I mean, that is the whole point. They know 8GB is the bare minimum and they're giving us that. Very similar to how iPhones were stuck for ages with 16GB that let you install 2 apps and get a song or two on there. Particularly maddening is the fact that it's a device for Pros with a Pro price. It's absurd, but that's Apple. They have always been like this and as long as people buy the upsell instead of switching brands they'll keep doing this, as it's been a successful business model for them selling hardware that's obsolete the day it hits the shelves.
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.
Apple Announces New 14-Inch and 16-Inch MacBook Pro Models With M3 Series Chips
- Thread starter MacRumors
- Start date
- Sort by reaction score
Rest in pieces. Won't miss that questionably useful feature that no one asked for.
Status symbol? I go so far as to always get whichever colorway is shared with old models (so I never got a space grey machine) to minimize it standing out.
The M2 Max was the first time I was able to afford the top chip upgrade in a MacBook; it was a significant investment for me. No, it’s not less capable than it was on Sunday, but the fact that it’s turned out being a stopgap chip does mean I’ll be able to keep using it for less time.
Nope, Apple believes there’s a market for people that want the chassis of the MacBook Pro but with the power of the air and not pay anything more. That is this model, and they have a choice to update the ram if they want to. Is it for connectivity, the screen, active cooling, it’s a choice. I would have considered it once.
all those people that can’t think outside their own mindset..bleh
also those “if I max it out, it will last longer“ people, it’s not true. Apple will change the architecture of their chips every 4-5 years. Probably the next will be a quantum engine and then there you are with your neural engine, opening some browser tabs, excel and word or pages and your fans blowing.
I saw a statistic in the media (can’t remember exactly where) yesterday that said fewer than half of Mac owners have upgraded to Apple Silicon. That’s why Apple was pushing so hard on the performance differences between Intel and Apple Silicon. I’d imagine their verbal presentation also emphasized M1 verus M3 for the same reason instead of M2. M2 owners are the least likely to upgrade, so while Apple did give us the M2 comparisons in their charts, they didn’t emphasize it.
Back in the olden days (prior to 2000), CPU manufacturers and programmers emphasized single cores, enhancing performance through the process of making a chip by shrinking die sizes and by pushing higher and higher clock speeds. Once clock speeds hit somewhere over 3GHz, that became impractical due to heat problems, so CPU manufacturers started increasing the number of cores. Software can run many tasks simultaneously, and each core can handle a different task at the same time, hence the term multitasking. Since clock speeds could not easily be increased, core counts multiplied. The more cores there are, the more tasks can be processed at the same time, but there are limitations to that since some processes must wait for others to complete, so there isn’t a one-to-one improvement in additional cores, i.e. two cores aren’t necessarily twice as fast as one. They can be, but they often are not.
The same goes for GPU cores where tasks do tend to be more isolated, so there is closer to a one-to-one increase in performance per GPU core. For instance 40 GPU cores will be close to twice as fast as 20 GPU cores, say maybe 90%, but 16 CPU cores may only be 75% faster than 8 CPU cores.
Then there’s the issue of performance cores versus efficiency cores. While Apple didn’t pioneer this, they were among the first to use the concept for the wide consumer audience. Efficiency cores (aka e-cores) handle simple tasks well, like web browsing, text editing, etc. while their main emphasis is on conserving power over performance. Performance cores (aka p-cores) take advantage of additional power and are ideal for performing more complex tasks. The combination of p-cores and e-cores helped Apple Silicon handle the balance between high performance and great battery life. The OS looks at the various cores available and the tasks at hand and determines which core gets which job. A well written task manager will do well to make the computer blazingly fast while at the same time maximizing battery life.
So when looking at the various offerings this time around, you can assume that the more cores an SoC (system on a chip) has, the faster it will be, but you have to look at the number of p-cores versus e-cores. For instance, the M3 has four of each, so it is a good balance between power efficiency and high performance, but won’t be as fast as its Pro and Max cousins due to lower core count and higher percentage of e-cores. The Pro has 6 e-cores in all configurations, but only 5 or 6 p-cores. That means the M3 Pro also maintains a balance, emphasizing neither speed nor battery life. It does have more total cores, so it’ll be faster than an M3, CPU-wise. Graphics-wise, the M3 Pro can have upwards of 18 GPU cores while the M3 has only 10. Since graphics scales better, the M3 Pro is not quite 80% faster.
The M3 Max, with only 4 e-cores and upwards of 12 p-cores, is built for speed, sacrificing some battery life, though it seems the OS does a pretty good job on power savings. With more cores and a higher percentage of p-cores, the M3 Max is by far the fastest of the three. GPU-wise, the M3 Max has 30 or 40 cores, with the higher version having more than twice the graphics performance of the M3 Pro, which maxes out at 18.
While core count is a good indicator for comparisons within the same family, it’s harder to compare to previous generations, since Apple makes improvements to each generation by optimizing its chips better. So a p-core, e-core, and GPU core on an M3 are going to be individually faster than the equivalent on an M2. How fast, we have to look at what Apple says and how reviewers run tests to verify Apple’s accuracy.
This is quite long, so I’ll stop here, but hopefully this gives you a better idea of all the numbers being thrown around.
Yes this is the worst fallacy perpetuated on these forums. I bought new 2017 12” retina MacBook and 2018 MacBook Pro with 16GB RAM with the “future proof” mentality. Three year later they were both eviscerated by my 8GB M1 Air at half the price.
@tobybrut, you're awesome! Thanks for that quick rundown of how to think about cores. No, it was not information that I knew at all even though I have heard of the term "efficiency core" throw around! Much appreciated!
For the kinds of things I do -- hardly ever any video editing -- which is mostly writing, reading, doing academic research and databases (DEVONthink, Scrivener, OneNote, Obsidian), and simple photo editing none of it requires super duper heavy lifting. I was thinking of shying away from the M3 though largely just because of the typical advice of not to get the lowest level, but a level up. But that said, I don't know if M3 Pro makes sense for me. Power efficiency is definitely important to me and maybe a little bit of oompf for power, but the M3 might just be sufficient. But I wouldn't want the fact that if I chose M3 over M3Pro that I miss out on some future macOS feature that makes use of something about the M3 Pro or M3 Max that could be a fun side thing.
Thanks, I’m glad it helped. From the Apple presentation, I don’t think you’d miss out on features since all the chips in the family use the same cores, same neural engine, and same video codecs. It seems to me, unless Apple didn’t cover it, that the only real differences are speed and efficiency. The neural engines are the same on all three chips, along with the video codecs, so you won’t lose out on any features that may appear in the future by choosing a different model within the same family. Apple’s OS updates or feature limitations are usually due to different families (e.g. M1 versus M2) and usually revolving around the neural engine and its capabilities. One example is the ability to do hardware ProRes encoding/decoding, present in the media engine on the M2, but missing from the M1.
Not pay anything more? 400 and 200 euro difference with 13" and 15" 512GB MBA (even with the overpriced upgrade to 512Gb and discounts not withstanding). 8GB was standard in a 2014 MBP...But defend this practice if you like. It costs Apple next to nothing to upgrade these machines to 16GB. It is just their way of earning a good ~80% margin on what you pay for the upgrade to 16GB. Also note that the higher ram+ ssd configs are seldom on sale (e.g. MBA 512+16gb), but the base models are. My guess is that more and more people are fed up with these meagre base models and are thus a hard sell.
I meant I’m not a “creator”, and I don’t do much video rendering or editing on a computer anymore. Macs have been promoted by Apple to be the “creator’s” computer. Maybe so but they suck at being the computer for the power users in business, science and engineering, compared to Linux or Windows. The only good thing is the battery life (and quietness, both result from power efficiency where x86 sucks).
Even if that statistic is accurate, it's hardly a bad thing. It means they've influenced 50-60 million users to upgrade their Mac in the last 3 years (considering Apple's Mac user based topped 100 million 2018). Considering the lifespan a Mac, Apple must know that whilst their product line up has switched over in 2-3 years, it will take their eco-system the best part of a decade.
But in terms of marketing/positioning you're spot on. They're targeting to Intel Mac owners and probably looking to shave a year or two off that 10 year transition period.
Base models are NEVER on sale/discounted direct through Apple. The discounts you see from third party re-sellers are them simplying shaving a cut off their own share of the sale to get shoppers through their doors.
More GPU cores directly means more performance. That's because graphical tasks can be broken up into smaller independent tasks and thus can be distributed across as many cores as you have.
That's not generally the case for CPU cores, however. As @tobybrut explained, CPU mfrs hit a power wall when it came to increasing per-core speed, so their only way to significantly increase CPU performance was to add more cores.
However, the overwhelming majority of apps continue to be single-threaded only. Thus, since consumers typically aren't running more than 6-8 apps at once, any (performance) core count beyond that is simply wasted (gives no increase in performance), unless you happen to be running a multi-threaded app (because those apps can use multiple cores). And the only apps that are multi-threaded are those that contain tasks that can be easily broken up into multiple independent pieces, which mostly means apps for video, photography, other graphical tasks, and multi-channel audio.*
[*The one exception I can think of is Spotlight indexing, which all Macs do, and which can run on all cores; but that typically runs in the background.]
Bottom line: Unless you are running one of those heavily multi-threaded apps, you probably won't benefit from any increase in CPU core count beyond 6-8 performance cores. Though even if you're in that latter category, you might still need to get a Max or an Ultra—not because you need the large number of CPU cores they offer, but because you need their higher levels of I/O connectivity, GPU power, RAM capacity, or memory bandwidth.
As I explained above, that would be the case only if you are using programs that can take advantage of those added cores, which many apps don't.
Last edited:
Amen to this. Back when I started as a programmer in the early 90’s, programmers were told to write single-threaded apps whenever possible and to avoid multi-threading. First of all, CPU’s had to time slice in order to run multiple threads, which simulates simultaneous processing but really isn’t, and second, multi-threaded apps are a PITA to debug since things are happening all at the same time from a human’s perception. Multi-threaded apps are much more bug prone since you can’t always predict how one thread can affect another (e.g. changing the same object two threads are accessing, or overwriting results, etc).
When CPU makers started making multi-core CPU’s, programmers had to reverse course and start writing multi-threaded apps or else performance would stagnate since single cores can only run so fast with the clock speed limitations. A lot of programmers still try to avoid writing complicated multi-threaded code, or if they do, they significantly isolate each thread from the others to prevent problems from threads finishing out of order.
One thing to note on all three M-proccessors I didn’t mention in my long post is that each individual core runs at the same speed. An M3 Max running a single task on a single thread isn’t going to be any faster than a single task on a single thread running on a base M3. The cores are all the same, only differing in type and number. This is why tests like Geekbench all show very similar single-core scores but very different multi-core scores.
That is mostly true, which is what make AS's peformance on thin-and-lights like the Air so notable--it gives them about the same SC performance as the big top-end desktop processors from Intel and AMD.
One interesting exception, though, is that, starting with the M2, Apple did give modestly higher SC speeds to the upper-end SKU's. Specifically, the Ultra, and the Max in the Studio & 16" MBP (but not the 14" MBP), run at 3.7 GHz. All other M2's run at 3.5 GHz.
We'll need to wait to see what they do with the M3.
Obviously YMMV. After carrying a 6 pound 17" MBP all around campus and field work for years I find the 4.7 pounds of the 16" M2 MBP no problem. And I am old.
True and I upvoted this. But what I wonder is what happens with advancing graphics apps, ray-tracing, etc. Those kinds of things must multi-thread, yes?
It's more like "If you have an Intel Mac, upgrade now because Intel Macs won't be supported for much longer"
Still doesn't change the fact that Macs, and for that matter iPads, have quite a bit longer upgrade cycles vs iPhones.