Taiwanese chip maker MediaTek’s upcoming Helio X20 is expected to be one of the first mobile processors with 10 CPU cores. But the company is already working on a few more deca-core chips.
The Helio X22 is basically a faster version of the X20, while the Helio X30 uses a different arrangement of processor cores to offer higher performance.
Like many multi-core chips for mobile devices, the Helio X series basically combines a number of higher-performance processor cores with a group of lower-performance cores that use less power. The arrangement can help provide a burst of power when needed, while relying on the less power-hungry components to conserve battery life when you don’t need turbo speeds.
What makes these chips different from most on the market is that they combine three or more groups of CPU cores instead of two.
Here’s a run-down of the way those cores are set up in different Helio X series chips:
- 4 ARM Cortex-A72 cores clocked at 2.5 GHz
- 2 ARM Cortex-A72 cores clocked at 2 GHz
- 2 ARM Cortex-A53 cores clocked at 1.5 GHz
- 2 ARM Cortex-A53 cores clocked at 1 GHz
This chip will have ARM Mali-T880 graphics and support for up to 4GB of RAM, and eMMC 5.1 storage.
- 2 ARM Cortex-A72 cores clocked at 2.5 GHz
- 4 ARM Cortex-A53 cores clocked at 2 GHz
- 4 ARM Cortex-A53 cores clocked at 1.4 GHz
MediaTek’s first deca-core chip features ARM Mali-T800 graphics, support for 4G LTE, 802.11ac WiFi, and an ARM Cortex-M4 co-processor for always-on speech recognition, among other things.
Details about this processor are scarce, but it’s expected to have the same type of arrangement as the Helio X20, but higher clock speeds for at least some CPU cores.
You’ll probably have to wait until early 2016 before you can buy a device powered by any of these new processors, but MediaTek is expected to begin shipping samples of the Helio X20 later this year.
via PadNews, GizChina, and Weibo
Given sufficient Linux support, that X30 could power a full-blown desktop environment with the six Cortex-A72 @2Ghz+.
I would definitely buy a Chromebook or a Linux-based netbook with that SOC if its not too skimpy on the speed of RAM and the SSD.
I saw the benchmarks of mt8173 vs goldmont and arm a72 is at par with the latest atom architecture now. Where are the mt8173 tablets?
goldmont? that will release next year… you do mean aitmont, the successor of silvermont, dont you?
Are there any results on how effective the big.LITTLE configuration is? With this kind of chips, are there any tests like various server loads, sustained high loads and any highly varying types of work? That is, if this can be accurately compared to non-big.LITTLE chips.
Big.Little is about power well gating for leakage. Mobile chips are on battery power and server chips are on Wall power. Server cpu leakage is just an electricity expense, and the customer would probably want more big cores than small cores to reduce the electricity cost. Mobile leakage is very important as it dramatically increases battery life. In summary, Big.Little is only for reducing leakage in high performance mobile cpus.
The gates/cells used for the high performance cores are very different to the power efficient cores. The leakage can be 10x larger. This means that the power well needs to be completely gated or the battery could be dead in an hour or two. a72 at 2.5GHz is only possible in a mobile device with Big.Little (or the device would need to be completely down when the screen is off).
You do realize that this doesn’t make any sense and that the source has no real credibility.
The assumption about timing also comes from someone that is not even aware of publicly available info.
As for X20, Mediatek said on Friday that very small volumes will ship in Q4. They are going 16ff in the first half of 2016 for high end and in the second half for the mainstream (and that’s official info) but beyond that there will be 1000 random rumors from China and vast majority are clearly very unlikely.
16ff in the mainstream next year is kinda exciting, it’s sooner than expected , the competition is heating up and users will benefit. I do hope MTK or some competitor go for 2 big cores +2(or 4) small cores instead of many small cores in midrange.There is no need for midrange SoCs to be slow on the CPU side, they can offer better than iphone CPU perf easily in 150$ phones, someone just needs to make that SoC.
I wonder when Intel will take notice? They are trying to hey in this space. If they are not careful they will be a distant memory. Yes I know today media tek is not the large that but they are not standing still
I think Intel is doing OK for now. Every generation they get substantially better here. Atom is pleasant enough compared to higher end stuff (lacking in 3d maybe though?). I guess Atom cores in an i5/i7 laptop chip would be interesting, but that’s a higher power level playing field.
I’m actually more curious where the “tier 1” ARM companies are at – Samsung, Qualcomm?
Qualcomm’s Snapdragon Krait was ahead of it’s peers, being their own design. Last chip from qualcomm was a standard ARM design. I’m hoping to see that ingenuity again.
I’d like it if they’d finally enter the US market since they seem to offer a lot of value for money.. every time i see a MediaTek phone it never has US based LTE bands… they have to be better than Qualcom’s overheating disaster which kinda defeats the purpose of an ARM chip..
Actually Mediatek is very big, there are only a handful of chipmakers bigger than them , excluding memory makers (Samsung, Hynix, Micron, Toshiba) and even Intel since Intel has a monopoly allowing for very high prices. Excluding those only Qualcomm, Avago/Broadcom (Avago and Broadcom are merging), Freescale/NXP (these 2 are merging too) and TI are bigger in revenue.
Intel has no relevance in phones now really They do have a minority stake in Spreadtrum and Spreadrum shipped 100 million units smartphone SoCs lats year- low and very low end. Their Atom has no relevant market share in phones. I do wonder if the next iphone isn’t using an Intel modem but not going into that subject now. Intel could try to pay some 30 billions to buy Mediatek but it’s not really worth it for them to do that. The mobile segment is not worth it at all for them. In PC they are free to charge a lot for their products and they don’t have to make any good products, they can pretty much sleep at the wheel while the market keeps declining. In mobile there is competition and the chips are much smaller so much cheaper too. Plus they would have to make good products and they can’t even do that. Without a monopoly Intel is in trouble ,the vast majority of the money they make are because of it and there is no good way out for them. They’ll be just another chip maker, instead of being ridiculously oversized – something enabled by our lovely regulators/politicians and nothing else.
Comments are closed.