Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
With the Max moving to 12 performance cores they could have left the Pro with 8 performance cores and then the Pro would have been a perfect midpoint between standard and max.

Can you acknowledge that maybe, just maybe, the machine will work better for a broad range of users with more E-Cores. Is it possible that for most computing tasks, especially on mobile, having more E-Cores can accomplish the same tasks with far greater power efficiency while having an imperceptible difference in day to day perceived performance.

Apple employs some of the brightest silicon engineers on the planet, lets give them a little credit, that they do actually know better and therefore their designs are actually better for most users.
 
The thing is, no one was really complaining about battery life in M1 Pro or M2 Pro. I suspect people will now come out of the woodwork to praise this decision and talk about how poor their battery life is

Mine (M1 Pro) isn’t poor, but it isn’t fantastic either. The e-cores always seem fairly busy. Making them faster and more plentiful makes sense to me.

Further, if battery life was such an issue they would not have made the m3 max the way they did.

Well, no, this gives people more choice. Max out performance, or balance it.

Realistically, the M1 and M2 Pros were chopped versions of the Maxes. This time, they’re a separate design.
 
  • Like
Reactions: smulji and ender78
Why does the m3 max exist in the configuration it does then? Why isn’t it a 8p + 8e core chip?

Because for many people on laptops, all day battery life trumps a marginal performance increase.

We use Lenovo Thinkpads at work, on those machines with 16GB or RAM and 512 GB SSDs I see people that can barely get a few hours of battery life let alone 22. You don't need P-Cores for a Teams Call or to edit a document in Word or Excel.

You have an opinion that any solution that does not increase P-Core count is useless. Apple's engineers disagree.
 
  • Like
Reactions: tekksan
nope.
You do not ask Apple to make an M3 as a straightforward next step from M1 Max just because it is new gen. Why ? Because you do not identify M1 max user as the target for M3.
Same here. M2 pro users are not the target. It does not have to be straightforward for them.

So, why make the new product ?
M3 and M3 max seems to be more substantial updates. I suppose it is simpler for Apple to move the whole chip family instead of keeping the M2 pro until M4 pro is ready.

Is it a downgrade ? Does people who are looking for a new computer in this performance class have to rush on the remaining M2 pro products ?
It does not seems so.

People does not complain on M2 pro perf, nor battery life. When there is no problem there is no solution. Then comes the complains that apple bring no solution to the non existent problems.
It is about the fact that the pro was repositioned for the same money: we complain because they decontented the chip relative to the previous pro.

The value proposition of the pro is worse than it was in previous generations. I know many will defend Apple regladles of how they change the value proposition but the changes this generation are not praise worthy …
 
Can you acknowledge that maybe, just maybe, the machine will work better for a broad range of users with more E-Cores. Is it possible that for most computing tasks, especially on mobile, having more E-Cores can accomplish the same tasks with far greater power efficiency while having an imperceptible difference in day to day perceived performance.

Apple employs some of the brightest silicon engineers on the planet, let’s give them a little credit, that they do actually know better and therefore their designs are actually better for most users.
It is not enough of an upgrade relative to the standard m3 which I suspect is the point. It makes people who need more than the standard m3 more likely to look at the max than they used to because the standard pro is now a worse value proposition.
 
Apple most certainly does not sell every one they make at retail. If that were the case, there wouldn't be tons of deals on the M1 Pro MBP14 and 16 every other day for $1000+ off.

The retailers choose to discount them, does not mean that Apple is taking a hit.
 
Because for many people on laptops, all day battery life trumps a marginal performance increase.

We use Lenovo Thinkpads at work, on those machines with 16GB or RAM and 512 GB SSDs I see people that can barely get a few hours of battery life let alone 22. You don't need P-Cores for a Teams Call or to edit a document in Word or Excel.

You have an opinion that any solution that does not increase P-Core count is useless. Apple's engineers disagree.
So again, to be clear, from a performance perspective you are defending a decrease in value compared to last generation.

The m2 to m2 pro was a fantastic value proposition, m3 to m3 pro is worse.

Battery life on the m2 pro wasn’t in need of improvement, this looks like nothing more than a cynical move to increase margins. Balance is in the eye of the beholder and I would have rather seen the m3 pro become the mid point between m3 and m3 max rather than becoming something only marginally better in CPU than the standard m3
 
Then why M3 Pro have the same battery life as M3 Max?
Maybe the quoted battery life specs are for workloads that only use the 6 or 4 efficiency cores (video playback, wireless web). So you wouldn’t expect as much difference as there would be under full load?
 
Hmm, brings back memories of Jobs bragging "And it eats Pentiums for breakfast!" Sounds like Tim Cook has fallen down the same historic Apple processor rabbithole with diminishing returns. Not good. Now we know another reason why Cook isn't doing any "Bake-Offs". I hope I'm wrong, ,but I'm feeling deju vu here people.
 
  • Like
Reactions: bcortens
Battery life on the m2 pro wasn’t in need of improvement, this looks like nothing more than a cynical move to increase margins. Balance is in the eye of the beholder and I would have rather seen the m3 pro become the mid point between m3 and m3 max rather than becoming something only marginally better in CPU than the standard m3
How is 6-14% a "marginal" increase? Just because some guy added "barely" to his clickbait title?
Intel will gladly sell you 6% performance increase for 15% price increase, you can always move.
 
  • Like
Reactions: Axelle838
So again, to be clear, from a performance perspective you are defending a decrease in value compared to last generation.

The m2 to m2 pro was a fantastic value proposition, m3 to m3 pro is worse.
It’s just different. I’m struggling to see why you think the value is so much worse. The performance difference will vary based on the architecture. If this proves to be bad for sales, or real world performance Apple will adjust.
 
Average M2 Pro score is around
2675 / 13000. That’s about a 20/12% improvement for the M3 Pro.

In fact, the M3 Pro is faster in ST and slightly faster in MT than the M2 Max.

What does that tell you? That the Max model is no longer a pure GPU upgrade as it used to be. Like it or not.
This is also clearly visible from the die shots (Pro no longer a cut down Max model).

Hard to believe the intellectual level on this forum to be honest.

It wouldn’t suprise me at all if the Studio from now on won’t get the Ultra version, but only the Max. Ultra reserved for Mac Pro. Max reserved for Studio. Mac Mini gets base or, maybe, even Pro.
The mini will definitely get the Pro chip. But the Pro chip is just going to be a better version of the base chip and will largely be consumer and casual level chips. It won't be either a CPU or GPU powerhouse. The Studio is there for those chips. The Studio is pretty large and these chips run pretty cool, so I don't see much reason why Apple can't keep putting an Ultra in the Studio. If/when Apple makes a chip that needs more cooling than can be fit in a Studio, then it will only go in the Mac Pro. But doubling up a Max, which goes in a laptop and runs off a battery, is unlikely to stress the thermal capabilities of the Studio. So as long as an Ultra is basically just two Max chips, I don't see a reason to not have it in the Studio.
 
  • Like
Reactions: Chuckeee
Could it be better, sure, is more than 8GB/256 required for some users, absolutely not. The fact that there is a broad sector of the market where that base configuration is more than what most people need is a realization not an excuse that that config makes sense. No one is forced to buy the base config. Why should those users that don't need more, those that will NEVER use more than 8GB or RAM be forced to buy a machine with a higher base than what they need.
I think the biggest issue with storage is that Apple can jump from 256 to 512 for pennies, but they hold it back to nickel and dime us. For a device anywhere near the price of MacBook Pros they should come with 512 or 1tb as base, even if a lot of people don't completely fill 256.
 
How is 6-14% a "marginal" increase? Just because some guy added "barely" to his clickbait title?
The M3 Pro is essentially tied in code compilation performance relative to M2 Pro, this is what I care about so seeing stagnation here bothers me. I am also bothered that they are selling a smaller chip for the same money as the old one.
Intel will gladly sell you 6% performance increase for 15% price increase, you can always move.
As I said, two things can be true, the m3 pro can be a great chip, but it can also be a worse value than the m2 pro was.
 
  • Like
Reactions: ric22 and Chuckeee
It’s just different. I’m struggling to see why you think the value is so much worse. The performance difference will vary based on the architecture. If this proves to be bad for sales, or real world performance Apple will adjust.
Performance per dollar the m2 pro was a more substantial upgrade relative to m2 than m3 pro is from m3 - that is the change. Those numbers are just fact. You can say the tradeoff doesn’t matter, you can claim more sustained battery life is better choice. But relative position in the lineup is worse than M2 Pro.


Of course they are going to sell well, the Max is hugely more expensive and people don’t have a choice to buy a mythical 8+4 CPU + 20 GPU m2 pro.
 
The M3 Pro is essentially tied in code compilation performance relative to M2 Pro, this is what I care about so seeing stagnation here bothers me. I am also bothered that they are selling a smaller chip for the same money as the old one.
So far we have no idea about performance of E-cores and if it will affect build times. Get Max if you are worried, programming is one of the high-revenue professions anyway.

It was obvious from the start Pro and Max being the same CPU is a huge marketing mistake, and they corrected it. Days of maximum CPU performance on cheaper builds are over.
 
I would never need 512 Gb unless I change occupation.

I have 256Gb and have 143 Gb free.


With cloud services you don't all or the full version of your data on your computer.
It really is almost only media files that requires a lot of space, as a regular office worker I was hard pressed to create more than 10-15GB per year with .doc, .xls, .ppt and such… but I guess on MR that means those people are not “Pros”
 
This seems late to me, in terms of the 13" MBA, which is their best selling laptop by far, but totally plausible.
You may be right, but June 2024 will be 1 year for the 15" MBA and 2 years for the 13" MBA between updates. In the meantime, the 13" price will probably continue to drift down to the magical $999, which has been par for MBA models after a few years in productions. It also gives some time for M3 production to ramp up.

If you look at the MBA user base of the most popular configuration (probably the base model), I really don't believe many of these consumers would ever notice the difference between the M2 and M3 for their day to day use......That is assuming they knew M2 and M3 chip existed in the first place.
 
  • Like
Reactions: Chuckeee
It really is almost only media files that requires a lot of space, as a regular office worker I was hard pressed to create more than 10-15GB per year with .doc, .xls, .ppt and such… but I guess on MR that means those people are not “Pros”
Couple backups of your iPhones and you're done on 256Gb
 
  • Like
Reactions: Someyoungguy
So far we have no idea about performance of E-cores and if it will affect build times. Get Max if you are worried, programming is one of the high-revenue professions anyway.

It was obvious from the start Pro and Max being the same CPU is a huge marketing mistake, and they corrected it. Days of maximum CPU performance on cheaper builds are over.
They didn’t have to cut down the pro to differentiate it from the max because they made the max so much better.

We already can look at the GB6 Clang sub-test which shows tiny to non-existent improvements depending on which m2 pro you compare with.
I get it, you’re okay with moving the pro down the stack, that’s fine, but that makes it a worse value.
 
This whole line up screams rushed
I don't see why they would feel the need to. The machines being replaced are less than a year old (save the iMac that was 2 generations behind and there's no reason why they couldn't have released an M2 iMac). They could have held off until January without raising complaints, I'm sure. The aforementioned previous machines came out in January after all. Maybe they were aiming for the Christmas rush? But if you're not going for a stock build then you'll probably have to wait until after Christmas to get it anyway, so...
 
  • Like
Reactions: bcortens
Couple backups of your iPhones and you're done on 256Gb
I would never backup a personal iPhone on a work computer ( remember I said “office worker”) …
And even on my personal Mac I use close to 4TB with media files but less than 50GB in other files, in well over 20 years
 
No - Intel's problem is that they have been "too big to fail" and able to succeed through politics rather than innovation ever since IBM "made" them by picking their kludgey stopgap pseudo-16/32-bit 8086 for the original IBM PC. (This was paired with Microsoft's kludgy CP/M knock-off operating system to make a deeply mediocre machine who's only notable feature was the three magic letters on the front and an army of pin-suited salespersons with an existing office equipment and mainframe empire). Sure, Intel arguably got the whole ball rolling in the 1970s with the 8080 - but by the 80s they'd lost ground to the Zilog Z80 and even the 6502 and several "proper" 32 bit chips were in the pipeline. Intel and Microsoft managed to lock the industry into 1970s CPU architecture for a couple of decades - when we could have had 68k, a clean 32 bit instruction set, a Unix-based OS and high-level source compatibility

Even the current x86-64 instruction set was designed by AMD, not Intel.

The first ARM in 1987 wiped the floor with the contemporary Intel 80286 performance-wise - but it didn't run Windows (except under emulation), so it went nowhere.

Intel's Unique Selling Point is legacy compatibility, and this is now being undermined by a perfect storm of industry changes: There's Intel's almost total failure in the emerging mobile market, which is now dominated by ARM. The rise of Linux and web-based tech with its roots in Unix/Linux - and its culture of hardware-agnostic software design and open source (there's some advantage to running Linux on x86 but most of the key Unix/Linux software packages have supported ARM and other architectures since forever) - is giving Intel serious competition in the server space where power consumption is also a big issue. Apple - even if they remain the underdog in the personal computer market - have demonstrated to the world that you can make high-performance laptops and low-end desktop workstations without x86, and Microsoft seem to be having a serious attempt at Windows on ARM this time. Then, in general as computers have got more powerful, more and more software is written in portable, high-level code and uses operating system frameworks for things like graphics, multithreading, vector processing and neural networks, rather than direct hardware access. An increasing amount of stuff is written in scripting languages like Python or Javascript Even modern Windows ships applications as Common Language Runtime bytecode rather than x86 or ARM binaries, Android apps typically ship as bytecode and even the Apple App store has the capability to distribute "compile on delivery" bytecode. Apple have also demonstrated - with Rosetta 2 - how effectively x86 binaries can be translated to ARM once you've kicked out all of the legacy stuff. I think Windows on ARM's main problem is with bits of legacy Win 16, Win 32 etc. binaries still floating around.

Put simply, Intel's business model relies on people needing to run legacy code, and people ain't writing legacy code any more - so the clock is ticking.

The one area where Apple Silicon has failed to convince is in the high-end personal workstation market - the new Mac Pro satisfies a very limited niche of people who need high PCIe bandwidth but not discrete GPUs. If you really needed the 2019 Mac Pro an x86 tower is probably still the tool for the job. Trouble for Intel there is that the only people buying Intel are the dogmatic "workstations need Xeon because they just do" crowd and AMD are slaughtering them on price/performance. Plus, the whole idea of a high-powered personal workstation isn't for the ages, as the industry shifts to cloud computing (and NVIDIA have some nice ARM-based datacentre-grade iron such as Grace/Hopper to show you).

Whatever you do with x86 tech, it is always going to be carrying around the extra weight of supporting the complex x86 instruction set(s) that more modern RISC-like ISAs don't need. Intel's future is probably to make ARM, RISC-V or some other ISA - and rely on Rosetta-esque translation for x86 support. Trouble is, then, they're going to have to compete with multiple competitors on technical merit, which they haven't needed to do for 40 years.

While a part of this is down to Apple shaking up the industry from time to time, it certainly isn't about what Apple released recently - it goes back as far as the Newton (when they invested money in ARM) plus of course the iPhone's role in promoting the modern mobile scene.
My only quibble with your post is it was the 8088 and its 8/16 bit architecture (not 16/32) that made the original IBM PC. The 68000 was 16/32 and was my favorite line (Amiga 68030 was my last computer before having to “grow up” and use what the job used). I’m old.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.