I have always gone by the rating, 65 or 105 watts. I then saw PL2 on product articles this past year so that 105 or 125 Watt CPU may actually want a lot more than that and it would be nice to know when you are spec'ing out a system. The power and heat of Apple Silicon is quite attractive to me, at least during the summer when an inefficient system can heat up my basement. I don't really mind it in the winter though it's better to be efficient overall. I can run my mini around 300% CPU and the air coming out isn't even warm. My i7-10700 is that way most of the time too. I can get it to warm up, though at 20% CPU. It has a huge air cooler so the CPU stays fairly cool but you can feel the warm air coming out.
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.
M1X worst case scenario
- Thread starter JMacHack
- Start date
- Sort by reaction score
I have always gone by the rating, 65 or 105 watts. I then saw PL2 on product articles this past year so that 105 or 125 Watt CPU may actually want a lot more than that and it would be nice to know when you are spec'ing out a system. The power and heat of Apple Silicon is quite attractive to me, at least during the summer when an inefficient system can heat up my basement. I don't really mind it in the winter though it's better to be efficient overall. I can run my mini around 300% CPU and the air coming out isn't even warm. My i7-10700 is that way most of the time too. I can get it to warm up, though at 20% CPU. It has a huge air cooler so the CPU stays fairly cool but you can feel the warm air coming out.
Previous generation iPad Mini was the A12. So I need to find the deltas between the A12-A13-A14.
Edit: I just read that the A14 offers 40% higher CPU than the A12. Which means no improvement from A14 to A15 if this is all correct.
The Apple A14 Bionic is a System on a Chip (SoC) from Apple that is found in the iPhone 12 and iPad Air (2020) models. It was announced late 2020 and offers 6 cores divided in 2 performance cores (Firestorm) and four power efficiency cores (Icestorm). Apple states that the A14 offers a 40 percent higher CPU performance than the old A12 in the iPad Air 2019. That should translate to a modest gain of around 20% compared to the Apple A13. The A14 integrates a fast 4-core-GPU and a 16-core Neural Engine that is able to reach a maximum of 11 TOPS. The memory controller supports LPDDR4X and 4 or 6 GB are directly above the chip (package on package - PoP).
Processor Comparison - Head 2 Head
Apple A14 Bionic vs Apple A12 Bionic - Benchmarks, Tests and Comparisons
Previous generation iPad Mini was the A12. So I need to find the deltas between the A12-A13-A14.
Edit: I just read that the A14 offers 40% higher CPU than the A12. Which means no improvement from A14 to A15 if this is all correct.
The Apple A14 Bionic is a System on a Chip (SoC) from Apple that is found in the iPhone 12 and iPad Air (2020) models. It was announced late 2020 and offers 6 cores divided in 2 performance cores (Firestorm) and four power efficiency cores (Icestorm). Apple states that the A14 offers a 40 percent higher CPU performance than the old A12 in the iPad Air 2019. That should translate to a modest gain of around 20% compared to the Apple A13. The A14 integrates a fast 4-core-GPU and a 16-core Neural Engine that is able to reach a maximum of 11 TOPS. The memory controller supports LPDDR4X and 4 or 6 GB are directly above the chip (package on package - PoP).
Processor Comparison - Head 2 Head
Apple A14 Bionic vs Apple A12 Bionic - Benchmarks, Tests and Comparisonswww.notebookcheck.net
Yep. Which again suggests that A15 prioritizes the battery life over the unnecessary performance improvements. At the same time, Apple tends to be very conservative with their performance estimates, so I’m sure there will be plenty of things to discover.
Hmm, during the debate about whether the next Apple silicon Mac SoC is going to be based on the A14 or the A15 cpu cores, I've been of the opinion that they would be the next generation A15 CPU cores. It hadn't occurred to me that the question might be moot. If the A14 and A15 cores are the same then both sides of the argument were right and both were wrong. ?Previous generation iPad Mini was the A12. So I need to find the deltas between the A12-A13-A14.
Edit: I just read that the A14 offers 40% higher CPU than the A12. Which means no improvement from A14 to A15 if this is all correct.
The Apple A14 Bionic is a System on a Chip (SoC) from Apple that is found in the iPhone 12 and iPad Air (2020) models. It was announced late 2020 and offers 6 cores divided in 2 performance cores (Firestorm) and four power efficiency cores (Icestorm). Apple states that the A14 offers a 40 percent higher CPU performance than the old A12 in the iPad Air 2019. That should translate to a modest gain of around 20% compared to the Apple A13. The A14 integrates a fast 4-core-GPU and a 16-core Neural Engine that is able to reach a maximum of 11 TOPS. The memory controller supports LPDDR4X and 4 or 6 GB are directly above the chip (package on package - PoP).
Processor Comparison - Head 2 Head
Apple A14 Bionic vs Apple A12 Bionic - Benchmarks, Tests and Comparisons
Or I guess that one way to reuse the M1 cores and not have people complain that the single core performance is less than the iPad mini is to down clock the A15 cores so that they aren't faster.
Back in February 2021, Twitter user "Longhorn" posted some presumed Apple Silicon codenames they claimed were in macOS:
If they are accurate, then the performance cores for A15 and "M2" would be different than the "Firestorm" cores in the A14 and M1.
Interesting thing about the Apple t6000, is could "C" stand for "JadeC-Chop" and "D" stand for "JadeC-Die" which are the codenames Bloomberg said are for the 16 GPU core and 32 GPU core "M1X", respectively.
See my edit. But from at least the A15 specs given in the keynote, it doesn't seem like the performance cores go much beyond the A14 CPU cores. It could be that the A15 used in the iPad mini and iPhones is deliberately trading performance for battery as @leman suggests.
If the CPU performance is the same between the iPad Mini 5 and the iPad Mini 6, then I don't see a strong reason for upgrading.
USB-C. Cellular 5G. Pencil 2. Much improved screen to bezel ratio. Overall just a better design with the smaller bezels and the touchID in the power button.
I used my DTK refund to put in an order for the iPad mini with 5G. It is probably going to replace my 2020 iPad Pro which I don't use much since the release of the M1 MBA.
We still mainly use Lightening and don't need cellular though it's a consideration. I'm not an artist. It's overall hard to justify with how I use the iPad.
Yeah they did. I remember a bit of worrying happening here back then when they compared the iPad Air with the old one and the improvements they were posting were either the same as or not much more than the difference between the A12 and A13. Then the iPhone 12 was announced and they only compared the A14 to the competition as well. Then the benchmark numbers came out and it was in-line with the expectations at the time. And then the M1 came out and we know how that went. In fact, if you look at the 12 Pro page on archive.org you'll notice they made no direct comparisons to the A13 in the A14 section.
I found some interesting wording on the iPhone 13 Pro webpage:
"All-new 5-core GPU: delivers up to 50% faster graphics performance than any other smartphone chip"
Don't want to read too much into this but that wording would seem to include the A14 Bionic too.
"All-new 5-core GPU: delivers up to 50% faster graphics performance than any other smartphone chip"
Don't want to read too much into this but that wording would seem to include the A14 Bionic too.
I saw a leaked benchmark last week on iCaveDave indicating higher GPU performance. I don't recall the net number though. The better GPU may be necessary for the camera improvements.
The Mini has just gone through a 2 year dormant cycle. If it is going into another two year dormant cycle assigning it the A15 is more so to cover the gap. Pretty doubtful it is downclocked if they have done a good job.
When the Macs get M2 upgrades then they'll be faster than the M1 systems also. The whole "single core benchmark score envy issue" is only an issue inside of extreme tech porn forums; not with the vast majority of system buyers.
The A15 is more likely a slightly TMSC N5 node tweak ( N5P ) . The caches are substantively bigger. They could have bumped to LDDR5. There is little need to downclock them to get more battery life out of a updated iPhone or iPad. In most cases, better and faster "race to sleep" would be a lower overall power reduction.
The 5G mm coverage range of the iPhone 13 is larger. So highly likely there are new radio chips in this. If this is Qualcomm on Samsung N5 instead of N7 then that is probably helping on battery life and that has nothing to do Apple sprinkling magical pixie dust on their own Silicon.
The screens on the iPhone 13 are different. If dialed back from 120Hz more most of the time could get a net decrease in battery consumption. Apple's battery tests are watching video ( pretty close to zero need for 120Hz screen ) and audio playback ... definitely zero need for 120Hz screen.
Finally, for iPhone 13 the whole device is thicker and heavier. Probably a bigger battery is a contributor to longer operating lifetime also.
As long as people are throwing out wild guesses, I might as well throw one out also.
The A15 CPU cores seem to be designed, not for higher IPC or clock rate, although those numbers might go up a few percent from the A14. The cores were designed for significantly better area and power efficiency. Thus more CPU cores can be put on a single die. And more CPU dies can be put in a single package without exceeding system thermals. Thus the M1X is more A15 cores of some combination, and the M2 is multiple M1X dies tightly interconnected in a single package, with double (or more?) the total system memory and IO bandwidth. So the M1X and M2 are the same and different at the same time.
The A15 CPU cores seem to be designed, not for higher IPC or clock rate, although those numbers might go up a few percent from the A14. The cores were designed for significantly better area and power efficiency. Thus more CPU cores can be put on a single die. And more CPU dies can be put in a single package without exceeding system thermals. Thus the M1X is more A15 cores of some combination, and the M2 is multiple M1X dies tightly interconnected in a single package, with double (or more?) the total system memory and IO bandwidth. So the M1X and M2 are the same and different at the same time.
Not much hard evidence of that yet. Apple could have just bloated an 88 mm2 die size out to 98-108 mm^2 to add the extra GPU core and system cache. The larger system cache would help with raising IPC without huge core (function units and pipeline ) changes.
Need to see how big die got and if they stayed with N5 or took a "short hop" to N5P.
A14 88 mm^2 ( N5)
A13 99 mm^2 ( N7P )
A12 83 mm^2 ( N7)
A11 88 mm^2 ( N10 )
A10 125 mm^2 ( N16 )
They have done it before to use a '1/2" node ( N7P) to go to bigger die but use the more modest processor reduction to not go much bigger. So not Apple doing "area / power" as much as picking up area/power from the node tweak doing most of the work. A GPU core for A13 was about 3.5mm^2 so if go from 80 to 90 with a 10mm bump it would fit along with some other stuff without major issue.
The size of the A series die has bounced around within a range.
The M2 is very likely going to be in the same die size range as the M1.
M1 121mm^2
A12X 127mm^2
A11X 97mm^2
A9X 129mm^2
If it still has to fit inside of an iPad Pro logic board there is just a cap to how big it can get.
Back in the A series there was about a 40mm^2 between regular A and extra-large (XL) 'X' variant.
Same thing is likely true for M1. The 'X' is a substantively bigger die than the starting point. If go from 8 GPU cores to 32 GPU cores that is going to be a much bigger jump ( for example if a GPU core is 3.5mm^2 ; 3 * 8 * 3.5 = 84 mm^2 .. that is about the total size of a 'plain' A series just for that subsystem addition. going to need cache and other infrastructure to go with that. )
The M2 is likely going to have that classic delta area change different relative to the A15. It is just in the "bigger than an A" state. From there Apple will build a bigger building block.
If going to join dies on a MCM package then the each die will need some inter-die connectivity logic. That also will raise the die size. The chip will be more expensive so Apple will generally put it into more expensive systems ( so a smaller fraction of total cost).
For the M2 they aren't going to want to "pay" that overhead cost (both in logic board space and overall system cost).