POST A COMMENT

37 Comments

Back to Article

  • [-Stash-] - Tuesday, May 06, 2014 - link

    I want a revamped Nokia 808 with an 808 in it :) Reply
  • Taneli - Tuesday, May 06, 2014 - link

    What's wrong with arm11? Reply
  • [-Stash-] - Tuesday, May 06, 2014 - link

    Then I wouldn't be able to get the trifecta and video a Roland TR-808 with it… Reply
  • SydneyBlue120d - Tuesday, May 06, 2014 - link

    I'd like to see a performance (and watt/performance) comparison between Qualcomm Krait 450 CPU Architecture in the 805 and the "Vanilla" A53 CPU Architecture in Snapdragon 610/615... I think clock difference and better architecture will be enough to keep the lead in Krait realm... or not? Reply
  • Krysto - Tuesday, May 06, 2014 - link

    Cortex A53 is most definitely better perft/Watt wise than Krait. Whether Krait will be better than Cortex A57 in perf/Watt that remains to be seen, but it will definitely be weaker in performance. Qualcomm has milked Krait for far too long. Not to mention you get a lot of other more specific performance advantages from the ARMv8 architecture. Qualcomm has nothing to show but repacked chips with repackaged names this year, and they won't for a while (mid-2015). Reply
  • beginner99 - Tuesday, May 06, 2014 - link

    Yeah but if only the core/CPU mattered and not the whole SOC (especially included LTE modem) Intel should have the lead and we all know that isn't the case. Reply
  • mczak - Tuesday, May 06, 2014 - link

    I wouldn't quite say they milked Krait for far too long but yes the 450 is one iteration too many (Qualcomm missed the armv8 boat badly). I suspect running 32bit code it would still have better perf/w than A57 but it is indeed possible the A57 gets competitive there (something the A15 was not) due to be able to run armv8 code. Reply
  • Morawka - Tuesday, May 06, 2014 - link

    "(Qualcomm missed the armv8 boat badly)"

    haha man they did so bad. its gonna cost them the company i think.
    Reply
  • Morawka - Tuesday, May 06, 2014 - link

    if not the company then a lot of frackin money Reply
  • drexnx - Tuesday, May 06, 2014 - link

    or none at all, because they're an ARM licensee. If Krait isn't competitive vs. vanilla ARM for the next year, QC can roll vanilla ARM products until Krait's successor comes on stream.

    QC is in EVERY flagship android phone right now - those kind of OEM relationships don't go away because their core isn't the best of the best for 12 months. I.E. when scorpion (snap S3) was blatantly inferior to A9, HTC, samsung, etc. all still rolled Snapdragon S3 dualcores anyway.
    Reply
  • michael2k - Tuesday, May 06, 2014 - link

    QC is in every flagship android phone right now because no one else has a workable replacement (ie, everyone is in the same boat) so they might as well leverage QC's radio integration. That however isn't a static relationship as both Intel and NVIDIA have their own integrated radios and their own higher performance SoC as well as competitive GPUs coming out this year. Reply
  • bigstrudel - Tuesday, May 06, 2014 - link

    They're using stock designs to make up for it. I'd say they absolutely missed the boat that Apple launched and left without them. Reply
  • drexnx - Tuesday, May 06, 2014 - link

    and yet they're still capturing design wins to boot, so no skin off their back. They're still selling chips.

    They can't expect to be ahead for 2+ years on one uarch anyway.

    look at it this way:
    Scorpion was better than A8 for about 18 months, then inferior to full A9 (A9 with NEON. Tegra 2 doesn't count since it traded blows evenly with higher clocked snap S2s, and until the OMAP4430 and Exynos 4210 came out, the snap S3 then killed the Tegra 2) for a year before Krait came along. Krait is better than A9 and A15, and has been since launch in 1-2012... so A57 will beat Krait for 9-12 months... big deal, Krait's successor will be out soon enough anyway.
    basically, QC's custom core designs have been better than vanilla ARM for something like 66-75% of the time since scorpion came out way back in 2008 - that's pretty damn good.

    Seriously, are there any viable flagship non-proprietary SoC competitors left anyway? TI is gone, Nvidia is a joke, ST-Ericsson dissolved
    Reply
  • drexnx - Tuesday, May 06, 2014 - link

    I guess that's why QC has 64% marketshare in SoCs and the next closest competitor has 12%.

    they're really doomed.
    Reply
  • tuxRoller - Wednesday, May 07, 2014 - link

    Clock for clock, krait is pretty inferior to a15. Reply
  • CharlieSD11 - Wednesday, May 07, 2014 - link

    How is it missing the boat when there isn't a 64-bit OS released yet to pair with a 64-bit CPU? If Android or Windows Phone was 64-bit today then I would think you could say they missed the boat and should have seen this coming. Reply
  • bigstrudel - Wednesday, May 07, 2014 - link

    By miss the boat I mean MediaTek beat them to 64bit. Reply
  • tuxRoller - Wednesday, May 07, 2014 - link

    Do we have any idea about how different A7 cores are from a57? Reply
  • bigstrudel - Tuesday, May 06, 2014 - link

    Krait is ancient. It's hardly changed from its first iteration in the Snapdragon S4 Dual SoC. Just more cores and clock speed. Better software and browsers also bloat gains between Krait generations.

    Any SoC using Armv8 instructions will be more efficient than Krait.
    Reply
  • da_asmodai - Tuesday, May 06, 2014 - link

    I'm curious how the "LITTLE" from the 64bit generation (A53) compares to the "big" from the current 32bit one (A15) on the same process technology node and with 32bit workloads. Are the architecture improvements enough to close that gap significantly? Reply
  • Wilco1 - Tuesday, May 06, 2014 - link

    According to the Geekbench graph above, A53 in 64-bit mode has 75% of the IPC of A15. Of course it runs at lower frequencies, so you'd probably get about half of the performance of today's high-end A15. Reply
  • extide - Tuesday, May 06, 2014 - link

    Look at the pictures in the article, this exact scenario is displayed. Reply
  • darkich - Tuesday, May 06, 2014 - link

    As far as I know, A57 can be clocked over 3GHz.
    So chips with, say, quad 3GHz A57 CPU and Maxwell Tegra GPU configuration could be one helluva prospect for tablets and serious chromebooks
    Reply
  • bigstrudel - Tuesday, May 06, 2014 - link

    Stock Arm cores for all new models for Qualcomm? Lazy. Custom cores are much more interesting. Reply
  • TheJian - Tuesday, May 06, 2014 - link

    Nvidia has already announced IN HOUSE Denver cores which 810 is NOT. I don't understand your Qcom qualifier here as if it's special or in a great position (their move to 64bit is a mistake they have to fix, not a well executed plan). It's the exact opposite as they missed the 64bit boat and had to choose off the shelf IP for the cpu's until 6-12 months later than usual when they can get their own cores out he door. I can't see off the shelf IP beating NV's custom cores in battery and we already know the gpu is top notch with kepler which will further add to the quality of their games (tegrazone has a huge lead in optimized games that look better than other hardware). In a tablet it will be tough to argue for Qcom this year at xmas and a good portion of next year probably too. If google really is re-tooling their 10in tablet as a gaming tablet with T4/K1 NV will start to catch on at just the right time for gpus to take over modems in the food chain. Now that they i500 certified NV's follow ups to T4i should be impressive on all fronts. Modems are becoming moot points (2GB caps make faster speeds pointless, that isn't even a single 720p movie) and Qcom isn't in house like krait/swift now for their next chips so their battery life won't look like krait/swift as before. There is a reason Krait/Swift have great battery/perf combo.

    Qcom's two main advantages are gone with 810 and will take another gen to catch back up. A software modem may actually end up being better than hardware based also, but we'll have to wait and see how that plays out. More importantly NV's gpu's will be hard to beat from here out as every desktop chip gets plopped into a soc a year later with all game devs knowing full well how to use them inside & out. What qcom did to the mobile industry with modems NV is about to do to it with gpus. Soul Craft, Heroes Call, Blood Sword: Sword of Ruin, Raven Sword, Bounty Arms, King's Bounty Legions, etc etc show you where mobile is heading (straight into console's worlds, they are way deeper/better than angry birds these days). They already are moving over games like Mount & Blade Warband (a 2010 PC game), GTA San Andreas (the whole 70hr game) etc showing many won't have a need for console this year or next as their mobile device is capable of playing a few hundred gamepad games already (330+ as of april, surely will head on up with shield rev2 coming next month or july). Cut down or not effects wise these will be drawing lots of users and the experience is best on NV hands down and I can't wait to see what K1 games bring. It doesn't have to be new content to win gamers who have never seen the old stuff on xbox360/ps3/pc.

    The coming crop of games will make an xbox360/ps3 not necessary for most and the die shrink after 20nm will make consoles moot for a majority of the rest especially the ones with pc's already. I don't see all these lost console gamers moving to qcom mobile chips without some major work on Qcom's part (where are qcom optimized games?). You can already see the quality starting to come from Tegrazone games. Next gen sales will be ruled by the best gpu's/games. With 70% of our time on mobile being spent on gaming it's easy to see why devs massively went mobile (sheer units force you to pay attention, 1.2B per year already) and why the GPU will rule vs. modem going forward.

    People see Tegra as a failure but don't understand the first 4 were just about making noise and getting noticed until the real war began with K1/desktop gpus and a modem that is good enough to get into phones (leveling the playing field, time to compete in NV's wheelhouse: GAMING). AMD needs to hurry up, as there is plenty of room for both NV/AMD in eating Qcom's $6B+ profit pie for a while, not to mention the share mobile will steal from notebooks/desktops soon in x86 world (already getting 21% of notebooks, so Intel pie can be had to some extent also as shown by chromebooks). AMD makes nearly nothing and NV is only pocketing ~$600mil/year right now. Merely getting 10% of Qcom's market gets tegra to double NV's income (never mind how much this would help AMD). It only takes a few good design wins to garner a good chunk of sales. NV can bleed Intel's lawsuit payments until this happens (which is what they've been doing, Intel's payments cover tegra losses/R&D) as their top gpus move into socs. I don't see anyone else even on the gaming map yet, even apple doesn't seem to get it so far. If gaming isn't your #1 priority on mobile you're going to lose the next war as gaming moves from consoles to mobile and the rest back to PC (as GDC 2013/2014 shows). PC is now #1 with devs with mobile just a decimal behind (console FAR behind both)! The 808/810 are nothing special. I have no need for faster modems these days (well at home maybe, but that is unlimited, pointless on phones however) but I can always use better gaming. NV's software modem might be upgradable to 300 anyway (already did once to 150). You've already said the modem IP isn't the winner in that 808/810 article already so not sure how you think they win a gpu game with NV way out in front no matter how you slice it. Qcom is now behind with no custom A57's until 2h2015. That means someone else will be the top choice for at least tablets and maybe phones if NV can get a modem into a K1 chip soon enough to take real advantage of Qcom's 64bit mess up. The tablets should start to be a no-brainer soon.

    I hope you guys are planning on covering NV's quarterly report like you did AMD's (still waiting for the NV portal to show up...LOL). Leaks show NV had a great quarter yet again (166mil income) and this is still with them eating tegra losses that should stop as soon as K1 kicks in. Net profit nearly double last year's quarter and margins up again.
    Reply
  • wintermute000 - Wednesday, May 07, 2014 - link

    too bad all the T2/T3 misses burnt all their bridges with all the OEMs. None of the big players trust Nvidia now. Thats a bigger problem than anything technical. Reply
  • darkich - Wednesday, May 07, 2014 - link

    Awesome post.

    A bit away from the point about gaming imo - mobile games aren't going to be replaced with pc ports, and we are not going to start using smartphones and tablets as consoles. The fact is mobile games are a new, strongly established entity and will only continue to grow and develop in the taken direction instead of trying to copy other platforms - but again, awesome post.
    Reply
  • phoenix_rizzen - Wednesday, May 07, 2014 - link

    We'll see.

    Fool me once (T1), shame on me.
    Fool me twice (T2), shame on you.
    Fool me thrice (T3), why are we still talking?
    Fool me a fourth time (T4), is anyone even paying attention?
    Fool me a fifth time (K1) ... oh, wait, we all gave up a long time ago; there won't be a fifth time.

    nVidia may have the greatest SoC since sliced bread, but who is going to go to bat for them a fifth time?
    Reply
  • tuxRoller - Wednesday, May 07, 2014 - link

    I guess we all have our dreams. Reply
  • sonicmerlin - Wednesday, May 07, 2014 - link

    So how does the A57 compare to a Core 2 Duo? Reply
  • Wilco1 - Wednesday, May 07, 2014 - link

    More interestingly, how does it compare to a low clocked Haswell, say the ones used in Chromebooks? Benchmarks showed those were ~2x faster than the original Samsung Chromebook (1.7GHz dual A15). The slides above say A57 will be twice as fast as the latest 28nm A15s, so should be able to match or beat those. And that is very interesting as it shows ARM is getting the level of performance of a typical cheap laptop/PC. Reply
  • darkich - Wednesday, May 07, 2014 - link

    ..even if it doesn't match a low power Haswell in performance, it will absolutely trump it in perf./watt.

    But I am more interested in seeing the A8 from Apple.
    That one should beat the most frugal Haswell variants even performance wise
    Reply
  • errorr - Wednesday, May 07, 2014 - link

    Qualcomm will be fine even if they just have vanilla cores and the customer comes later. Premier tablet chips are a tiny niche market and the money is made in smartphones. They might lose some market share outside the US but there won't be a premium US phone that moves away from QC because of their modem and IP integration. Nobody can beat QC in price/performance in the US because of CDMA and 2g legacy IP which it luckily doesn't have to pay itself for.

    If VoLTE was fully deployed then maybe an LTE only flagship may debut but that is at least a year from being viable on Verizon and even longer on other networks. QC has to be the choice for US markets and unless you are Samsung it isn't worth building 2 phones to market to non US customers. Also QC pays the freight for network tests and certification required by the big US telcos which is just another reason QC dominates on price.

    We won't have tried SOC competition until the bog networks start shutting down their 2g bands and migrate fully to LTE. VoLTE will be the real game changer as the licensing costs for the IP is cheaper and easier to implement with a lot of competing modem techs available.
    Reply
  • CiccioB - Wednesday, May 07, 2014 - link

    1. High level tablets use SoCs that have more profit margins than those used in phones
    2. Phones SoC are very low cost and very low margins, and Qualcomm position on low market (the one with bigger selling numbers) has already been taken by Chinese and Tawainese SoC producers. If SoC producers like Chinese and Taiwanese ones can offer a SoC + separate modem at a lower price then Qualcomm integrated SoC then Qualcomm advantage is gone.
    2. Modem integration is still what gives advantages to Qualcomm in mid/top market. But take into account that USA is not the only market on the planet, as Europe and Asia greatly out numbers it by selling numbers and Qualcomm USA only modems are not required there.
    Reply
  • darkich - Wednesday, May 07, 2014 - link

    "High level tablets use SoCs that have more profit margins than those used in phones"

    On what planet??

    Ever since the very introduction of low voltage ARM soc's, the highest end ones are going to Smartphones first.
    It is by far the most dominant platform and tablets don't come even close
    Reply
  • CiccioB - Wednesday, May 07, 2014 - link

    On what planet what?

    In tablets SoCs can be bigger or of less "quality" than those used for phones, as in tablets power contraints are a less critical. If you speak for numbers, yes, tablets have a smaller market, but it is not sure that they have such a smaller profit.
    Reply
  • darkich - Wednesday, May 07, 2014 - link

    Your quoted statement is wrong unless you are referring to Haswell chips.

    I repeat - high end phones get the best and most powerful ARM chips.
    Reply

Log in

Don't have an account? Sign up now