Today, Qualcomm announced a number of details in the Snapdragon 820, specifically about their Kryo CPU. Given that the Snapdragon 810 was a somewhat standard 4x Cortex-A57/4x Cortex-A53, it was clear that that this chip was a stop-gap for a future fully custom design. With the Snapdragon 820 announcement, the first major bit of information that we received was that this would be a return to a custom CPU core design, and today Qualcomm is finally unveiling a bit more information on Kryo.

The two main spec details that are being disclosed today is that the quad-core Kryo CPU in Snapdragon 820 will reach up to 2.2 GHz, and that the SoC will be manufactured on Samsung’s 14nm FinFET process. It isn’t stated whether this is the 14LPP process, which will give up to 10% transistor performance improvement over 14LPE which was seen in chips like the Exynos 7420, but it’s a safe bet that it is. As a result of the new architecture and new process node, Qualcomm is claiming up to a 2x increase in performance and up to a 2x increase in power efficiency compared to Snapdragon 810.

The final part of this announcement is Symphony System Manager, which is said to be designed to deal with heterogeneous compute in an efficient manner. This is likely to be a kernel-level mechanism that ensures that the SoC is well-optimized for use in a smartphone or any other application. Given the focus on heterogeneous compute for this launch, I wonder if Qualcomm is going for some form of heterogeneous CPU design as well.

Source: Qualcomm

Comments Locked

60 Comments

View All Comments

  • jjj - Wednesday, September 2, 2015 - link

    Mediatek is going 16ff in the first half of 2016 in the high end (and in the second half in the mainstream) so X20 will likely have the lead for a few months ahead of SD820 and then we'll see when x30 or w/e arrives and if it can compete with SD820. We'll see if X20 can hit it's 2.5GHz clocks, would be great if it does but i guess it would be ok at 2-2.2GHz too before their 16ff product arrives. Huawei doesn't seem quite ready to compete so i wouldn't hope for too much. Samsung and LG should have their own SoCs too next year and maybe Spreadtrum goes for the high end on 14nm Intel.
  • Buk Lau - Friday, September 4, 2015 - link

    I mean the Kirin 950 with quad A72 with T880 seems promising, but much like Samsung they are for in-house products. I think the biggest game changers next year will be MediaTek and Intel. For now it's even a 50/50 for whether Samsung will continue to their Exynos or not for S7
  • jjj - Friday, September 4, 2015 - link

    I was aware what you mean but so far Hisilicon hasn't been so great. On the memory side, on the GPU side so i wouldn't expect them to be better competitive even if they use the same core and GPU as others. At some point they will ,maybe even as soon as now but based on their record i'm reluctant to expect too much from them just yet.
    Oh and Hisilicon is actually trying to sell to others. As for Samsung,they'll always go with the best solution no matter if it's internal or not.
    A game changer wouldn't be in the high end. A dual A72+ some A53s (instead of 8xA53) at 20$ would be a game changer. Would enable Xiaomi, Meizu and the likes to make really fast 150$ phones.
  • tabascosauz - Wednesday, September 2, 2015 - link

    HAHAHAHAHAHAHAHA

    "Nv getting into high end phones"

    Nice try, buddy. When phone manufacturers have endured Tegra 2, 3, 4, and 4i, you can bet your a** that they're not going to try anything with Nvidia again.
  • lilmoe - Wednesday, September 2, 2015 - link

    It would be hard for them. NVidia has lost its hype edge. Marketing a device as having a Tegra inside doesn't have a big impact anymore. There's also the issue of integrated modems that NVidia dropped.

    It's going to be VERY hard for them to compete with QC and Samsung in the high end without a complete package. It's also very hard to compete with Mediatech and other Chinese chips in price for the mid to low end...

    I believe they should be focusing more on tablets. Microsoft, in particular, should strike a good deal with them to build a chip with a huge, relatively efficient, GPU to be the minimum requirement for Windows 10 Mobile (tablets and phones). They have the edge in Windows for GPU drivers, which would then be a standard for Games to evolve on the platform with DX12 and most games would target that performance level for optimization. But it also seems that Microsoft is going the Qualcomm route.... Oh well.
  • tuxRoller - Wednesday, September 2, 2015 - link

    Nvidia had quit the consumer market in mobile, aiui.
  • mforce - Thursday, September 3, 2015 - link

    Well Nvidia gave up on their modem division and on the whole smartphones idea because they said they couldn't make enough profit from it. In reality they didn't seem to manage to make a powerful SOC so of course they couldn't charge as much as they wanted for their average ( at best ) ones ...
    Mediatek is a force to be reckoned with and the only real competition for Qualcomm right now. Intel is trying something but they're always late with their SOCs so by the time theirs are out they can't really compete with the top ARM SOCs.
  • 0razor1 - Thursday, September 3, 2015 - link

    They simply used old nodes for decent cores. Denver, had it two A53 cores ( they were out then), would have been great, They actually beat Apple's custom cores at the point they came out. Their SW issues have plagued them and general wider adoption.

    When you have a reference device, you make it perfect. The nexus 9 could have, if executed well, pushed the Denver cores into a lot of chromebooks ( at the very least) and other android tabs.
  • 0razor1 - Thursday, September 3, 2015 - link

    I meant two a53+ their custom cores. Where's the EDIT button ?? :D
  • tuxRoller - Thursday, September 3, 2015 - link

    Mediatek isn't too interesting, IMHO. They don't balance cpu/gpu well. They don't create their own designs. They seem to be pursuing hype over user experience with the use of such high clocked little cores (way out of their max efficiency range). Their, unreleased (illegal given the kernel's license) kernel changes don't seem to be very good.
    LG might be interesting.

Log in

Don't have an account? Sign up now