Apple comments on A9 chipset performance: Negligible real world difference

There's been a lot of chatter about potential performance differences between the Apple A9 processors in the iPhone 6s and iPhone 6s Plus supplied by Taiwan Semiconductor Manufacturing Company (TSMC) and Samsung. Synthetic benchmarks run in unscientific tests have shown some variance in battery life between the two when under artificially heavy load. Under normal usage conditions, however, very little variance has been found. When reached for comment, Apple provided me with the following statement:

"With the Apple-designed A9 chip in your iPhone 6s or iPhone 6s Plus," an Apple spokesperson told iMore, "you are getting the most advanced smartphone chip in the world. Every chip we ship meets Apple's highest standards for providing incredible performance and deliver great battery life, regardless of iPhone 6s capacity, color, or model."Certain manufactured lab tests which run the processors with a continuous heavy workload until the battery depletes are not representative of real-world usage, since they spend an unrealistic amount of time at the highest CPU performance state. It's a misleading way to measure real-world battery life. Our testing and customer data show the actual battery life of the iPhone 6s and iPhone 6s Plus, even taking into account variable component differences, vary within just 2-3% of each other."

While Apple has been designing their own custom chipsets for years, the company doesn't own any fabrication plants (fabs) and so manufacturing is handled traditionally by Samsung and more recently by TSMC. This year, Apple is sourcing from both.

Since different fabs use different processes, yet Apple requires uniform results, all the chips are tested in just the way you'd imagine they would be—hard. That includes internal tests, popular public tests, and more importantly, data from actual devices being used by real people under real-life conditions.

If something fails to meet Apple's standards, it doesn't ship. Sometimes that causes shortages or minor delays, like with some models of Apple Watch, and sometimes it causes longer delays—the white iPhone 4 being the most famous example. And that was for the color.

With something as important as the processor in the iPhone, which is core to Apple's business, it's better for everyone involved to play it extremely safe. Apple's not interested in just selling you the current iPhone. They're interested in selling you iPhones now and far into the future. The brand is more important than the product.

If one supplier or another sometimes manages to exceed requirements in one dimension under one specific set of circumstances, that's great. Not good vs. bad—literally great vs. slightly greater under those specific conditions.

And again, real-world tests—including some of the ones posted—show only a 2-3% difference. Radio reception can create far greater variances than that, as can temperature, background task load, and more.

To be clear, if you're a tech enthusiast who is super into extreme workloads and synthetic benchmarks, it's an interesting subject to knock around. If you're a mainstream consumer, ignore the noise. There's no meaningful difference between iPhones and your time is much better spent enjoying your new phone.

Rene Ritchie

Rene Ritchie is one of the most respected Apple analysts in the business, reaching a combined audience of over 40 million readers a month. His YouTube channel, Vector, has over 90 thousand subscribers and 14 million views and his podcasts, including Debug, have been downloaded over 20 million times. He also regularly co-hosts MacBreak Weekly for the TWiT network and co-hosted CES Live! and Talk Mobile. Based in Montreal, Rene is a former director of product marketing, web developer, and graphic designer. He's authored several books and appeared on numerous television and radio segments to discuss Apple and the technology industry. When not working, he likes to cook, grapple, and spend time with his friends and family.

  • Typical canned manufacturer response. But I do agree that benchmark tests are fairly meaningless. Those who are getting worse battery life with their 6s need to wait a few weeks to start worrying about battery life. Most people who get a new phone use it alot more than the phone they just replaced. As for myself, I am getting phenomenal battery life out of my 6s plus. Even better than my 6 plus got.
  • Maybe, but the trouble tyou have is that benchmarks can be consistent. ‘Real world tests’ are not and it leaves a lot open to the manufacturer to skew. Notice that the Apple statement did NOT say that there is only a 2-3% difference between chips, only phones, (in THEIR testing).
    Also there are certain people that will run their phones at continuous heavy load, I used to and I was just gaming. What Apple need to do is get together with the other industry players and devise a suite of tests that are lab based, are consistent and do go some way to mimic real world usage. It’s not hard for any of them if they wanted to.
  • So according to the independent tests, which chip performed better, on average? We had a similar issue with the iPhone 6 flash storage last year - different supplier, different random read/writ times, if I recall correctly. Then there's this;
  • You know what. I thought it was my headphones or I. Exactly the same problem exists on my Mac.
  • So much about battery life is subjective that people will never be happy with any response from Apple. Illusion plays the biggest role here. The top comment on 9to5Mac's version of this story is some dude who is already extremely angry about it because his wife's iPhone battery "lasted much longer than his" which "proves" that there is a serious issue. Reason has already left the room. One only has to think of all the "bendgate" nonsense which happened despite there not even being an issue of iPhones bending during normal use, to realise that this chip thing (which is a slight but actual, real difference), is going to cause a shitstorm that won't end until the iPhone 7 comes out. It does make me wonder why Apple even went this route though. They should have seen the horrible mess that using two different chips (on two different processes) in the phone would cause. They must have known that chipworks would tear it down as they always do.
  • It's made worse by Apple's history of saying something isn't a problem only to admit later on there is an issue. You end up losing some credibility after a couple of times.
  • It's hard to complain about Bendgate when Apple fan sites were trying so hard to make Penghazi become a Twitter Trend. Apple can fix this issue by putting higher capacity batteries in their devices. Instead, they go the complete opposite way to save money and practically waste the advances in power efficiency of the upgraded components to maintain the status quo. Instead of the user upgrading and getting a battery life upgrade, they're upgrading and maintaining the status quo (unless they went from 5S to 6 Plus or 6 to 6 Plus, etc.). I don't think the variances are that huge. I think being in bad reception and things like that have a far bigger impact on your battery than most other factors. I know I can go to a training facility where I can barely get a signal and my battery will dip from 100% to 85% in an hour of standby. At home I can talk on the phone for an hour and it's still at 98-99% when I hang up.
  • I'm not sure I'd consider 12-18 minutes of battery life "negligible", but that is just a nit pick. Also odd that Apple pulled all the apps that identify the chip..
  • Apple pulling the apps is not odd at all given that this can only lead people to continue to obsess over a non issue. This is a mental insanity that happens every time a new phone comes out. There's a sudden peak in hysteria as evidenced in actual Google searches on "battery life issues" immediately after every launch of a new iPhone, and then those searches taper down and fall way down to an insignificant level. Each and every time this happens because when a new phone comes out, people use it much more than normal and suddenly their battery dies sooner, wow, what a correlation lol. So Apple pulling the apps is just a way to try to control this hysteria that is baseless and will go away in the next couple weeks. You'll get a bunch of idiots running to Apple demanding to exchange their perfectly good phone because an app told them they had the Samsung chip, when in reality, there *might* be a 15 minute difference in total overall battery life, which is completely negligible given your day to day life will vary more than that anyway.
  • If you read the statement carefully, it seems clear that Apple is saying that the 2-3% difference is between iPhone 6s and iPhone 6s plus, not between the two versions of the chip. Either that or they are saying that there is only a 2-3% difference between *any* two phones of the current generation tested, which is the same as saying the margin of error on the testing is 2-3%. No where do they actually say that there are two "versions" of each phone and that the difference is between one of these versions and the other.
  • The plus has several hours more battery life compared to the regular 6S. Check on their website. Apple is referring to the two chips in their statement. Sent from the iMore App
  • It seems that you're right and that there is an 8% or so difference between the two sizes. I think the only only reasonable interpretation then is my (original) *second* option, which is that Apple is saying that the overall difference between any two models (of the same size) varies by "up to 3%," and ignore the tense mistake (life versus lives), which implies that the comparison is between the two types instead. This still doesn't actually, (necessarily) address the chip issue though. Still, given that, if the iPhone 6 plus lasts one hour longer (8%) and the possible difference between two phones of the same type is a max of 3%, means that the total difference in battery life between differently chipped phones would be roughly 20 minutes? 25?
  • There's more than an 8% difference between the two sizes in real world usage.
  • But Apple didn't pull all the apps that identify the chips.
  • i have the samsung chip in my 6s and I'm not happy. i was going to stick with my iPhone 6 because the6s had a smaller battery but i fell for the power efficient spin. very annoying knowing i forked out a lot of money for an unlocked phone and now there is a better version out there. these new lumias are becoming more tempting by the day.
  • Yes. Lumia has always be the hidden gem. I used one 2 years ago. And I still find the Lumia is very much under rated. Posted from my Samsung Galaxy S6
  • Wait how is this insignificant? This means that if you have the Samsung chip, you can't do as much of anything that requires performance: recording 4K video, playing games, heavy internet surfing.. And the difference in these high load examples is not 2-3%, it is potentially up to 25%. And that is huge. If this is something we should not worry about, then we should also not worry about very loud fans in notebooks that turn on at high loads, or overheating phones/laptops, or speakers that sound distorted only when playing the loudest notes..
  • All it means is that if you are a super-picky wiener who thinks they need "performance," then you should go out of your way to buy a phone with a TSMC chip in it. Probably on the grey market though, since you won't get a choice of chip from Apple. You will have to pay a premium as well since those selling them will have seen you coming a mile away. Otherwise, you should probably just chill about the whole thing. Seriously, the majority of what's happening here is in your head. I know I feel (irrationally) better simply because I have the TMSC version, but the reality is that I'm not actually better off and you are not actually worse off. It's a game you're playing with yourself in your head.
  • Oh right, I should not be picky about a device that costs $1100 CAD and just be happy because Apple said it's all good.
  • You know the author. He always play down the impact on Apple's product. But a similar issue happen on any other competitor's product it's like the epic fail of all time. Posted from my Samsung Galaxy S6
  • Well we can all rest that the iPhone 7 will only have tsmc chips at least by early reports. Guess waiting for the design won't be so bad after all. Sent from the iMore App
  • "You're holding it wrong."
  • I suppose there has to be some kind of "gate" with every model or people simply wouldn't be satisfied.
  • Bill Gates Posted from my Samsung Galaxy S6
  • No, "you're testing it wrong".
  • I have the iPhone 6S with 128gb of int-storage, my brother has the 6S with 64gb of int-storage. We are running different chips, I'm running the so called "bad chip" (N71AP). My "bad chip" is getting better battery than his so called good chip, (N71MAP). After looking how he had his phone set up I changed a few things and now we are both getting just about same battery life. I changed all the settings on his phone to match as close to mine as possible. It's amazing how changing a few things around can make a huge difference. Just thought I would throw that out. My conclusion is I'm digging the hell out of my iPhone with the so called "bad chip". My 6S is easily taking me threw a full day.
  • What settings would they be?
  • Just checked. My wife's iPhone 6s Plus has the Map chip and my 6s is stuck with Samsung. Even if the reports are elevated I'm still getting 2-3% less battery life than I would get with the Taiwanese chipset. First world problem, but I didn't pay 2-3% less for my phone. Why do I get less performance? I imagine if I head down to my local Apple Store, they won't do much about it. I could care less about bend gate, but this is something even Apple admits to. Seems like I should be able to switch my phone. Sent from the iMore App
  • Apple is making some untrue statements trying to undermine the Geekbench test, and of course iMore (a caricature of an Apple fan website if there ever was one) parrots them without hesitation. Funny considering this site loves to post benchmark tests when they favour the iPhone against the competition. It does not "spend an unrealistic amount of time at the highest CPU performance state" According to Anandtech, a site that actually understands the tech in these phones, the test is a mix of high and low loads with periods of idle in each interval, translating to about a 30% usage level. At this level it in facts isolates the chip as the main factor in power consumption. Apple's idea of "normal usage conditions" are basically light browsing and iMessage, etc; that is, usage that stresses the display more than the cpu and therefore probably why they had little variance in these cases. And sure, this might mirror many peoples typical use of the phone, but I'm sure there are many who also love their phones for gaming, video recording, movies, etc. And they might be getting the short end of the stick if they get the Samsung chip. I'm willing to bet Apple won't ever try to dual source their CPU again if they can help it...
  • Yeah Apple is trying to downplay it because they were thinking the average consumer won't notice or care.
  • In my Twitter discussion with John Poole (Geekbench guy), he said that the GB battery test sets the screen brightness to minimum, and runs a constant workload loop from 100% to shut off. The proper way to run the GB battery test is literally to let it run for 4 to 8 hours (depending on phone). Tough benchmark to run. Running the test this way magnifies the differences in power consumption of the SoC, and ignores the power consumption from the screen and radios, which are 2 of the 3 biggest users of power in a phone.. Runtime ~= batter capacity in WHr / (power consumption of screen + SoC + radios + other components in Watts) For most of what you do, the brightness of the display is the biggest driver in run time life. If you use your phone "normally", the runtime of your device is gated by the brightness of your display and the signal level of your radios. The SoC spends 90% of the time in an idle state. If you use your 6S/6S Plus at minimum screen brightness, you may notice a difference if your usage is similar to the GB battery test's workload (doubtful it is). Once you set the screen brightness to medium or higher, more and more power is used by the display and the differences between the TSMC & Samsung SoCs will become quite small. If you are in a 0 to 2 cellular signal dot location most of the time, unfortunately, this is where you'll see <10 hr runtimes. Low signal conditions combined with being outside and using GPS or some SoC taxing app, tough tough conditions. 2% to 3% variation in run time performance is quite small and has to be referring to the SoC power consumption in Apple's "normal" workloads. For the whole phone, there will production variation for all components from each supplier and variation in components for multi-sourced components. Nearly every high power consumption component in the 6S/6S Plus is multi-sourced except for the radio. When accounting for all this: the display, the SoC, the radio chipsets, your environment and your workload will drive a ±50% variation in runtime performance let alone 2% or 3%. When Apple added background downloading in iOS 6 (?), Instacast implemented them, but it was buggy initially. This resulted in my phone getting hot and the battery meter going down a percentage point every minute or so. Yeah, that's when I deleted Instacast and switched to another podcast app. Run-time performance is multi-variate and these differences between TSMC and Samsung A9 SoCs is just not that big a deal for 99.93% of users out there. I would be highly skeptical of anyone who says that can pin down the difference in run time performance to the SoCs. It still really is the screen brightness, the lighting conditions you are most in, the signal conditions you are most in, and this or that backgrounding apps that pull data in bad signal conditions.
  • I nominate this comment to replace/amend the article.
  • I think the displays are fine. The Auto Brightness Algorithms err on the side of caution, are super fast and respond very very quickly to changes in light. I never felt like my display was a big factor in my battery life. I feel like the Cellular Radios are the achilles heel of these phones. IN great reception I feel like I have absolutely amazing battery life and unlimited standby times on my 6 Plus. Once I go into a building that blocks signals or into a market with crowded LTE Networks, my phone will drain 8-10% an hour without the screen off, in standby. Honestly, it was this way on Android Phones as well. I have a SIM-less M8 that I keep all the radios on, but it doesn't connect to a cellular network. It has none of those issues in the same building, so the issue is certainly not the screen or the SoC. The Cellular Connection seems to be the biggest problem. You profit when it's amazing, and you frown when it's dubious to awful. Before I moved, I lived with bad reception that would tank my phone if I didn't leave it plugged in over night. Lay it down at 73% or so, wake up and your phone was dead. Screen was off entire time, nothing was really running, but that's how much of an effect the reception issues can have on a phone. Happened to iPhones, Windows Phones, Android Phones, etc... Except those without a SIM card in them. Can leave them sitting in a drawer for a week with BT/WiFi/NFC and everything else turned on and they seem to last forever even with an active internet connection and accounts hooked up to them.