Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
For the M2 pro to beat M1 Max in both SC and MC after just 1 generation, its impressive

The M1 Max is virtually the same in terms of CPU as the M1 Pro; it's basically an M1 Pro with more GPU cores (and more memory bandwidth to support them).

I wonder if M2 Max can approach the level of M1 Ultra, in SC yes, but in MC?

It'll handily beat it in SC. Even the MacBook Air's M2 (with no suffix) does.

In MC, no. The Ultra scores 23325; the 15013 don't come close to that. Which makes sense — the M1 Ultra has 16+4 cores, whereas the M2 Pro and M2 Max have 8+4.

But i still suppose the gpu will be more powerful on the M1 Max compared to M2 Pro

Probably, but it could be close!
 
For the M2 pro to beat M1 Max in both SC and MC after just 1 generation, its impressive
I wonder if M2 Max can approach the level of M1 Ultra, in SC yes, but in MC?
But i still suppose the gpu will be more powerful on the M1 Max compared to M2 Pro

The M2 Max is going to be very fast compared to the M1 Max – if this result for the Pro is true, I think we can expect a 30% increases in GeekBench for multicore. That really will make a big difference to all those out there compiling large chunks of code, or outputting 4K/8K video. If I were working in those kinds of professional fields, I'd be looking to upgrade. But for the rest of us, well, it's basically like the second gen of a car being able to go at 180MPH, while the older version can go at 150MPH. Nice on paper. But functionally meaningless.

I must admit I'm impressed. I thought this would be a small speed bump but it looks like it's very respectable. I still want to see those real world battery figures, though, because that increase in clock speed and the more core counts don't come free of charge.
 


While we've yet to see Geekbench scores for the updated MacBook Pros, M2 Pro's performance with the Mac mini is unlikely to differ greatly from its performance with the new Mac notebooks.

This is because Geekbench does not expose the thermal limitations of the notebooks.

For intense real-works tasks, the Mac Studio outperforms the MacBook Pro due to the Studio’s improved cooling, which means throttling occurs much later (if at all) on the Studio compared to a MacBook Pro with identical configurations.

It will be interesting to see if the same holds true for the Mac Mini with the M2 Pro.
 
creditable but not earth shattering, and I still have gut feeling that Apple were not that impressed with the M2 in general, especially with delays causing it to be on the heels of the M3 where I believe a much greater efficiency in both energy and productivity will occur. Also question why they bother to have different implementations of m1/m2 rather than go for the best to start with and possibly go to the dual chip implementation for higher end. Costs a lot less producing one implementation. I know its SOC and I know much of the benefit is by virtue of it being SOC, but sure Apple could sort that out. I know with the M1 the top end was basically a dual chip, but cost and even performance could be enhanced by having one M3 chip based on the best Apple could do, as the performance hike on every device would be worthwhile but still not preclude the power hungry users having dual chip etc.

I was very unimpressed with the base M2 air both from its SSD slow down and its less than impressive performance hike.

After testing the MBA M2 I've postponed any thought of updating the iMacs, Mac mini's and MacBooks we have, as they still do their job well but will revisit it when the M3 shows up in a device.

I don't upgrade for the prestige, not the colour of the device, and it does make me smile when I see the only criticism is 'they don't do it in x colour'....
 
…and since the bulk of Apple’s Mac business is ultra-portable laptops and ultra small-form-factor desktops, that’s the key measure of their success. Also, this is just Geekbench - application benchmarks that take advantage of the media engine, neural engine etc. may show a greater advantage over Intel.
There is a Create ML benchmark: https://github.com/jcampbell05/ml-training-benchmark

Would be interesting to know the difference between M1 and M2 neural engines.
 
So raptor lake 13600k is faster than this… same with zen 4.

x86 is still impressive and will only get better as Intel ramps intel 4, 3 and 20a, and amd ramps designs using tsmc 4nm and 3nm.

M2 pro vs M1 Pro/ max is a decent uplift but still “meh” compared to M1 Max vs intel skylake / comet lake. I’m going to skip m2 ad wait for m3 3 nm. But those upgrading to m2 pro should be happy.
 
For the M2 pro to beat M1 Max in both SC and MC after just 1 generation, its impressive
I wonder if M2 Max can approach the level of M1 Ultra, in SC yes, but in MC?
But i still suppose the gpu will be more powerful on the M1 Max compared to M2 Pro
M2 pro and m2 max have the same cpu core complex of up to 12 cores; 8 perf and 4 efficiency. M2 max has a better gpu than m2 pro.
 
  • Like
Reactions: 4odomi
Yep, you just need to a bolt a huge heatsink to it or an AIO watercooler and put all that inside a large chassis 😅
Not true a water cooler helps yes with raptor lake, but a 13600k does not need a water cooler a noctua fan cooler will tame that beast.

And the amd 7900 (non X) part is even more power efficient out of the box before you apply PBO, and it has a much lower tdp of only 65W. Runs cool under load. And is faster than m2 pro.
 
For the M2 pro to beat M1 Max in both SC and MC after just 1 generation, its impressive
I wonder if M2 Max can approach the level of M1 Ultra, in SC yes, but in MC?
But i still suppose the gpu will be more powerful on the M1 Max compared to M2 Pro
That's not a surprise. The CPU in the M1 Pro and Max are identical, as are the CPUs in the M2 Pro and Max, so it's no surprise that the M2 Pro beats the M1 Max in single core and multicore. M2 Max benchmarks will be about the same as the M2 Pro, give or take margin of error/slight differences when running the benchmarks. M2 Ultra is where we'll get the next multicore boost.
GPU is what separates the Pro and Max chips, not CPU performance.

M2 Pro GPU, while a hell of a lot better than the M1 Pro, doesn't quite hit the M1 Max GPU performance. M2 Max is likely to be an equally large gain over the M1 Max though, so that's quite exciting.
 
For the M2 pro to beat M1 Max in both SC and MC after just 1 generation, its impressive
I wonder if M2 Max can approach the level of M1 Ultra, in SC yes, but in MC?
But i still suppose the gpu will be more powerful on the M1 Max compared to M2 Pro
it's not really impressive considering the Pro and Max models share the exact same cpu
 
That is a VERY good score (usual disclaimer about synthetic benchmarks).

Here's the score my M1 Max MBP got in August last year (haven't tested recently). The new M2 Pro leaves it in the dust!

View attachment 2144580
Here's what I got on my M1 pro 2021. So not sure what Macrumors is basing their scores on ...

Screen Shot 2023-01-19 at 6.17.03 AM.png
 
If M2 Max scales the same way as last gen for the GPU in Metal that would be around 86k that is M1 Ultra territory.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.