediaTek to be developer and community friendly. But they are businessman, and do not care. They steal open-source code and projects and never release their drivers. It is not changing, or not in any appreciable way. MediaTek goes as far as blackmailing and threatening their own clients; the OEMs, from releasing sources. MediaTek believes if they release sources, then a competitor will steal their drivers. Ironic, isn’t it?

        The Helio X20 was “okay” it was not “Great”.
        It had great idle power, and its top performance was good… but its sustained performance wasn’t that good at all. I can see what they were going for, but there’s a HUGE difference between theoretical effects and practical effects.

        A single-cluster of Cortex A73 is a BRILLIANT idea.
        Why?
        It’s because that is what the chip was actually designed for.
        It is NOT a “BIG chip” like the Cortex A15, A57, and A72.
        It is a “medium” chip like the Krait 400, Krait 450, and Cortex A17.

        Theoretically a 4x Cortex A53 + 4x Cortex A72 is faster AND uses less power than 4x Cortex A73. IN THEORY. However, if you’ve been following the market closely, you would’ve noticed that Android and Apps have a lot of trouble when shifting threads from one cluster to another. There is a noticeable lag, and this slows the entire system down (so you lose your “faster” advantage here). And as the system tries to fix and finalise the migration of the thread, you have moments where both clusters are working whilst not doing any actual processing (so you lose your “less power” advantage).

        In practice, the simple nature of the single cluster wins.
        Yes, the LITTLE.big concept was a flop.
        It’s been around for quite a long time, and things have only really improved this year on the matter (thanks QSD 821 and Google’s AOSP Pixel).

        So you either:
        – Stick to one cluster
        – Stick to a dual-cluster that is simple, and has a lot of bandwidth and optimisation to it.

        I tell you this, I rather have a phone:
        – Built on 16nm
        – 4x Cortex A73 (0.6GHz – 2.8GHz)
        – Mali G71MP16 (900MHz)
        – 4GB DDR4x RAM
        – 64GB UFS 2.0
        – Optimised

        Over a phone that is:
        – Built on 10nm
        – 2x A73 (1.9GHz – 2.4GHz), 4x A53 (1.4GHz – 1.9GHz), 4x A35 (0.3GHz – 1.4GHz)
        – Mali G71MP8 (850MHz)
        – 8GB DDR4 RAM
        – 128GB eMMC 5.1
        – Not Optimised

        Mobile tech is my name, and Efficiency is my game….

        1. What I understand in Korean jugement is that Qualcomm sells some patents to Mediatek but not All they need.
          Qualcomm refuse to sell any patents to Intel.

          The goal is to make the phone manufacturers (and not the soc Vendor) paid for the missing patents. So that they can force manufacturers to not buy others soc.
          Mediatek is interested in others markets…But market is closed by unfair patents usage.

          Agree that Mediatek must better release source code…With presure of campus brands it will go Gaster.

          Agree that A73 is “smaller” than A72/57 or A17 is “smaller” than A15…
          But you can’t compare A17 and Krait…launch a simple JavaScript test on both…Krait is far far far behind.

          Agree that storage Speed start being important…But Kryo weakness kill the avantage of UFS in real use…And X30 is able to use UFS2.1

          Agree about RAM, but X30 is able to use Lpddr4x…P20 too when S625 not.

          Not agree about GPU, bref to test before. Mp8 that can sustain performances can be better than mp16 that can’t.
          Mp16, 900mhz, 16nm is suicide on term of sustained performances…So you’ve got your benchmark score but nothing more in-game

          Agree about optimisation…Some X20 are faster than X25 despite juge différence on frequency.

          Disagree about lag during cluster migration…There isn’t noticeable one on X20. Corepilot 3.0 and Mediatek Coherent Interconnect prevent lag. X30 hasard corepilot 4.0.
          Corepilot is great, if you test 2.0 on mt6752 and compare it to S615, you know what I talk about…615 is far far far slower even if both are Octa A53.
          You should really try X20, it’s a great 20nm soc…It’s not a great soc because others use 16nm

          1. I’ll have to look into it more specifically.
            Is there any particular Helio X20 phone you recommend?
            I’ll try to get my hands on one and run some C++ scripts, maybe something that can tax the CPU, GPU, and DSP. It’ll be hard to run a controlled test on this, as I know its hard to tinker with MTK/Hwei chipsets to manually control each thread/core/governor.

            From my previous testing Qualcomm’s simple design won because it was practical.

            I’ll also see if I can get a loaner Pixel phone from Google to compare.
            Like I said, if you go on GitHub you can see plenty of OS-Kernel optimisations by El Goog. And this benefits other devices greatly that use LITTLE.big style computing.

  2. I’d love to see this in an Android gaming handheld. Something like the JXD S192.

Comments are closed.