Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Was kinda hoping single core would break 3200… I’m on the fence about upgrading from my M1 Pro to an M3 Pro or Max.
Sounds like: Wait 18 months and get the M4. My take is: If you are happy but want the new thing, wait. If you are struggling with what you have, buy. This will make you much happier about each upgrade, because it will actually feel like a difference.
 
Yep and by time m4 max is ready, these same laptops will be on discount for $600 or more off.
At some point it is very likely the Studio and Pro will pull away from the laptops. They just have so much thermal potential. I have a hard time believing Apple didn’t design them for a lot more than they currently have available.
 
  • Like
Reactions: falkon-engine
Apple has a good leverage here with Microsoft. Port all your games to Mac with full native support and we will make Bing the default in Europe on iPadOS.

Make native versions of Microsoft Project, Visio and Access and we will make Google optional on iOS in Europe.
 
how do you fit desktop into backpack and keep it under 3.5 pounds?
you will be able to buy similar Intel based machines with 14 series processors and Nvidia chipsets. When it comes to getting real computer work done that requires speed.. no one cares about colours and size and power draw.
 
  • Haha
Reactions: polaris20
I would be more interested to find out how close the TDP of the M3 has returned to M1 level values (13.8W). The extra six watts of TDP for the M2 was apparently a bridge too far for Apple to properly handle in the MacBook Air thermal design.
 
How did Apple neglect to mention this as well:
 

Attachments

  • unnamed-1.png
    unnamed-1.png
    182 KB · Views: 171
All I know right now is that they are not "Pro" for the outrageous prices they are charging. Generationally they are still far behind. When it comes to render power, you can buy a fully decked out Desktop PC with a soon to be released 14900K and 5090 GPU. That system will be expandable, serviceable and upgradeable. Not to mention running faster PCIe Lanes. The M3 is still PCIe4! The 5090 will dance circles, multi-fold, around whatever boost to the GPU apple has added to the M3. Apple has really painted themselves into a corner with moving to a phone SOC for their laptops/desktops. So sad. Someone should be fired for making the choice to do that.

Interesting that in your screenshot you compare the M3(Base) to an I9-14900K. Based on another post here, the M3 MAX which is a much better comparison has a score of nearly 22000 (which is still lower that the 26K from the i9). Big difference is that you can use the MacBook pro unplugged and consumes a fraction of the power. Do you even know that the 5090 will be released ever ?

Apple is acting as a system integrator and is picking components that are GUARANTEED to work together. Try the same with your custom built PC. Oh no, did you not know that you cannot use this cooler with this case, or that this memory will only work at 80% of its Max in this motherboard with this video card. Some people love this compatibility game, you have the right to purchase what you like. Mac buyers are happy to pay the "Apple Tax" to not have to put up with that bull.
 
getting 20% better performance by increasing clock speed by around 17% is to be expected but that's what has some people worried. That Apple is basically relying on TSMC's process technology to increase performance by as opposed to to improving performance via IPC. You can't keep increasing clock speed indefinitely without consuming more power.
I expect that M4 will bring about the same level of gains at about the same clock speed (maybe a small increase), meaning that they will probably introduce significantly improved cores. I think they focused a lot on new GPU architecture this time, and let the process technology (and my association, clock speed) bring CPU gains.
 
  • Love
Reactions: smulji
If my research is correct, the M2 Max mac studio was released on June 13... only to be obsoleted less than 4.5 months later by M3 max. If people who bought the m2 max mac studio had known m3 max was coming so soon, perhaps some would have delayed their purchase, no?


I don't want to ruin the surprise, but there will be an M4 Max too.
 
That's because there is little to rave about for this update.
Apple themselves during the keynote advertised it as an upgrade for mostly Intel users and maybe some M1s, they dodged most comparisons against the M2 line.
I don't see how you can say that. I am a self acknowledged Apple hater. Yet even I have to admit that a 20% performance increase, in just a year and a half... when they were already pushing right against very edge of the physical laws of the universe with the M1... that is IMPRESSIVE.
 
At some point it is very likely the Studio and Pro will pull away from the laptops. They just have so much thermal potential. I have a hard time believing Apple didn’t design them for a lot more than they currently have available.
This! There is so much thermal headroom in the studio chassis, and even the mac mini chassis. I'm surprised Apple doesn't push them further.
 
So let me get this straight, a laptop with m3 max will be faster than a mac studio with m2 max? Seems like anyone who bought a mac studio m2 max in the past 1-2 months got taken for a ride.
Are you suggesting that Apple nerf newer products so previous customers don't feel gypped?

I have no problem with new products coming out that smoke my current ones. Apple shouldn't hold itself back, ever.
 
Of course, another way to go would be to put iMac guts on a card inside an AIO screen. When the guts get too old, replace the card like we used to be able to replace/upgrade RAM in iMac.
  • Open panel/remove back,
  • remove old iMac card,
  • insert new iMac card,
  • replace panel/back,
  • enjoy up-to-date iMac.
Bonus: maybe engineer that so that an iMac without a card is ASD 2 monitor. Then one product frame + screen + camera + speakers could cover BOTH bases.

Yes, I know that is extraordinarily unlikely from Apple, but it would be a mother-nature-friendly way to NOT doom AIO and overcome the "screen outlasts the tech guts" issue.

Personally, with long-term love for iMacs I used for well over a decade, there's NO way I would ever buy another without overcoming the issue of screen life vs. tech guts life. Go card-based approach, resurrect TDM functionality, etc or bust IMO. And I voted with my wallet on this topic by embracing separates and NOT even buying the screen from Apple.
Name three AIO's that you can do this in on the market today, or name one.

Different logic board, different port configuration, different cooling needs, a modular AIO is a dream.
 
  • Like
Reactions: ignatius345
The medians are 3018 / 11671. That's up 14.7% / 19.8% from the Mac mini's M2.

Geekbench lists the clock as 4.05 GHz, but I wouldn't fully trust that yet; it sometimes merely guesses what the clock is. But if true, then the clock is up 16.4%, so single-threaded IPC is actually 1.5% worse than before (which is surprising, given that memory bandwidth on the non-Pro hasn't changed), and multi-threaded IPC is only up 2.9%.

Which… for a two-generation jump doesn't seem so great? But again, it's possible that Geekbench is wrong about the clock. Or that these results don't reflect final shipping machines. Or that it is not, in fact, a two-generation jump; that these are actually A16-derived cores (but the GPU raytracing feature would suggest otherwise).

Of course, ignoring clock changes, a 15%/20% improvement in 18 months is pretty good. The clock steadily increasing just worries me because, well, they're not gonna be running the M7 at 6 GHz…
Remember what happened with the first A17 benchmarks...

EVERY DAMN TIME what happens with these early release benchmarks is that they are executed on a machine that was plugged in 20 seconds earlier and is doing all its "first power-on" tasks at the same time as the benchmark.

Expect this to be, likewise, followed by breathless claims of power use that, once again, evaporate when the same measurements are performed by responsible adults under appropriate conditions, not in a desperate attempt to be FRIST!!!!
 
Interesting that in your screenshot you compare the M3(Base) to an I9-14900K. Based on another post here, the M3 MAX which is a much better comparison has a score of nearly 22000 (which is still lower that the 26K from the i9). Big difference is that you can use the MacBook pro unplugged and consumes a fraction of the power. Do you even know that the 5090 will be released ever ?

Apple is acting as a system integrator and is picking components that are GUARANTEED to work together. Try the same with your custom built PC. Oh no, did you not know that you cannot use this cooler with this case, or that this memory will only work at 80% of its Max in this motherboard with this video card. Some people love this compatibility game, you have the right to purchase what you like. Mac buyers are happy to pay the "Apple Tax" to not have to put up with that bull.
Because there are definitely no PC systems integrators out there.
 
  • Like
Reactions: canon-cinema-0r
If my research is correct, the M2 Max mac studio was released on June 13... only to be obsoleted less than 4.5 months later by M3 max. If people who bought the m2 max mac studio had known m3 max was coming so soon, perhaps some would have delayed their purchase, no?
Apple sells PRODUCTS not chips.
Just FYI you cannot buy an M3 Max studio right now, and who knows when you will be able to...
 
Gotcha. I think this article might clarify exactly what we're talking about... Potentially still A16 CPU (Everest/Sawtooth) but with A17 GPU because that GPU was originally designed to go with the A16.
Correct. And this also explains why the M3 doesn't have as many neural cores as the A17pro. Touting ML where the models can fit into RAM (which is great) while the neural engine is 1/2 the performance of your latest phone is ...interesting marketing to say the least.

M3 Ultra, if they make interesting packaging changes (like stacked dies for example) or M4 Max will be when I jump-in to Apple Silicon, probably.

Edit: I wonder if they added support for 64-bit atomics finally, this is in the a17pro.

Watch this link for updates: https://developer.apple.com/metal/Metal-Feature-Set-Tables.pdf it's a pretty important feature. The tables haven't been revised to show M3 data yet.
 
Last edited:
Because there are definitely no PC systems integrators out there.
There are and some charge a very significant premium for that build. Desktop PCs are modular, Macs are not. Most people don't care about machine upgradeability, they buy a new machine when required. The SOC that Apple has developed has benefits and trade offs, vote with your wallet.
 
IMO, we’ve gotten to a point where no one really NEEDS to upgrade if they already bought into Apple Silicon. Buying a Mac has become like buying an iPhone: you can wait and buy it when you need it and not sooner. My M1 machine is still fantastic and I feel no need to upgrade. It will be at least a couple of more years before I buy another Mac.
I think that explains why Apple made many comparisons to Intel Macs.
 
  • Like
Reactions: Orange Bat
Point being?
The M3 is not worth the price for a "Pro". I am seriously not understanding how people are accepting the marketing hype of these keynotes when the real world benchmarks are showing it is not an impressive chip for the price for Real Pro's. 8GB of ram in a pro machine with reduced memory bandwidth. This is "Pro"? In what world are we living. The Pro moniker needs to start with 32GB Ram and 1TB drive space. Period. These low end Pro machines are being created to allow people to run around pretending they are in some social class. Apple is all about social status and maximizing profits now... It's just gross.
 
  • Disagree
Reactions: AlphaCentauri
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.