Mediatek showed off the ARM Cortex-A17 bearing MT6595 SoC during MWC but all was not as it seemed. What could be the most advanced SoC shown publicly was tainted by dishonest marketing.
Lets start out with what the Mediatek MT6595 is and that is a very impressive SoC. Starting out with the core you have four ARM A17s along with four A7s in a big.little configuration. Those eight cores are backed by an Imagination Series6 Rogue GPU of unknown shader count backed by a similarly vague H.265 unit, no word if it is hardware, software, or a combination of both. This video decoder is capable of 4K video encode and decode, no easy task in a phone form factor.
53.15Mbps stream decodes on a phone
As you can see from the demo devices above, they are capable of decoding 50+Mbps H.265 streams. Unfortunately the display controller can only put out 2560*1600 images, more than enough for a phone or <10″ tablet, but not good enough for 4K TV output. It wouldn’t surprise us if the display out was capable of more though but Mediatek didn’t give out many tech details at MWC. In any case the ISP can support cameras up to 20MP.
On top of this add in just about every radio you would care to think about, 802.11ac, LTE in Rel9 Cat4 guise, DC-HSPDA+, TD-SCDMA, an a few more too. GPS, GLONASS, Beidou, Galileo, and QZSS are there for positioning plus Bluetooth LE for accessories. Basically the MT6595 has almost every mainstream radio standard you could want in a phone, not necessarily bleeding edge LTE stuff but damn near everything else.
On top of that you have a wireless charging, by itself nothing new but the MT6595 does it one better. Instead of picking one technology, inductive or resonant, this new SoC does both. Mediatek wasn’t clear if this is on the SoC die or on a separate chip. Either way if you are interested you can buy the same functionality in a MT3188 chip which was also shown off at MWC. Either way you can build an MT6595 device using whatever charging standard you want.
So if the MT6595 has just about everything you could want in a phone, why are we down on it? Is it the fact that Rockchip claimed to be several of the same firsts with the RK3288? Nope, in fact Rockchip was a no-show at MWC, not to mention they only announced the device, there was no silicon publicly shown at CES or MWC. Mediatek was definitely the first out of the gate with real working silicon.
Update April 3, 2014 @ 11:00AM: The situation is quite a bit more complex than this, the A17 is an A12 with significant pipeline improvements. On top of that the uncore and interface is massively updated to support big.little/coherent 4+4 CPU comfigs which the A12 does not support. So it is a bit more than the A12.1.
RTL for the A17 was not out at MWC but beta/non-final RTL was. It is SemiAccurate’s understanding that for a SoC to be released and to be called (insert ARM IP name here), it needs to use the final ARM RTL and be signed off on. The final A17 RTL hadn’t been released at the time of silicon showing and couldn’t have been signed off on.
This means that the MT6595 could be a pre-release A17 but it will need to be updated to match the final A17 RTL before it can be released. What you consider this to be is up to you but it looks like the situation wan’t as bad as we first thought. Apologies for the confusion. End Update.
The problem is the core, supposedly the worlds first ARM Cortex-A17. Why? Because the A17 doesn’t exist, it is a Cortex-A12 in that device. To be fair the A12 was called the A17 until shortly before CES when there was official word as to the differentiation between the two names. As you can see from the Rockchip RK3288 banner, some impromptu last minutes updates were needed. The A17 name is reserved for the first major update to the A12, a departure from previous ARM naming standards. Yes ARM is to blame for this stupidity but it is done and that is the law of the licensee land now.
Unless you are Mediatek. SemiAccurate has been lead to believe that they are purposefully confusing the two for a marketing advantage. How do we know that the MT6595 isn’t using the revised A12/Owl core that is now known as A17? Because the RTL for the A17 hadn’t been released by ARM to SoC designers as of MWC but was scheduled to be released before you read this. Mediatek did not have A17 designs much less silicon bearing it back from the fab, period. What they did have is the desire to stand out and claim something for marketing purposes regardless of what is actually in the device.
With that in mind, it takes the Mediatek MT6595 from the first demo’d A12 silicon, first demo’d SoC with H.265 encode and decode at 4K resolutions, and several more firsts too and turns the showing into something else. It should have been a triumph, what Rockchip announced first Mediatek released first. Instead their marketing did something akin to a car manufacturer claiming a model has a V8 when there are only four cylinders under the hood, it is at the bare minimum unethical. With that they took a triumph and destroyed any good will and accolades they would have gotten. What a shame.S|A
Have you signed up for our newsletter yet?
Did you know that you can access all our past subscription-only articles with a simple Student Membership for 100 USD per year? If you want in-depth analysis and exclusive exclusives, we don’t make the news, we just report it so there is no guarantee when exclusives are added to the Professional level but that’s where you’ll find the deep dive analysis.
Latest posts by Charlie Demerjian (see all)
- More on Intel’s 10nm process problems - Sep 17, 2018
- Intel puts out another 14nm 2020 server platform - Sep 11, 2018
- Why Can’t Intel Supply Enough 14nm Xeons? - Sep 10, 2018
- Intel can’t supply 14nm Xeons, HPE directly recommends AMD Epyc - Sep 7, 2018
- AMD reintroduces the Athlon name with two CPUs - Sep 6, 2018