M4 Pro Chip Benchmark Results Reveal an Extremely Impressive Performance Feat

This isn't true.

For example, AMD Threadripper, which is the high-end workstation brand, is currently two generations behind AMD's desktop CPUs.

Similarly, Intel often launches new microarchitectures on laptops first, desktops second, servers even later. For example, Intel Xeon Granite Rapids has started rolling out, and uses Redwood Cove as its microarchitecture. The consumer variant of that is Meteor Lake, which rolled out for laptops last year, and is already obsolete.

You’re right. Honestly I forgot about Threadripper because for a while there were rumors that AMD abandoned it, and I haven’t kept up with PC CPUs. Nvidia has released their flagship and upper tier cards first for the past few generations, though.
 
Last edited:
You’re right. Honestly I forgot about Threadripper because for a while there were rumors that AMD abandoned it,

Yeah, that might be. I don't keep up with AMD that closely.

But, Intel absolutely has had a pattern for a long time of staggering their CPU releases, usually mobile first, and very-high-end last. So Apple isn't unusual in that regard.

Nvidia has released their flagship and upper tier cards first for the past few generations, though.

You're right, but I think that's a different market segment. Nvidia can't compete in the mass market, because CPUs already include mid-range GPUs these days. So, unlike Apple and Intel, Nvidia pretty much do high end only, and I think that makes them not comparable. This was different in the nForce days.
 
There are several reasons:

-UI does not scale well to big screen, because design guys did not think about it..many don't even target iPads, so even if we wanted to do it, it does not provide good user experience, so let's not do it.

Yep. I find that iPadOnMac apps are… not a great user experience, to put it mildly.

-technical reasons: macOS as a development platform is not as good/nice as UIKit on iOS, and SwiftUI on macOS is not as good as on iOS.

I think this was discussing specifically the "take an iPad app and run it on an ARM Mac" thing, though. In that case, the development platform aspect isn't relevant; you're just developing an iPadOS app (in UIKit, SwiftUI, or a mix of those) and having someone run it on a Mac, unless you've specifically disallowed it.
 
Same reason why Apple keeps adding stuff to their chip with Max Ultra Pro Extreme suffix? People bought them so they can have the fastest Mac ever, for whatever reasons.

Problem is 4090 is NVidia's end game for 2 years now and kept its promise. Apple just kept bashing their last gen Pro Max Extreme chip like a cheap toy. I have no problem with Apple kept pushing the limit, but having an M4 Pro running circles around M3 Max in just a year? Apparently a "Max" level chip means nothing.
This is such weird take imo. "Max" simply means the best in that generation of chips. It makes no promises about how the next generation will be like. People buying the MBPs with the Max chip knows that they're paying a premium for the best performance they can get at that moment in time and believe that the money is well spent getting that extra performance early. Would you rather Apple held back the extra performance each year just so customers can brag that the M(n-1) Max is still faster than the M(n) Pro?
 
It's worth it if you need better GPU performance (it will be way better than on the Pro), or much more RAM.

It probably isn't worth it for the higher CPU score alone.
Yeah considering gpu usage are fairly limited on mac. Really only used for modeling work.
Video production uses the decoder/encoder/cpus.

And Mac doesn’t have much games, need for machine learning training and other use cases where upgrading gpus are more justifiable.

I do a lot of work on my mac but never found gpu to be the bottleneck. Even after few years down the road, the cpu usually gets saturated first.. my intel i9, m1 pro is getting slower and I can feel the lags. M3 max going strong. The radeon 5500m and the m1 pro gpu was fine
 
Wait, what?

I didn’t read the M4 being faster than the M3 Pro when the M4 iPad Pro launched. Are Mac’s M4 running at a faster clock than the ones in the iPad Pro? Do they have more cores?

The M4 being so much faster than what I initially thought would make me choose the M4 + 1TB + 32GB configuration instead of the M4 Pro + 512GB + 24GB of much faster RAM
I think it has to do with the cooling. Clock speed appears to be unchanged.

iPad Pro M4 with no special SoC cooling: 3805 / 14572
iPad Pro M4 with liquid nitrogen cooling: 3977 / 14905
Mac16,1 M4 (maybe iMac but not sure): 3864 / 15288

ie. There seems to be about a 5% multi-core CPU performance delta between the Mac and the iPad.

P.S. I think you should decide on RAM and storage amounts first, and then decide on the SoC. As you already know, t's generally not a good idea to skimp on the RAM or storage, whatever SoC you get.
 
This is a fair point. I think future-proofing the RAM amount is tricky.
Indeed it is. But IMO future proofing is bad semantics, because we are just talking about the life cycle of a box. That life cycle of a new box is indeed future, but it is future starting day one of purchase. Not some far distant fantasy future impossible to plan for.
 
Indeed it is. But IMO future proofing is bad semantics, because we are just talking about the life cycle of a box. That life cycle of a new box is indeed future, but it is future starting day one of purchase. Not some far distant fantasy future impossible to plan for.

Well, the question was about ten years. Who knows what we do on computers in ten years? When I got the M1 Pro, I had no idea LLMs would become a big thing.

And, generally, RAM use does go up. macOS 25 will use more RAM just idling than macOS 15 does, which uses way more RAM than Mac OS X 10.0 did. But how much more? Hard to say!
 
Hence, I said this is a bad time to get an M series chip. I don't want to go back to "that point" when people kept getting left out 100% slower just after a year. I'd just wait for things to go to "recent days of x86"?


Not for long.. apparently 14 + 20 M4 Pro is enough to outperforms that a year later. Max chip is just that, nothing.
But your computer does not become 100% slower….it will stay just as fast as it was when you bought it.

You seem to be very concerned with measuring your property against other people’s. This mindset is likely to lead to disappointment and dissatisfaction.

Try to view a computer as a tool, not some kind of precious object. In 10-15 years (or less) it will in the trash or hopefully recycled.

I’ve found that buying an older model (1-2 years) can save a lot of money and removes the FOMO of not having the latest and greatest, because you are starting out with an older model.
 
Last edited:
There are several reasons:

-UI does not scale well to big screen, because design guys did not think about it..many don't even target iPads, so even if we wanted to do it, it does not provide good user experience, so let's not do it.

-phone does have some sensors/HW mac don't. We can't do it, sorry.

-support of new platform is cost for developer, more testing, more customer support and we don't have a budget for it

-business reasons (any other reason considering $$$ and people making it)

-technical reasons: macOS as a development platform is not as good/nice as UIKit on iOS, and SwiftUI on macOS is not as good as on iOS. SwiftUI also does not scale very well to complex user interface, where you have much more UI elements on screen, meaning you see a lot of performance problems. This makes development harder, it takes longer to make, so it cost more, and on platform is much less people than on iOS, so it is not worth it. Many just failed to make it work properly.
And this one is completely Apple's fault. SwiftUI is not the silver bullet many of us (developers) think it would be. It is great on watchOS, good on iOS and tvOS, but not as good on macOS.
Excellent. Thank you.
 
I think it has to do with the cooling. Clock speed appears to be unchanged.

iPad Pro M4 with no special SoC cooling: 3805 / 14572
iPad Pro M4 with liquid nitrogen cooling: 3977 / 14905
Mac16,1 M4 (maybe iMac but not sure): 3864 / 15288

ie. There seems to be about a 5% multi-core CPU performance delta between the Mac and the iPad.

P.S. I think you should decide on RAM and storage amounts first, and then decide on the SoC. As you already know, t's generally not a good idea to skimp on the RAM or storage, whatever SoC you get.
I hope thats the 8core imac, cause ipad got 9 cores, and mac mini got 10 core

As the base
 
This is a fair point. I think future-proofing the RAM amount is tricky.
Within two hours of the announcement, I ordered a MM with M4 Pro, 12 Core, 24 GB RAM, and 1 TB SSD. After reading through this thread, this morning, I contacted Apple to change the configuration to the 14 Core M4 Mac Pro, not due to my current needs, but for future-proofing. The very nice Apple Guy suggested pretty strongly for my use case (MM will be new hub for my hobbyist music studio) that I'd be better off leaving the processor at the lower core count, but increasing RAM to 48 GB even though I don't need it now. I went through the ordering process, and was about to pull the trigger on the new configuration, but then backed out. I am torn about paying 25% more for a computer just for 24 GB of RAM. I also thought that, maybe I get the cheaper option now, and then, if in 5 years 24 GB or RAM doesn't cut it, maybe at that point, the M8 MM for the same price as today will be 3x faster and come with 128GB RAM???
 
Dunno about M4 Max, and we only have a single (but plausible) M4 Pro result, but I took ten measurements each from the AMD Ryzen 9 9950X (Granite Ridge¹, 16p), the Intel Core Ultra 9 285K (Arrow Lake, 8p16e), and the AMD Threadripper¹ PRO 7995WX (Storm Peak¹, 96p) and did an arithmetic mean (if someone can convince me why a geometric mean would be better, I'd actually be quite curious to hear it), and ended up with

Single-coreMulti-core% better (single)% better (multi)
AMD Ryzen 9 9950X34202206514.8%2.7%
Intel Core Ultra 285K32392244221.2%1.0%
AMD Threadripper PRO 7995WX29032794135.2%-18.9%
Apple M4 Pro392522669

So, the M4 Pro may not be a wrestling match contender since its name just isn't hardcore enough, but it beats all of them in single-core, and is comparable in multi-core, with the exception of the Threadripper Pro, which needs 96 p-cores to beat the M4 Pro's 10p4e setup.

(Granted, the Threadripper Pro is two generations behind.)

An M4 Max probably still wouldn't beat the Threadripper; even if we assume linear scaling (which we can't for multiple reasons²), the best-case score would be "only" 26,447.

¹ these are, somehow, real names
² including: 1) four cores are e-cores; we'd have to first separate those out 2) Geekbench 6, by design, penalizes high core counts 3) Apple's scheduler also isn't that great at scaling
Thank you for providing those benchmarks; there's a handful floating around in the Geekbench database, but they're both remarkably inconsistent and uncommon.

I agree that a reasonable guess for an M4 Max would be on the same general order of the 10% gain between the two M3 Max versions with the same core count, although there is now at least one result in, putting it at just over 26,000. Which is closer to 15%, a number I'm willing to believe based on it having double the memory bandwidth (Apple may also have improved multi-core efficiency).

So yes, it looks like it does come in a little below the Threadripper, although 26,000 is still an incredible result for a laptop CPU--that the M4 Max can even get close to the Threadripper with 1/6 as many cores and 1/6 the power draw is rather impressive.
 
Last edited:
Because it creates confusion as when should I buy a Mac? Imagine spending $4000 for the most expensive, fully maxxed out dream setup Mac and just a year later Apple can squeeze the same performance on a base line Mac Mini?
It’s not confusing. If Apple Silicon is advancing quickly, then the best strategy is to upgrade each year or two, rather than buy a fully maxxed out model that you plan to keep for 5 years.

I’ve never followed the fully maxed out, long term purchase strategy. I’ve always gone with the best CPU I could afford, and enough ram and SSD to last me a year or two.

“Unfortunately” I got stuck at the 10 core 14” M1Pro MacBook Pro with 8P cores and 2E cores. My DAW only uses P cores, so the equivalent M2Pro and M3Pro really weren’t much of an upgrade…especially the 12 core M3Pro being 6p6e.

Now that there is an M4Pro chip with 10P cores I’m interested in upgrading! If I do, it would be the 14core M4Pro Mac Mini with base ram and SSD. That will be a kick ass machine for a couple of years until we get an M series pro chip with 12P cores.
 
It’s not confusing. If Apple Silicon is advancing quickly, then the best strategy is to upgrade each year or two, rather than buy a fully maxxed out model that you plan to keep for 5 years.

I’ve never followed the fully maxed out, long term purchase strategy. I’ve always gone with the best CPU I could afford, and enough ram and SSD to last me a year or two.

“Unfortunately” I got stuck at the 10 core 14” M1Pro MacBook Pro with 8P cores and 2E cores. My DAW only uses P cores, so the equivalent M2Pro and M3Pro really weren’t much of an upgrade…especially the 12 core M3Pro being 6p6e.

Now that there is an M4Pro chip with 10P cores I’m interested in upgrading! If I do, it would be the 14core M4Pro Mac Mini with base ram and SSD. That will be a kick ass machine for a couple of years until we get an M series pro chip with 12P cores.

That strategy seems to work for this kind of situation. Go with lower or mid tier and upgrade more often instead of buy the higher end and keep it for longer.

Yes I'm still on my PC and I'm just so used with latter move. I've kept it for 4-5 years with little to no upgrades, because it reaches to a point where generational boost doesn't seem worth it so I can keep my PC longer and the higher spec compensates any offsets of new, faster hardware.

I'm sorry if this "Mac situation" seems a bit strange for me.
 
Yes its pretty much a full computer in a little box.

M4 max, which is not in the mini, will be pretty good for gaming on a mac. But 'on a mac' is a massive caveat because emulating or simulating windows /x86 environments and frameworks (e.g. DirectX) takes a massive performance hit. So yeah, some older gaming, specific games that get native ports (i.e. Civ 7), maybe WOW you could have a great time on the mac. If you are expecting to be on Steam and enjoying a large number of games you will be very disappointed.

For the time being, if you dont need the GPUs otherwise, you'd be much better off investing the cash difference from an m4/m4 pro down from the max into a dedicated gaming machine. Console or steam deck would probably give you best bang for your buck.
What’s in the mini?
 
M4 Max Geekbench scores are even more impressive



M4 Max CPU.png




M4 Max GPU.png
 
It’s not confusing. If Apple Silicon is advancing quickly, then the best strategy is to upgrade each year or two, rather than buy a fully maxxed out model that you plan to keep for 5 years.

I’ve never followed the fully maxed out, long term purchase strategy. I’ve always gone with the best CPU I could afford, and enough ram and SSD to last me a year or two.

“Unfortunately” I got stuck at the 10 core 14” M1Pro MacBook Pro with 8P cores and 2E cores. My DAW only uses P cores, so the equivalent M2Pro and M3Pro really weren’t much of an upgrade…especially the 12 core M3Pro being 6p6e.

Now that there is an M4Pro chip with 10P cores I’m interested in upgrading! If I do, it would be the 14core M4Pro Mac Mini with base ram and SSD. That will be a kick ass machine for a couple of years until we get an M series pro chip with 12P cores.
Believe the maxed out strategy worked rather well in the Intel era, cpu/gpu/ram/hdd/ssd wise, showstopper would be transfer rate improvements for a subset of users.

Apple M-series kind of turned it all around, the only bummer in my opinion was USB A exit and the downgrade in ssd bandwidth from M1 to M2 for budget macs.

With M-series, Apple are capable of changing processor configurations with serious impact upon performance, and "get what you need for a planned shorter timespan" makes a lot more sense. Thunderbolt 4 or 5 would be a lower threshold to consider. Thunderbolt 5 will be here for a while, perhaps useful to consider for cable choices....
 
Within two hours of the announcement, I ordered a MM with M4 Pro, 12 Core, 24 GB RAM, and 1 TB SSD. After reading through this thread, this morning, I contacted Apple to change the configuration to the 14 Core M4 Mac Pro, not due to my current needs, but for future-proofing. The very nice Apple Guy suggested pretty strongly for my use case (MM will be new hub for my hobbyist music studio) that I'd be better off leaving the processor at the lower core count, but increasing RAM to 48 GB even though I don't need it now. I went through the ordering process, and was about to pull the trigger on the new configuration, but then backed out. I am torn about paying 25% more for a computer just for 24 GB of RAM. I also thought that, maybe I get the cheaper option now, and then, if in 5 years 24 GB or RAM doesn't cut it, maybe at that point, the M8 MM for the same price as today will be 3x faster and come with 128GB RAM???
The Apple guy is right, especially for audio production, more RAM is a better pay off. It'll handle more tracks at the same time with ease over hitting the bottleneck and using swap which, while fast on Apple Silicon machines, still isn’t as fast as straight RAM. For audio production any Apple Silicon CPU is going to be quick and not much of a noticeable improvement over a new chip while RAM is still king 👑.
 
Yeah, the Apple Guy (a Genius?) is right. However, the price jump between the 24GB and 48GB M4 Pro Mini is almost 500€. It’s insane. There needs to be an intermediate step of, say, 36GB. That would be my optimal setup.

But staying at 24GB or bumping it to 48GB for almost 500€ is just too much.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Back
Top