Can you do a "real life" test? For example, play youtube for 2 hours, then 2 hours websites browsing (you would need a script to do that), then something else..until both battery end.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
TSMC's A9 Chip Outperforming Samsung's in Early iPhone 6s Battery Benchmarks
- Thread starter MacRumors
- Start date
- Sort by reaction score
Can you do a "real life" test? For example, play youtube for 2 hours, then 2 hours websites browsing (you would need a script to do that), then something else..until both battery end.
Hard to agree with you,
First, FPS games are heavy loaded in CPU and GPU; you have to both parts are in AP
Second; with Geekbench can running 4 hours in battery test, but, in Game, it just running 2:35 hours, Dut to the CPU& GPU are running together, I feel, it will makes the result is worst
Third; i do not think GeekBench is using 100% of CPU load, who said it using 100% ? If 100%, the benchmarking itself become invalid due to the throttling (heavy dependency on device & environment)
Forth; if running car at the difference speed, yes, result should be different; but, people concerning the car is running in the similar speed, and it shows much more than expected (2-3% apple)
No, sorry, I cannot, because as I said, I already brought it (my Samsung chip based 6S) back.
But the Geekbench Benchmark is the same for all devices, so I think, it is very well-suited for comparing them. Furthermore, as mkimid already said, I also think, that the Geekbench doesn't use 100% CPU load, so it nearly is a "real world scenario". Perhaps a little bit more stressful than streaming youtube, but not 100% full load.
Is there a test showing any negative impact to performance with the "slower running" chip?
So many pundits here saying this is a huge issue. But the funny thing is not one is saying how it affects the overall performance.
The battery issue.
I have the TSMC and my initial battery life was 6.2 hours from 100% to 10%.
I restored as new, manually synced then ran the device until black screen.
Recharged it over night and things are significantly better.
I spend very little time at 100% CPU utilisation. So do most people when they aren't running benchmarks
Last edited:
Yeah I think geekbench is not far from a heavy load user real life usage, at least not totally just burn cpu to 100%
I just ran the battery benchmark for a few minutes while monitoring my iPhone using the instruments developer tool. The benchmark seems to hold the CPU usage at around 55-60%
According to Anandtech, the CPU load for the geek bench battery test is 30%. http://anandtech.com/show/9708/analyzing-apple-statement-for-tsmc-and-samsung-a9
I found different results:
https://www.dropbox.com/s/z4mdom3h1irls9z/InstrumentsGeekbench.mp4?dl=0
Perhaps my testing is flawed, or perhaps Geekbench's tests are completely inaccurate. That's probably not something they'd own up to if that were the case.
The point you are missing: Did Apple promise anything that they don't deliver?
A while ago I read that US hospitals would be in trouble if one of their X-ray machines broke down, and they couldn't find an identical replacement, but only a newer, improved model. Because then they would have different X-ray machines of different quality, and anyone being X-rayed with the older model could sue them for not getting the best possible diagnostics...
1. The justification is that the battery is quite obviously the most significant component to determine battery life. After all, it is called "battery" life so you would expect it to be related to the battery. 2. IF it turns out that swapping the battery changes the results, then we know for a fact that the chip isn't solely responsible.
Last edited:
Nice job with testing. With anandtech they were told by the developer it was 30% CPU load however in your actual testing, it's closer to 60%. You should share your findings with Anandtech as they have been known to call out developer, OEMS, etc for BS information. Anyway, thanks for sharing good information.
Apple heard of them. And Qualcomm, NVIDIA, AMD, MediaTek, Marvell, Broadcom, Intel, Texas Instruments, Xilinx, Altera and so on.
Except that you have no idea whether all your "advice" is nonsense or not.
Apple designs its phones to work best for the average user. This test doesn't even measure how much work the processor does, it only measures how quick it can possibly uncharge the battery. Apparently the TSMC chips are not very good when your goal is to drain the battery
That's really all we know. But this battery test is nowhere representative for what the average user does. So neither phone nor chip are in any way optimised for this case. It's like measuring the fuel efficiency of a car by running it on a race track at maximum speed and measuring who long it takes to empty the tank.
For the curious. My Geekbench battery results.
http://browser.primatelabs.com/battery3/130680
From 100% to 0% (shutoff) Screen dimmed, airplane mode on.
Do bear in mind however the CPU utilisation they report is incorrect, it is roughly double this.
http://browser.primatelabs.com/battery3/130680
From 100% to 0% (shutoff) Screen dimmed, airplane mode on.
Do bear in mind however the CPU utilisation they report is incorrect, it is roughly double this.
Except that you have no idea whether all your "advice" is nonsense or not.
Apple designs its phones to work best for the average user. This test doesn't even measure how much work the processor does, it only measures how quick it can possibly uncharge the battery. Apparently the TSMC chips are not very good when your goal is to drain the battery
If tsmc car is much more energy saving with same speed as the samsung car in top speed racing, why do you expect the samsung suddenly energy saving ratio is much higher in lower speed racing? The samsung chip has no design different in general except it just products more heat for the doing same thing and waste energy.
Apple just uses the term "up to" in iphone 6s/6s plus battery spec, that means Apple takes no responsibility at all, even the phone just has 10 minutes life then it is not against the law.
I got the good chip, booyakasha!
As an aside, this is confirmation that either the Taiwanese are just still better at making chips or Samsung intentionally tried to cripple the competition ;P
As an aside, this is confirmation that either the Taiwanese are just still better at making chips or Samsung intentionally tried to cripple the competition ;P
I think it was samsung good salesman skill to make Apple believed samsung could use 14nm to do the same thing with much cheaper cost then fail to deliver, then Apple decides not cut it but not take any responsibility, just make a no prove real life similar performance claim.
Yes, I think so!
On the Apple page it says "Up to 11 h WLAN usage". I mean, "up to" doesn't guarantee anythink, okay! BUT: All the TSMC users easily make it up to 9-10 hours, whereas the Samsung users can just score up to 7, maximum 8. I couldn't even score 7 (!) with just surfing and messaging at 25% display brightness, no games, no video editing, nothing special! And THAT is way too far away from 11 hours, which apple says.
I nearly got the same result: 4:05 with my equal Samsung chip based model.
My friend got 5:51 with her TSMC, enough said. -.-
If I were you, I would bring it back!
Best wishes
Casey
It's past the return period, so I'll be very much at their mercy but I'll drop in later and see what the score is. Chances are I might just get another Samsung though
I had agonised over this possibility myself. I've came to the conclusion that a software fix isn't possible.
The only software/firmware optimisation which could level out the two chips would be throttling.
Given that the Samsung chip doesn't perform higher to any significant degree, such throttling would only result in lower performance.
The chip seems to be "leakier" - this isn't something software can resolve.
Apple Care told me on the phone call, that they admit the low battery usage time as a defect/deficit and as a consequence of this, they would exchange the iphone as many times until I am satisfied with my phone, irrespective of the return period.