New MacBook Pro Reviews: Hands-On Look at Performance and Upgraded Specs

[chucker23n1]

I don't why MacRumors keeps posting these articles.

Because it generates clicks, and because people are interested.

 

[Michael Scrip]

I don't why MacRumors keeps posting these articles.

Because it's about Macs?

 

[theorist9]

Ah. Not sure how things are going on macOS, but I know Geekbench had to move to guessing the clock on iOS long ago, because iOS won't reveal it any more.

Looks like on my M1 Pro Mac, sysctl hw.cpufrequency is no longer set (I think it was on x86 Macs), so maybe Geekbench has the same problem as on iOS that the API just doesn't exist. Instead, you can have powermetrics sample the CPU for a while and give you an average. I guess Apple feels that frequency varies too wildly for it to be meaningful. (There's also a bit of a Schrödinger thing going on. Measuring the frequency affects — ever so slightly — the frequency. Not to mention that the e-cores run on a different frequency than the p-cores…)

TL;DR: I wouldn't rely too much on that number either. And I hope Geekbench 6 allows for showing heterogenous cores.

I think the difference in frequencies is sufficently consistent to reflect a real difference between the models.

Here's a screenshot showing most of the first page returned by a search for "apple m2 max" in Geekbench's Recent Results (14,5 = 14" M2 Max MBP and 14,6 = 16" M2 Max MBP).

Even with the variation, it can be seen there is a systematic difference in base frequency between the 14" and 16" Max's—they appear to be 3.5 GHz (same as the base M2), and 3.7 GHz, respectively:

14": 3.5 GHz (6 results); 3.4 GHz (1 result)
16": 3.7 GHz (7 results); 3.6 GHz (1 result)

Scanning two more pages showed essentially the same thing, which gives about 20 results for the 16" M2 Max at 3.7 GHz, and 20 for the 14" M2 Max at 3.5 GHz, along with a few that are probably outliers.

I'm going to speculate that the difference isn't because the 16" models had high power mode turned on, since I suspect many who ran these scores didn't bother to do that; if so, the default base frequency for the 16" Max models is higher. That does make me wonder what the effect of high power mode is.

I didn't see this distinction when I searched for M2 Pro results—all models (14,9; 14,10; and 14,12 = 14" M2 Pro MBP, 16" M2 Pro MBP, and M2 Pro Mini) are 3.5 GHz.

 

[GuilleA]

You cannot run an Intel VM on an M1 or M2 device.

But you can run Windows 11 for ARM, which is also capable to run (almost all) x86 apps through an emulator.

 

[theorist9]

Passmark testing also reports a frequency difference between the M2 Pro (3.5 Ghz) and M2 Max (3.7 GHz).

I'm surprised no one has devoted an article or youtube video to this, since it does seem interesting.

Apple M2 Max and M2 Pro become PassMark's laptop CPU single-thread top dogs but Raptor Lake-HX is yet to strike

The new Apple M2 Max and M2 Pro chips have become the latest chart champions in PassMark’s laptop CPU single-thread benchmark. The previous leader, the Apple M2, is pushed down to third, and it is just ahead of a trio of Alder Lake-HX parts. The M2 Max and M2 Pro produced decent results but...

www.notebookcheck.net

 

[chucker23n1]

A 6% clock increase but a 2% score increase?

 

[theorist9]

A 6% clock increase but a 2% score increase?

That's not what's important here. A single pair of PassMark results with a 2% difference could be due to run conditions or normal variation. The question under discussion is whether the GB reports of the CPU frequency difference between the M2 Pro and M2 Max are "real". You were skeptical of this, but the fact that PassMark also reports them supports that they are.

Furthermore, if you look the first three pages of recent GB results for both the M2 Max and M2 Pro, pick the three highest SC scores for each (which should be those run under optimal conditions), and average them, you get a difference of {2076, 2075, 2075}/{1971,1970,1970} = 5.03%, which is quite close to the 3680/3480 = 5.07% expected based on the frequency difference.

[Note: As I mentioned earlier, based on the GB results, the higher clock appears to be confined to the 16" M2 Max specifically (model no. 14,6).]