New Geekbench scores are up for the M2 Max

[Retskrad]

Mac14,6 - Geekbench

Benchmark results for a Mac14,6 with an Apple M2 Max processor.

browser.geekbench.com

2027
Single-Core Score

14888
Multi-Core Score

 

[Juicy Box]

Better than the other ones that were leaked.

Those first leaks were pretty sad, imo.

 

[Xgx]

3.68 base freq 🤔

 

[purplerainpurplerain]

If it broke 2K single that’s real good. I expected 16-17K with 12 cores and that clock speed.

 

[theorist9]

3.68 base freq 🤔

Yeah, just noticed that myself. The last "M2 Max" was 3.54 with an SC of 1889; the production M2 in the 13" Pro and Air is 3.49. The variation in freq suggests these scores are from preproduction devices.

The SC score of this 3.68 variant is what we'd expect based on increased clock alone. GB gives an average SC value of 1899 for the production M2 in the 13" Pro; extrapolating gives us 1899 x 3.68/3.49 = 2002, which is within normal GB variation of 2027.

[As an aside, for the M-series, those listed frequencies aren't the base frequencies. They're the SC frequencies. The base (all-core) freq for the M-series P-cores is a bit lower. The M1, for instance, actually has a modest (7%) "turbo boost" of the SC clock over the base clock.]

 

[fakestrawberryflavor]

Could be that they have not finalized what they want the clocks to be, or, they are testing lower clocks for notebooks and high clocks for desktop devices.

Either way, I’d like a 2000+ SC but was hoping more for a 2200+ SC for at least desktop class workstations.

 

[Bodhitree]

3.68 base freq 🤔

Yeah I hope Apple aren’t going to get into the Intel game of pushing the frequencies as high as they will go. I thought the focus on performance per Watt and low-power usage was spot on, especially for laptops which are most of Apple’s sales. It would be a shame to see them move into the Intel camp.

It could be an attempt to overclock the entire SoC for desktops, mainly the Mac Studio where there is some thermal headroom as well as plenty of power from the wall.

 

[theorist9]

Could be that they have not finalized what they want the clocks to be, or, they are testing lower clocks for notebooks and high clocks for desktop devices.

Either way, I’d like a 2000+ SC but was hoping more for a 2200+ SC for at least desktop class workstations.

I'd like to see that as well but, with the current microarchitecture and process node, they'd need to increase the P-clore clock to 2200+/1900*3.49 ≈ 4.0+ GHz. No one not under an NDA knows whether the M2 can go that high and, even if it could, whether Apple would be willing to do that with its desktops.

At least the increased TDP at 4 GHz should be readily manageable in a desktop, given how low their TDP's are now, and given that increasing from 3.5 to 4 GHz shouldn't increase it much more than (4/3.5)^2 – 1 ≈ 30% (plus they could limit how many cores can run at the higher clock).

 

[leman]

That's much more reasonable given the expectation of a higher clocks. Maybe this new score is with the performance mode on? At any rate, this could be a very decent laptop for 2023. Especially if you get this performance on battery.

 

[OllyW]

Please continue in the 'M2 Max' Geekbench Scores Surface Again Online Ahead of Release in 2023 news discussion thread.