I know, I didn't say the watercooler was a must, rather an option. My point was that it needs decent cooling.
I've run a 13600K under a NH-U12A and NH-D15, but those are almost the size a a mac mini
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.
First Geekbench Scores for New Mac Mini With M2 Pro Surface, Beats M1 Max
- Thread starter MacRumors
- Start date
- Sort by reaction score
I know, I didn't say the watercooler was a must, rather an option. My point was that it needs decent cooling.
I've run a 13600K under a NH-U12A and NH-D15, but those are almost the size a a mac mini
ikr?
i was hoping for a bigger bump, been riding the M1 13" waiting for a significant upgrade on the MX Pro line.
Might ride a bit more for the M3.
13" 12-core 32/1TB clocks in at 3500€, it's not cheap.
since you say it share the exact same cpu...its impressive that the M2 pro is better that M1 Max in SC and MC
If the M2 pro was on new architecture 3nm, then yes, not so impressive since we would expect it to be better
That's not the point tho.
Better efficiency = less heat.
Less heat = better performance at same cooling capabilities.
More heat = stress on components.
Intels are pushing the envelope in terms of heat management.
Apple could've clocked these higher, but they would guzzle battery and make fans spin harder, or would need to be larger.
Really nice numbers.
BUT:
Realistically speaking. How much of a number increase would you recognize as an improvement in the real world scenario?
How is that impressive?
(10core) M1 Max vs M1 Pro difference was never in CPU.
M1 Max vs M1 Pro difference was extra memory bandwidth - which when tested was shown to be mostly to accommodate bigger GPU.
M1 Max vs M1 Pro scored roughly the same on CPU benchmarks, because they were more or less the same on the CPU side. The difference was on the Metal benchmarks because of the GPU.
Mwah, 10%-ish faster single core and 15%-ish on multi core? (1774/12327 on my MBP 16" M1 Max). Not impressed actually, M2 having more cores on the CPU and a higher clock-speed.
And the price increase is crazy, the M2 Max with (almost) the same specs as I have now will cost me up to € 500,- extra...
I think I will hold on to my beloved M1 Max for two more years, or maybe even more.
And the price increase is crazy, the M2 Max with (almost) the same specs as I have now will cost me up to € 500,- extra...
I think I will hold on to my beloved M1 Max for two more years, or maybe even more.
For Logic Pro and in my experience, real world improvements always correlated pretty much in line with Geekbench scores.
The only difference with Max over Pro is the GPU so not really. M1 Pro has the same SC and MC to the M1 Max. The M2 pro has 2 more cores and all of those cores are running faster. So its not impressive its inevitable but it will run hotter so interested to see what the difference is on longer processing tasks.
And while the improvements are conservative and less than what i have hoped for, the speed bump is still bigger than if you upgraded from the 4/4-core M1 to 6/2-core M1 Pro.
(god i wish it would become the norm to type out efficiency/performance cores)
(god i wish it would become the norm to type out efficiency/performance cores)
This shouldn’t be surprising, from a CPU perspective the Pro and Max chips are almost identical save for extra memory bandwidth on the Max. GPU is the key differentiation so you should expect a Pro variant from one generation to beat the Max variant from the prior in terms of CPU. The M2 Max should have similar benchmark figures. (EDIT: This is not counting the binned lower core count version of the Pro that’s available on lower end models).
Last edited:
Appearentely the new M2 Mac Mini has a better cooling system as it weight more than the M1 Mac Mini (accorind to MaxTech - > see
But I'd like to see how the temperature behaves on the MB Pros- with the new M2 Pro/Max, as they have less space than a Mac Mini and therefore the cooling needs to work even more efficient. Wonder if we finally will hear the fans?
There’s a link in the article to the benchmark entry. It’s the binned 6+2 core version of the M1 Pro that’s only available on the lowest tier 14” MBP. The full 8+2 core version that’s standard on all upper models of the 14” and all of the 16” ones is closer to the 12,000-ish score the Max pulls in.
"For comparison, the previous M1 Mac mini achieved a single-core score of 1,651 and a multi-core score of 5,181." LOL so funny, so wrong. according to Geekbench the scores are 1715/7442. https://browser.geekbench.com/macs/mac-mini-late-2020
Whoever did this first set obviously did not know what they were doing. Here is a primer:
1) make sure the system is not running any tasks (even indexing of hard drives), as other tasks will impact performance
2) make sure the system is at normal operating temperature
This is not rocket science. And a point to ponder, many tests seem to deliberately skew down the results, as Geekbench uses averages skewed down by poorly run tests which make ridiculously low scores (a typical process would either account for too much load on the CPU from other sources, or even remove the obvious outliers (both too high and too low). and this happens to every entry in the Geekbench average tables, not just Apple devices
Whoever did this first set obviously did not know what they were doing. Here is a primer:
1) make sure the system is not running any tasks (even indexing of hard drives), as other tasks will impact performance
2) make sure the system is at normal operating temperature
This is not rocket science. And a point to ponder, many tests seem to deliberately skew down the results, as Geekbench uses averages skewed down by poorly run tests which make ridiculously low scores (a typical process would either account for too much load on the CPU from other sources, or even remove the obvious outliers (both too high and too low). and this happens to every entry in the Geekbench average tables, not just Apple devices
I think the only people who say this are 13 year olds who don’t pay their own power bills. And no, I’m not turning up the heat in the house!
I think they will wait to upgrade the Studio to m3 instead.
Gives me time not to be tempted and save up for a Studio.
The M1 Neural Engine is rated at 11 trillion ops/sec from Apple’s marketing whereas the M2 is rated at 15.8 trillion ops/sec, so it should be a substantial increase. All variants except the Ultra have the same number of Neural Engine cores so this shouldn’t change.
Apple advertised Adobe Photoshop being 40% faster on M2 generation machines, and Adobe did some significant tuning to use CoreML and in turn the Neural Engine for certain operations when they updated PS to support Apple Silicon.
What the actual benchmarks look like remain to be seen but Neural Engine boosted CoreML tasks should definitely get a boost.
Impressive since is better than the M1 Max
If M2 Pro was slower or the same as the M1 Max...that would be the opposite of impressive
If M2 Pro was under "3nm" and gain the same difference, again not that impressive
Very disappointing! And there is no support for AV1 at the hardware level.
What was the point of leaving Intel I still don't understand, maybe Apple's arrogance.
Not calling out which chips are binned and not is lazy reporting, especially when you are throwing in a binned M1 Pro and comparing it to an unbinned M2 Pro.