The A17 is more efficient at the same performance level and it’s not the same architecture. This has all been explained.
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.
M3 Chip Generation - Discussion Megathread
- Thread starter scottrichardson
- Start date
- Sort by reaction score
I agree with the above, although I wouldn't phrase it in terms of marketing people forcing the chip division to do anything. It's just the next reasonable direction for Apple Silicon to go towards.
This is where you lost me. They did build a new microarchitecture for A17, and it is more efficient than the 5nm architecture. It almost sounds as if you have expected the 3N to double the performance at the same power level. Instead we got 20% higher clock at the same power. Which is kind of in line with what TSMC promised?
You want battery increases? Then you need either bigger batteries or you need to sacrifice performance.
That's two generations, the M2 uses the same processor cores as the A15.
Great battery life is one of the most coveted aspects of a new phone for potential buyers. The fact that the 14 Pro and now the 15 Pro have average to poor battery life twice in a row while the performance has increased by 10% year-over-year tells me something is wrong here. Don’t you think? Do people buy a new phone when it’s 10% faster or when it has 1-2 hours extra battery life? Apple chose 10% performance two years in a row at the big cost of battery life. Why do you think that is?
Marketing is not making decision on the clock frequency of the CPUs, rest assured.
As evidenced by... the new microarchitecture of the A17 Pro? You either don't believe new CPU has wider decode, improved branch prediction, more execution units in the integer side, bigger reorder buffer... or you don't believe that constitutes a new microarchitecture. Which one is it?
Again, as evidenced by... the more efficient A17 Pro? Or is your evidence that you are confusing efficiency with power output?
Plus something strange seems to be happening. The new Apple Watch S9 CPU is not efficient anymore. Yet the display is twice as bright? On the same battery? Where's that energy coming from? Don't the newer Apple Watches also measure body temperature now too? Doesn't Apple report the iPhone 15 Pro as having more battery life than the 13 Pro? Isn't the display also brighter now too? And both the newer Apple Watches and iPhones support dual-frequency GPS too, which requires more energy?
It kinda seems like Apple has somehow been able to cram more features on these inefficient chips without worsening battery life.
Imagine what would people on the forums say if Apple released a hone with the same performance as a two years old model but with one hour more battery life? If Apple wanted more battery they’d make a larger phone. Their marketing department seems to think that is not necessary.
What I’m more curious about is why Samsung and OnePlus can fit a 5000mah battery in the same chassis size where Apple can only fit a 4300mah one. That’s what gives the Android flagship an edge in the battery fight, not the peak CPU power consumption.
Last edited:
It’s because some people want ultra fast performance, some demand a slim phone and some want great battery life. You can’t please everyone. There will always be a compromise.
The wrong conclusion to come to, is that there is some terminal decline in the chip divison.
Quit lying.
iPhone 14 battery life tested — here's how long all four models last
The iPhone 14 Pro Max is the champ, but the Plus is a strong performer, too
As far as Mac battery life, nothing comes close without cutting the CPU/graphics performance nearly in half when running on battery mode.
The reason why I think Apple’s division has lost their mojo is because they are no longer disciplined. Pre M1, when all the best Chip designers were still at Apple, they had strict guidelines: fight tooth and nail to not increase wattage by 0.1. Now that those same best in class chip designers left after A14/M1, look at we have now: A16 and A17. Both have dramatically increase power consumption for small performance e increase.
What made Apple’s chip division special was they didn’t think like this. Now, Apple’s chip division is no different than AMD’s, for example. I mean, Jesus. The A17 peaks like an M1 chip. That’s ridiculous.
Given the timeline of silicon development to release being about 3 years, this narrative about Apple being doomed because some designers (who’s names you only know because the media is trying to make this a narrative) left doesn’t add up with this generation of chip…
Really this is all speculation. None of us know anything about the “discipline” of the team. Chip designers move around. Some have left Apple, some have joined. Given the Nuvia fiasco, there is no reason to think the ones who left are indispensable.
You keep repeating things about the A17 that aren’t true. I’m not sure why. It has delivered among one of the biggest increases in performance in absolute terms for single core scores. It’s very impressive.
I never said anything bad about the performance increase. Both A16 and A17 are solid y/y updates in performance. However, at what cost? The peak power consumtion of both A16 and A17 have skyrocketed. That’s two years back to back and most reviewers say the iPhone 14 Pro’s and 15 Pro’s have average to poor battery life. Apple’s competitive advantage was performance per watt. If we go down this path, this advantage is slipping away fast.
iPhone 14 battery life tested — here's how long all four models last
The iPhone 14 Pro Max is the champ, but the Plus is a strong performer, too
Why are you lying about the Pro’s battery life? Even the average battery life of the regular 14 beat out the Samsung Ultra.
iPhone 14 battery life tested — here's how long all four models last
The iPhone 14 Pro Max is the champ, but the Plus is a strong performer, toowww.tomsguide.com
Why are you lying about the Pro’s battery life? Even the average battery life of the regular 14 beat out the Samsung Ultra.
its no secret that the iPhone 14 Pro’s had poor battery life.
Last edited:
They have a seemingly wider performance range, and given physics, yes power consumption must increase. As indicated earlier, they are most likely adapting their strategy to accommodate desktop performance. This is a good thing. No longer will a Studio be forced to run close to the speed of a phone.
I disagree about battery life. We know little about the battery life of the 15 yet, and I’ve seen plenty of reports of good battery life of the 14. In any case, you do realise that the SoC is just one component of battery life. If battery life problems you claim exist turn out to be true, that doesnt necessarily mean the A17 is the culprit.
What would good battery life be for you?
Just as an anecdote. I got my 15 Pro yesterday. I didn’t charge it. I updated and migrated my data while I was in the gym and had no charger. I ran multiple demanding benchmarks and tests on it. I didn’t charge it overnight. I used google maps navigation on the phone today for about one and half hours and spent considerable time using the browser on 5G connection.
Right now my battery is at 40%. I don’t know whether it’s good or bad. But it’s certainly much more than I ever got out of my 11.
At any rate, I’m looking forward to the battery tests done by reputable reviewers, such as DxOMark.
From the drawings, more like M3 Extreme...!
- 64-core CPU (48P/16E)
- 160-core GPU (w/hardware ray-tracing)
- 64-core Neural Engine
- 960GB ECC LPDDR5X RAM (Sixteen 64GB chips)
- 2TB/s UMA bandwidth
I wonder why Apple hasn't improved heat dissipation if they were going to increase the power consumption.
My guess is they are trying to decrease scenarios where extended P core use is needed. We know that the E-cores are significantly better than before and they are promoting things like MetalFX using the ANE and RT cores for AAA games.
My guess as well. Allow the P cores to clock higher for short burst workloads, and then downclock them back to a more efficient 3.46GHz. Don't think Apple engineers intend to have the P cores to use the full 14W for long periods of time, regardless of thermals.