So I wasn't impressed with only a 10% difference between the A16 (5nm) and the new A17 (3nm). But what happened to the battery/ power efficiency that should've come with 3nm? It seems it's barely noticeable and in some cases it's even worse than before?
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.
iPhone 15 Pro What happened to 3nm efficiency?
- Thread starter MBX
- Start date
- Sort by reaction score
Apple seems to have chosen increased performance, which bodes well for M series.
Also, RAM consumes battery, and A17 Pro has 33% more RAM.
Also, RAM consumes battery, and A17 Pro has 33% more RAM.
Battery life on the iPhone is perfectly adequate for the vast majority; Apple used 3nm to improve overall performance. That's the right decision IMO.
What happened to using one of the already existing battery performance threads
To get the most from 3nm, a chip has to be designed from scratch for 3nm. The latest word from engineering circles is that the A17 is basically a ported version of the A16 with GPU enhancements. Next year’s A18 is likely to be significantly improved even on the same process because it’ll be designed for that process.
And this is why I will wait until next year for a purchase over my iPhone 12 Pro. Let the testers with the iPhone 15 Pro be guinea pigs first
No matter what Apple does with the A17 or the iPhone 15 Pro, it'll still have better battery life than my last phone, the iPhone 13 mini. And since that was more or less sufficient for me, the 15 Pro will present no issues for the next couple years until the iPhone 17.
Seems like they decided to push the A17 Pro harder while maintaining the same or slightly better battery than the 14 series.
Fine with me as the 14 Pro Max already had great battery life. It would be nice to have the option of more battery or increased performance. A toggle for performance/gaming mode would be cool but oh well.
Fine with me as the 14 Pro Max already had great battery life. It would be nice to have the option of more battery or increased performance. A toggle for performance/gaming mode would be cool but oh well.
If history is our guide - the jump to smaller transistors always promises increased performance (which is usually always attainted) and increased power efficiency. Almost always, the power efficiency is non existent to very minimal. It usually takes a refinement of the chip (next generation) to get the power efficiency that was promised initially.
At some point you just need a larger gas tank if you want more range. Apple, for whatever reason, sticks with relatively small batteries year after year.
I think they really need to address two things next year, the batteries and the heat, keeping these things cool, along with a larger battery of at least 5000mah should give a large boost in battery performance.
I think they really need to address two things next year, the batteries and the heat, keeping these things cool, along with a larger battery of at least 5000mah should give a large boost in battery performance.
Last edited:
Good thing rumors suggest a 6.9" model next year. They should be able to squeeze more battery inside.
Can't say I've felt my Pro Max getting hot or even warm. Battery life is very good as well.
I think one major problem here is how rumour sites and youtubers built up 3nm to the point where people expected a revolutionary gain. Meanwhile the industry rumours all said that N3B was a problematic node and the gains vs cost was such a terrible ratio that only Apple would even make a move on it.
The good stuff with the A17, that will trickle over to the M3 I'm sure, is that redesigned GPU. Hard to appreciate in a constrained device such as a phone, but once we scale up to the other devices I'm sure there will be some positivity and appreciation going around.
Makes me wonder yet again if there really is an M2 in that headset of theirs.
The good stuff with the A17, that will trickle over to the M3 I'm sure, is that redesigned GPU. Hard to appreciate in a constrained device such as a phone, but once we scale up to the other devices I'm sure there will be some positivity and appreciation going around.
Makes me wonder yet again if there really is an M2 in that headset of theirs.
I have a oneplus 11, and this is NOT the case. Phones getting warm is normal, but phones getting extremely hot to the point of being uncomfortable to touch is NOT normal. I was using a banking app to do some transactions on my 15 pro max, and the heat was insane. I had to close the app and restart the phone as the phone stayed hot even after closing the app with no signs of cooling down!
Well mines is cool and have no problems. Must be a bad batch you guys have.
It's obvious after watching the 15's tear downs, there's a lack of proper cooling and heat dissipation.
The great 3nm “efficiency” was never there to begin with. Many people just assumed it would be massive because people didn’t understand the physics behind the 3nm marketing term.
Terms like 5nm (A15), 4nm (A16), and 3nm (A17) are all marketing only.
3nm isn’t anything except a marketing term. There is not a single transistor in there that is 3nm. The naming only represents the equivalent transistor density and approximate improvement in performance or efficiency, as determined by TSMC. It is TSMC, Intel, Samsung, SMIC that decide what they should call 7nm or 3nm. Improvements have come mainly by material or structural changes, not size of transistor.
Transistors stopped shrinking in a linear manner since about 45nm. We used to have new nodes every year, then in increased to 18-24 months. Today, it will take 3 years before we get to a new "node."
Terms like 5nm (A15), 4nm (A16), and 3nm (A17) are all marketing only.
3nm isn’t anything except a marketing term. There is not a single transistor in there that is 3nm. The naming only represents the equivalent transistor density and approximate improvement in performance or efficiency, as determined by TSMC. It is TSMC, Intel, Samsung, SMIC that decide what they should call 7nm or 3nm. Improvements have come mainly by material or structural changes, not size of transistor.
Transistors stopped shrinking in a linear manner since about 45nm. We used to have new nodes every year, then in increased to 18-24 months. Today, it will take 3 years before we get to a new "node."
It's factually incorrect posts like this that give people false hope. A17 is very much a new microarchitecture already. You can't just port over a 5/4nm chip due to changes in design rules. It's certainly not a rehashed A16.
The reality is Moore's Law stopped working a long time ago and each node offers fewer and fewer benefits. That why TSMC and others are investing so much in chip stacking and packaging technologies for the future.
Where is the performance though? Seems really limited. +10% is not much in mobile terms, esp with a lithography change.
gpu core changes to me are pointless to me on a phone.
Agree overall that the iphone 15 pro looks really disappointing to me. At this point I'd rather just buy a renewed 14 pro or something and save $400.
Last edited:
Not sure what world you're living in, but roughly 10% is about what we typically see for any new generation of chips. The huge increases from the M1 was a rare event that happens maybe once a decade.