bit-tech.net

ARM announces mid-range Cortex-A17

ARM announces mid-range Cortex-A17

ARM's new Cortex-A17 boasts a 60 per cent performance improvement over its predecessor the Cortex-A9.

ARM has announced a new processor aimed at mid-range smartphones, tablets and embedded systems: the Cortex-A17, with the first devices expected to appear before the end of the year.

The ARM Cortex-A17 is designed as a replacement for the Cortex-A9, a common sight in low-end to mid-range devices. According to ARM's internal testing, the new design offers a 60 per cent boost in performance while also improving energy efficiency over its predecessor. It also includes support for ARM's big.LITTLE architecture, which pairs high-power primary cores with a suite of lower-power secondary cores designed to take over when raw grunt isn't required. An ARM Mali-T720 graphics processor is also specced as an ideal partner in ARM's system-on-chip (SoC) design, with the company plugging the part as a winner for set-top boxes as well as battery-operated hardware.

'We expect to see a rich set of innovation in the mid-range mobile phone segmentm which is forecast to become a half a billion unit market annually from 2015m and the Cortex-A17 processor will be a key component in that growth,' claimed Ian Ferguson, vice president of segment marketing, ARM. 'To date, the ARM Partnership has shipped more than 50 billion ARM-based chips and the continued broadening of our processor family will enable our partners to further optimise their offerings in existing and new product categories.'

The first company to confirm availability of parts based on the design is MediaTek, which has come out of the gate with a promise of an eight-core 4G-enabled smartphone SoC in the first half of the year. The MediaTek MT6595, the company claims, boasts four Cortex-A17 cores alongside four Cortex-A7 cores in big.LITTLE configuration, a PowerVR Series 6 graphics processor, an integrated Long Term Evolution (LTE) modem with 150Mb/s downstream and 50Mb/s upstream support, and H.265 hardware video acceleration for up to Ultra HD (4K2K) video streams.

MediaTek is to begin shipping the SoC in the first half of this year, and predicts the first devices featuring the chip to appear before Christmas. ARM itself, meanwhile, doesn't expect to see shipping products from other hardware partners until early 2015.

5 Comments

Discuss in the forums Reply
azazel1024 11th February 2014, 15:16 Quote
I am hoping the A17 is a bit more promising than the A15.

The A15 just doesn't seem to have the performance per watt ratio that works well, at least for phones. It might be higher performance than the A17 here, but it also is a fair bit thirstier on power, even in a big.little config, which also increases chip cost because of die area.

I am wondering if/when OEMs realize that maybe "octocore" designs aren't the way to go. I'd think, especially in a phone, the best setup would be 2 A17s and a single A7 if you need a big.little configuration. I can't imagine it would be that difficult for the OS to schedule it appropriately. Chip would cost a lot less than a 4+4 design and probably be roughly as powerful for a lot of workloads.
schmidtbag 11th February 2014, 15:26 Quote
So, is this basically just a more power efficient A15? It seems to be pretty similar to it in most ways, except for power consumption. Also, ARM is getting to be just as bad as Intel and AMD when it comes to naming products. Their naming schemes are much simpler but still don't make much sense.

On another note, I'm a little surprised its taking so long for the A9 to go away.
schmidtbag 11th February 2014, 15:31 Quote
Quote:
Originally Posted by azazel1024
I am hoping the A17 is a bit more promising than the A15.

The A15 just doesn't seem to have the performance per watt ratio that works well, at least for phones. It might be higher performance than the A17 here, but it also is a fair bit thirstier on power, even in a big.little config, which also increases chip cost because of die area.

I am wondering if/when OEMs realize that maybe "octocore" designs aren't the way to go. I'd think, especially in a phone, the best setup would be 2 A17s and a single A7 if you need a big.little configuration. I can't imagine it would be that difficult for the OS to schedule it appropriately. Chip would cost a lot less than a 4+4 design and probably be roughly as powerful for a lot of workloads.

I agree, the the octocores aren't ideal. I thought big.LITTLE was kind of a dumb idea - it's a pain for fabrication systems and software developers to work with; it was a sloppy solution with good market potential (8 cores in a phone or tablet).

But, what many people tend to forget is the A15, while not as power efficient as A9, still (generally) had better performance-per-watt compared to most Intel Atoms and similarly clocked Celerons.
rollo 11th February 2014, 15:38 Quote
A9 has stuck around for so long because its cheap and still more than enough for 90% of users.
GuilleAcoustic 11th February 2014, 19:09 Quote
As long as it has a good Linux support (GPU included) ...
Log in

You are not logged in, please login with your forum account below. If you don't already have an account please register to start contributing.



Discuss in the forums