No we were talking about smartphone chips and how Qualcomm's lack of innovation was also leading to Apple refusing to provide decent updates to their processors. This year they just gave up and shipped the iPhone standard with last year's chip.
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
The 23.7% for multi-core are great — but they're accomplished in part by adding cores, which Apple won't always be able to do.
It's the 12.1% for single-core that I'm more concerned with. There's also no surprise here; we already know what the cores are like from the Ax CPUs.
Recent Ax single-core improvements:
So you can see there's a bit of a downward trend. It seemed to stabilize around 20% for a while, but the A15 (relevant here for the M2 Pro/Max) and the A16 were a fair bit lower in their improvements.
Now, the M3 might be different, if it uses A16 cores but at 3nm, e.g. by running them at 4 GHz, or by adding a lot more cores.
Like I said, nothing terrible, but I'm hoping Apple has significant design improvements up their sleeve for the A17 or so.
"A little more RAM" like 48+ GB is only needed for heavier work, which means more heat removal, which means Studio. But we do not have an M2 Studio yet, which is problematic.
There is no 'battery life' measurement needed for a desktop. That is what a Mac mini is. You must be considering the MacBook Pro, which wasn't tested here
For anyone looking for raw peak performance, wattage be darned, though, there’s not an “either camp”
Well, unless it’s “either Intel or AMD”!
Intel made a huge mistake 10 years ago. Now 12,000 workers are paying the price.
Vox is a general interest news site for the 21st century. Its mission: to help everyone understand our complicated world, so that we can all help shape it. In text, video and audio, our reporters explain politics, policy, world affairs, technology, culture, science, the climate crisis, money...
www.vox.com
PC’s are not a bad business, don’t get me wrong, but imagine how much MORE money Intel would be making if they were in even just 30% of Android devices…
Not as far as the common benchmark suits would show. The cpu- scores are always within margin of error. Seems the bottle neck is the cpu complex and both memory subsystems are enough to saturate it.
Nice chart. Am hoping for v9 ISA for after A16/M3.
N3 will be nice, although based on A16.
Yup. The GPU benefits from the additional memory bandwidth; the CPU mostly does not. Anandtech was able to sustain 243 GiB/s when using all CPU cores (including the e-cores). That's a lot, but nothing close to 400.
And I guess in actual benchmarks (and also in real-world usage), memory factors in slightly less, so the result is within the margin of error of 200 — meaning the M1 Pro and Max CPUs benchmark the same. I wouldn't be shocked if we see the same for the M2 Pro and Max.
At the end of the day, this probably only matters to this crowd, the vast majority wouldn't care anyway, just thats its the newer model.
Interestingly enough, that’s also what Apple’s been doing for years. The A16 is more performant than the A15 because they increased the clock rate slightly and gave it an extra GPU core. “Adding cores for performance” is table stakes in this processor performance game. It’s not an anomaly that companies add more cores provide more performance, it’s one of the ways the game is played.
Intel STILL performs better, and likely always will, though.
I would think that the Mx Max variant Macs will have better real world performance compared to the Mx Pro variants. The 200GB/s bandwidth of the Mx Pro is the theoretical max bandwidth. Anandtech is able to push the M1 Max over 240GB/s during their memory bandwidth test. I would think that for workloads that pushes around large data sets across the two P CPU clusters, the Mx Max likely will perform better compared to the Mx Pro. I guess macOS have to be smart enough to pin the processes to the proper CPU clusters that is nearer to the physical memory it is crunching.
Looks like the binned (10 core) M2 pro scores 12074 with 16GB ram.
Mac14,9 - Geekbench
Benchmark results for a Mac14,9 with an Apple M2 Pro processor.
browser.geekbench.com
My predictions are a m3 Mac Pro and next year a studio. Updating the studio now would conflict with sales for a Mac Pro.
Well, I ordered one. Let's hope airdrop works on the new model. I always have to reboot my M1 Max MacBook pro for the airdrop to work.
In looking at the frequency for the M1Pro/Max compared to that shown in leaks for the M2Pro/Max - it looks like the frequency is upped from 3.2GHz/3.2GHz to 3.48GHz/3.68GHz, respectively.
We can see the speed effect on the benchmark results - but does anyone have an idea about effects on heat? battery life?
We can see the speed effect on the benchmark results - but does anyone have an idea about effects on heat? battery life?
I agree the mini is the ideal form factor and platform for enterprise developers. The studio is overkill. An m2 ultra with 64GB would be the perfect machine for under $2500 for most app devs.
Looks like the Max might be clocked 200Mhz higher than the Pro…
3.68 GHz. You couldn’t give us 0.01 more GHz, Tim? Nice 😩
This is purely comparing the CPU, which means Pro vs. Max is meaningless here and it's just M1 Pro/Max vs M2 Pro/Max in which case, you have newer cores and more of them. In other news, water is wet.
I’m thinking of replacing my son’s 2014 iMac 5k 27” with the M2Pro Mini. Would there be a noticeable improvement on the user experience front (apart from ditching the need to vacuum the ventilation slots weekly)?
I’d need a suitable monitor as well, or course. Any suggestions? Gotta be fit for gaming, too.