30% power reduction compared to N5 v1.0 maybe, but certainly not after N5P (A15) and N4P (A16).
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.
Other A17 Pro. A new step forward.
- Thread starter Populus
- Start date
- Sort by reaction score
30% power reduction compared to N5 v1.0 maybe, but certainly not after N5P (A15) and N4P (A16).
A16 was N4 not Pro
And even discounting both "refining nodes" upgrades as I mentioned a few posts back, you would have expected at least a tiny bit of power uplift for just 10% more performance, the same way it happened with N7 to N5.
This, I don't know. It's bad. Either it's Apple's fault or TSMC's, but it's bad.
Maybe there's still debate, but the reverse engineering pros consider A16 fabbed using N4P.
Apple APL1W10 A16 Bionic TSMC N4P FinFET Application Processor Digital Floorplan Analysis | TechInsights
Apple APL1W10 A16 Bionic TSMC N4P FinFET Application Processor Digital Floorplan Analysis
They mentioned a 4nm process in their keynote and not "second generation 4nm" like they did with the A15 and M2. So yeah, there's a lot of confusion. But if the A16 was N4P then, as right now with the A17, the results are extremely disappointing.Maybe there's still debate, but the reverse engineering pros consider A16 fabbed using N4P.
Apple APL1W10 A16 Bionic TSMC N4P FinFET Application Processor Digital Floorplan Analysis | TechInsights
Apple APL1W10 A16 Bionic TSMC N4P FinFET Application Processor Digital Floorplan Analysiswww.techinsights.com
The A17 Pro is obviously not that much faster.
www.golem.de
Benchmark-Boost: Geekbench Scores sind manchmal höher unter iOS 17
Die ersten Benchmarks zum neuen iPhone 15 sind schon da. Beim Vergleich mit Geekbench 6 muss jedoch aufgepasst werden: Die Ergebnisse sind offenbar nicht vergleichbar.
Oh my god, maybe I should edit the thread’s title to add a question mark at the end of it.
Apple may have fixed vulnerabilities in the processor design that result in a major performance hit. (Like Spectre & Co.) Does anyone know if there were any major vulnerabilities in the ARM design?
More like a step backward...
Another reviewer confirmed dramatically increased power draw.
Another reviewer confirmed dramatically increased power draw.
Many people were expecting A17 to be three steps forward. A17 is more like half.
Coming from a 13 Pro. Speed wise, no complaints, but do not want something that sucks more watts...
Yeah and even more were expecting (hoping) for the three steps forward because of what it might mean for M3 Pro/Max. Personally I don't care about iPhone ray tracing or anything like that but I did think M3's new 3nm process would be a huge jump vs. M1 Max.
We shall see. So far A17 numbers are not impressive.
Except Resident Evil Village isn’t available on iPhone yet, so how exactly were they able to install it and test its performance?
I'm waiting for a special sale in my country that occurs in November (200 bucks off), but so far I'm not that enthusiastic about the jump from my regular 13. Currently waiting for more reviews, especially about any improvement on the awful watercolor pictures in the processing pipeline.
But the Beta version I’ve seen (US reviewers) is locked at 30fps with no way to change it. This review from China says “Geekerwan displayed that the phone thermal throttles quite aggressively and the frame rate falls from mid-40s to around 30 FPS.”
So something isn’t quite adding up . It shouldn’t be hitting 40fps ever.
Geekerwan says they modified the config.ini file to change the resolution and FPS.
See, that’s a bad test, then. You can’t run some unofficial configuration and then expect it to give useful information about anything
Full review of A17 Pro, benchmark comparisons with previous chips, and power consumption. Very interesting video, now in English.
I would agree… except OLED panels are self-emitting and the display driver is what’s lighting up the panel (unlike an LCD, there is no separate backlight). Additionally, OLED panels use current, not voltage, for activation, which means power reduction in driver results in bigger power savings.
I believe the general consensus is that 2024 will be the year of 3nm.
A17 has been a nice way for Apple to flex its muscles, nice product, could've been a trainwreck and it isn't, but we're still not at prime time for the 3nm node, for both volume production and efficiency.
Next year it'll be probably scrapped altogether and we'll see binned A18 and full-fat A18 Pro.
I can see Apple not rushing the M-series on this node, probabily wouldn't scale well for the higher power chips (Pro, Max, Ultra).
A17 has been a nice way for Apple to flex its muscles, nice product, could've been a trainwreck and it isn't, but we're still not at prime time for the 3nm node, for both volume production and efficiency.
Next year it'll be probably scrapped altogether and we'll see binned A18 and full-fat A18 Pro.
I can see Apple not rushing the M-series on this node, probabily wouldn't scale well for the higher power chips (Pro, Max, Ultra).
Last edited: