• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Apple A10 Fusion is ** Quad-core big.LITTLE **

Page 14 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
krumme said:
Looking at the lukewarm reception of ip7 one have to wonder if apple shouldnt care more for the core product. But i guess they have another need for all that expensive cpu grunt comming up.
Click to expand...
Did you wormhole into bizarro world? Check Android profit ledgers and make sure...

iPhone 7 has been a huge hit everywhere not to mention us geeks who are focusing on the innovation inside
 
Arachnotronic said:
There is no information from Apple about how iPhone 7/7 Plus is doing yet.
Click to expand...
I've noticed that Forbes' tech "articles" these days are all basically blog posts. Maybe it's a move to get news and views out quicker, but the quality has really suffered. May as well read Gizmodo.

That said, I would expect that iPhone 7 / 7 Plus unit sales to trail 6s / 6s Plus sales, partially because of the loss of the headphone jack and partially because it is the third iteration of the same outward design, and most importantly because Apple this year suddenly decided not to release sales numbers. Apple claims the reason is because they couldn't accomodate demand, and I suspect that is true, especially for the new colours and the Plus. The Plus is the new king this year, whereas the prior two years the regular 6 / 6s were the bigger sellers. However, I also think the other reasons I've listed play into it.

But, I'm just guessing, and it seems Forbes' blog posts are guesses too. We'll find out more though in three weeks, at the earnings conference all. I do wonder though, even if the unit sales of the iPhone do fall, how much of that will be mitigated by the higher percentage of Plus sales. My GUESS is that overall iPhone revenue will still fall vs. last year, but it's not all doom and gloom like some want to believe, and things will change in 2017 with record sales again given that it will likely be a completely new form factor.
 
My wife just bought a new ip7 as an upgrade to her ip5. Verizon was pushing a large trade in value for the ip5, and an even larger amount for an '6' series iphone.
I have no idea if Apple is cutting a deal with Verizon or not, but this is a big push to increase sales (at least in the US). I'll probably wait for the ip8 (I want BT5.0 anyway).
 
Eug said:
That said, I would expect that iPhone 7 / 7 Plus unit sales to trail 6s / 6s Plus sales, partially because of the loss of the headphone jack and partially because it is the third iteration of the same outward design
Click to expand...
Anyone making predictions on sales numbers is just guessing. All we know is that even though Apple did not have to do any major retooling, the demand was so high that their was a backlog and there still seems to be a backlog on certain models. If you want to predict, just remember that iPhone 6/6Plus was 2 cycles ago and all those contracts have come up for renewal. Samsung has imploded and there does not exist a real competitor to iPhone in processor performance among any competing mobile phone.

But it's too early to tell and Apple guards its stats very well.
 
Antutu top 10 smartphones of September 2016

Antutu-Top-10-September-2016.png


In third there is the Le Pro 3, which uses Qualcomm Snapdragon 821.

Also, Google Pixel XL Geekbench 4 scores are out

Single core 1650, multi-core 4150

As a reminder, A10 Fusion gets about 3300 single core and 5500 multi-core.
 
So.. Apple destroys any ST opponent even Intel, but is brutalized in MT by everyone (even Mediatek and AMD), except Qualcomm? Interesting...
 
Nothingness said:
Cortex-A72 still does better with 1883/5420.
Click to expand...

LOL, Kryo is a joke. More like 64 bit Krait than anything else. Kirin 955 is 16nm A72+A53. At this rate, A73 based SoCs on 16/10nm next year are going to put the final nail in the coffin for Qualcomm's custom cores.

ARM is doing an awesome job on their reference cores lately.
 
.vodka said:
LOL, Kryo is a joke. More like 64 bit Krait than anything else. Kirin 955 is 16nm A72+A53. At this rate, A73 based SoCs on 16/10nm next year are going to put the final nail in the coffin for Qualcomm's custom cores.

ARM is doing an awesome job on their reference cores lately.
Click to expand...
Agreed, but even ARM is far behind Apple.
 
Yes, that's clear.

This puts it in perspective quite nicely:

Andrei. said:
What's so unbelievable about a core that's larger than anything out there? Assuming this estimate is anywhere near reality, then the an A10 core without any L2 is larger than a Zen core + L2. It's almost quadruple the size of a M1 core which is currently the fastest other ARM core, and that's on a less dense process. Remember that while it's faster its per-core power is also higher compared to other ARM cores.
Click to expand...

Higher power vs other ARM cores, yet it still fits in a phone. Amazing engineering, and insane to think such power can be held in your pocket. Hopefully the Android world will get to see a bigger fraction of this performance going forward with A73's successor in a few years, in the form of another off the shelf ARM core. Despite this, I think A72/A73 are already approaching the good enough area for what you can do with a phone, similar to Conroe back then...
 
.vodka said:
Higher power vs other ARM cores, yet it still fits in a phone.
Click to expand...
We had Snapdragon 810 devices routinely hitting current limits at 10W CPU (And that wasn't the theoretical unthrottled peak) power, the definition of what "fits in a phone" can be stretched to stupid levels.

Efficiency curves and dynamic range matters though. That's why introducing little cores in the A10 is probably part of why they were able to push clocks this much higher on the big ones.
 
antihelten said:
Kryo in Snapdragon 820 is also capable of hitting 1883, so I would be careful judging these numbers just yet.
Click to expand...
The only device which scores that high is the OP3 because it has a very aggressive kernel. But most devices with the SD820 and a more balanced kernel get around 1600-1700 ST.
 
Find me op3 kernel with 20% oc on big cluster. and actual oc btw, not just modified frequency table + other optimizations resulting 20% better score. you really have no idea about this have you ? gb4 made gb3 irrelevant, gb5 will make gb4 irrelevant. means gb4 is just as irrelevant as gb3. because it's just a freaking benchmark. just that.

or you know. just forget it. don't know why i'm still arguing when even apple agrees with me and still uses intel. wake me up when apple ditches intel. till then enjoy GB4 benchmarking.
 
kraatus77 said:
Find me op3 kernel with 20% oc on big cluster. and actual oc btw, not just modified frequency table + other optimizations resulting 20% better score.
Click to expand...
I don't know where you got that 20% from, but here's an OC kernel, one of the most popular ones:
http://forum.xda-developers.com/oneplus-3/development/kernel-elementalx-op3-0-01-t3404879

The OC commit:
https://github.com/flar2/android_ke...mmit/e441a0e96a75de0739e8ffa6e8a4ccfc20bbcd77

And the fun thing is that every GB4 result listed with a frequency of 1728 is guaranteed to be an custom kernel/frequency change: https://browser.primatelabs.com/v4/cpu/search?dir=desc&q=OnePlus+3&sort=score

The scores quickly fall back to median platform values (Such as MI5 https://browser.primatelabs.com/v4/cpu/search?dir=desc&page=1&q=xiaomi+mi+5&sort=score) a few pages in.
kraatus77 said:
you really have no idea about this have you ? gb4 made gb3 irrelevant, gb5 will make gb4 irrelevant.
Click to expand...
No, GB4 didn't make GB3 irrelevant. GB3 was irrelevant (at least the general score) by itself long before GB4. It's not my fault you don't understand benchmarks and that you keep using flawed arguments to justify your viewpoint.
 
Lodix said:
The only device which scores that high is the OP3 because it has a very aggressive kernel. But most devices with the SD820 and a more balanced kernel get around 1600-1700 ST.
Click to expand...

No OP3 is not the only device scoring in this range. Whilst the score I linked us admittedly the highest OP3 score I saw, the HTC 10 also scores in the 1800-1850 range. le max 2, Zenfone 3 deluxe and Mi 5s plus also hit this range (the latter two use the 821 just like the Pixel phones).
 
Lodix said:
They are overclocked.
Click to expand...

The le max 2 and Mi 5s plus results are not overclocked as far as I can tell, they do however seem to use a fairly aggressive governor since all of their results on geekbench are listed at the frequency of the faster cluster of cores.

As for the rest, here are non-overclocked versions, all scoring in the aforementioned 1800-1850 range (except the Zenfone, which only manages 1785):

Oneplus 3
HTC 10
Zenfone 3 deluxe

Also it's worth noting that the A72 score linked earlier is the highest A72 score on Geekbench, there are plenty of phones with A72 that don't go any higher than 1600-1700.
 
If Apple is really able to release an A10X at 2.8 GHz like some here think they will, that means it will probably have a Geekbench 4 single core score of 4000 and a multi-core score of 6500.

Those seem like nice round numbers, so let's go with those. I officially predict Apple will breach the 4000/6500 mark in early 2017.
 
BTW, someone asked if the A10 Fusion performance in a phone really is helpful. I mentioned a few metrics where the A10 is nice but in the end after several weeks usage I think if you're buying a 4.7" iPhone in 2016 the 6s is fine esp. if you're on a budget. (For the Plus though, the 7 Plus is highly recommended over the 6s Plus.)

However, one additional real world metric I've come across makes a difference though. Apple has supported HEVC h.265 in hardware ever since A8 and my iPad Air 2's A8X but doesn't expose this to developers. Consequently HEVC decode is all software, and unfortunately on my iPad Air 2, some 1080p HEVC content will stutter. This is better on A9 (wife's iPhone 6s), and even better on A10 Fusion (my iPhone 7 Plus). A10 has smooth playback but you can tell it's likely stressed by it since the video playback software's menu and control interface severely lags when playing HD HEVC content.

Mind you, this would be solved if Apple exposed hardware HEVC support to developers. Is this a licencing issue or something?
 
Last edited:
You must log in or register to reply here.

TRENDING THREADS

Back
Top