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.
M4 non-Pro vs M4 Pro - Help a guy stuck down a rabbit hole out?
- Thread starter Superman041
- Start date
- Sort by reaction score
Just because OP stated 2 options of the same price, doesn't mean there isn't a third option that meets all requirements and would cost $200 less...
And when you extrapolate 5 years into the future, it doesn't mean the base M4 chip wouldn't meet requirements without breaking a sweat.
It's the same performance as the M3 Pro.
Would you have told me 2 months ago that the M3 Pro isn't good for 5 years?
I mentioned the M1 because if that's your starting point as far as meeting computation requirements goes today, then the M4 should already be the chip to serve you very well for 5-7 years.
You are talking about something else, not the OP. You appear to be saying that Macs stay strong for years, which [quite obviously] is true, but is not the question. The question presented was:
A) M4 base chip with 1 TB SSD versus
B) M4 pro chip with 512 GB SSD
at similar price, and OP seeks a long life cycle.
The B) M4 pro chip is physically substantially larger, with
• twice as many performance cores
• more than double the memory bandwidth
• Thunderbolt 5 instead of TB4
External SSD capacity [which almost all workflows requiring >512 GB mass storage should have anyway, for backup purposes] is fast, competent, cheap and has been getting cheaper every year for 40 years.
IMO choosing B over A is an obvious no-brainer decision. It makes no sense to give up substantial performance and i/o capacity just to get 1 TB mass storage instead of 512 GB mass storage when external mass storage is so cheaply available.
Last edited:
If you dismiss the potential for having an interesting discussion with a broader picture and coming up with potential alternatives, then yes I agree - the M4 Pro 512GB is better value than the M4 1TB at the same price point.
You are 100% correct that "the potential for having an interesting discussion with a broader picture and coming up with potential alternatives" certainly exists. Unfortunately once we add in the used market place there become so many permutations and combinations that a coherent text based conversation becomes impossible.
Given you both are smarter and more knowledgeable than me. What would the extra core/gpu
The 14-core CPU choice you reference above is not insignificant, simply do the math of 14 being more than 12 and 20 being more than 16; an 18-25% improvement that will be relevant to folks pushing their workflows to need speed. To me that cost is very good value for an 18-25% performance upgrade, but probably most users will never notice it. Apple is really good at making price points right on the edge of where our individual cost inflection points are. <sigh>
IMO the important big jump is from base chip to pro chip: bigger chip, more transistors, more memory bandwidth, TB5.
IMO putting the equivalent extra money into a jump up in RAM is an even better value add than choosing the 14-core CPU upgrade, but that gets to individual workflows. IMO most images-related workflows would favor having more RAM, but one would need to A-B test a very specific workflow to make that determination [YouTube tests generally should not be used for reference because they are sensationalized to make entertainment and get clicks, not to get boring real data].
Personally I buy both the speed upgrade and the RAM upgrade because i tend toward long 6-9 year life cycles. Both upgrades make significant improvement as OS and apps become more hardware-demanding over time like they always do.
Last edited:
Instead of overspending today in order to make (probably incorrect) predictions about the future and trying to proof youself against it, wouldn't it be wiser to do this instead?
- buy just the configuration that you need right now
- look to upgrade in 3-4 years' time
- trade your machine in while it still holds a respectable value (usually drops significantly after 3-4 years)
- add in all the cash you would've plowed into today's future-proofing efforts
- repeat with the new machine
Likely spending the same amount of money in total, or even less.
Shorter life cycles are always a choice. But suggesting that buying new 2x-3x in 5-10 years instead of 1x in 5-10 years in my long experience is generally not "Likely spending the same amount of money in total, or even less." Constantly buying new stuff whether cars or Macs brings a loss of resale value immediately after any new device is purchased resulting in spending more, not "the same amount of money in total, or even less."
Also in my experience the issues we should care about are not having a "much more up-to date system on average," but rather about having a strong performing system on average. I have found over decades that the best way to do that is to buy strong hardware with lots of RAM and keep it for a long life cycle as defined by one's workflow over time. Coincidentally that also IMO results in least cost per day.
In the OP example buying at base level is not "a much more up-to date system." The larger M4 pro chip with
• twice as many performance cores
• more than double the memory bandwidth
• Thunderbolt 5 instead of TB4
will perform stronger for an entire [longer] life cycle.
Bottom line is that your commentary "overspending today in order to make (probably incorrect) predictions about the future" using prejudicial verbiage like overspending is IMO flat wrong. We have 40 years of Macs now, and the growth of hardware/software has been quite consistent, so any thoughtful person can reasonably predict that ongoing growth of hardware/software demands will occur.
Configuring a box optimally today for its intended life cycle [that starts tomorrow] is neither about "overspending" nor is it about "(probably incorrect) predictions about the future." It is just sound purchase decision making. We must be willing to think about what our future needs are likely to be. Even though that process is imperfect, to do otherwise is just lazy.
Last edited:
Or, in 4 years' time you might just realize you can just buy an entry level Macbook Air, or the previous year's discounted model, that not only blows the hypothetical "future proofed" M4 Pro you'd have bought back then out of the water, but meets your current needs even better. Potentially reaping other benefits along the way, with faster storage, a better display, a better I/O, better design or build, better battery etc.
And because you didn't overspend 4 years ago and used trade-in while values were still pretty good, you could actually end up paying about the same or less this way.
Seriously, think about it.
If the system meets your requirements today, that's a strong performing system for you.
At this pace of CPU performance gains, the base model M chip in 3-4 years' time will outperform the best Max chip of today. Hell, it seems like the base model M chips nowadays match the performance of the Pro chips from just a year ago.
So if you buy an M4 Pro today, and you're not even using a third of its potential today, then by the time you (maybe) start needing that power, you could as well have saved all the extra money in the meantime and upgraded when you need to.
And seeing the year-on-year gains, that upgrade would be much more substantial (reap multiple other benefits), potentially come in at a much lower relative price point and still blow the likes of today's M4 Pro out of the water.
Side note: RAM requirements are completely overblown by people's inability to understand how caching and preloading works, so they'll end up in a situation where they've got a 32GB system with nothing but a browser open and 15GB RAM usage, thinking they've made the right choice going for more.
In reality, the majority of people will get by just fine with the base 16GB of RAM, and that includes professionals as well.
As a professional software engineer, I'd say that trying to predict the medium-long term future of personal computing in general OR the future of your own personal needs/interests several years ahead - is a fool's errand at best. And by overspending on upgrades that you don't need today, you're essentially trying to predict both of these at the same time. And you'll most likely end up being wrong anyway.
Clearly we disagree. And it is 2024 so we have the benefit of hindsight to look at decades of past Mac usage. I personally lived up close and personal the growth in images apps/OS hardware demands for those decades. Go ahead, look back; growth absolutely has been consistent enough to be generally predictable for a reasonable 5-year life cycle. Imperfect predictability to be sure, but hella better than just looking backward at last year, since any new box is only used next year and beyond.
Misleading statement.
External TB 3/4 drive max. at 2GB/s, no matter how fast a NVMe drive put into it, that is less than half the speed of internal SSD running through NVMe interface.
Yet, with the new TB5 ports on M4, TB5 M.2 enclosure will be our best weapon fighting Apple’s greed.
Last edited:
Obviously the TB5 drives are about twice as fast, but TB 4 drives are still easily capable of 3Gb/s+ which should be plenty of speed for anyone.
The maximum theoretical bandwidth of TB4 is 40 Gbits/s, which is ~5Gb/s.
I'm a professional software engineer and have just gone for a M4 with 32GB RAM, because anytime I get properly busy coding and debugging on top of 60+ browser tabs with memory hogs like DataDog and Jira on my 16GB Intel MBP from 2019, I can see swapping-like delays happening and swap usage shows over 8GB. Over the next 8 years expected lifetime, if I hit a CPU/GPU limit in a way that really blocks me it'll be worth getting the latest chip then, in a new laptop. If I hit a RAM limit at 24GB I'll regret not getting the extra RAM at the start. The extra £400 to get 48GB on an M4Pro makes it too expensive, and as others have said the performance of a base M4 is awesome and unlikely to be a limit for 95+% of people.
For the OP's actual question, since their expected usage is likely to be fine with 24GB RAM, it's definitely worth going for the M4Pro with smaller disk.
Genuinely curious - mind sharing what takes up so much memory in your workflow?
Probably the most resource intensive part of my workflow is when I'm doing backend development in IntelliJ across 3 different (fairly sizeable) project spaces simultaneously, at the same time utilizing multiple docker containers, having about 20 browser tabs open and Teams - even all that only takes up like 9-10GB of RAM in total at most (5 for IntelliJ, 2 for docker, ~2-3 for browser, Teams pretty negligible).
I'm mean I'm somewhat mystified as well! "A gig here, a gig there, pretty soon you're talking about real memory"
6.5GB Memory used on a clean start with no applications open, although there is firewall and anti-virus starting by default.
Then my usual set of applications:
- 3GB for main Firefox process, plus 20+ Isolated Web Content sub-processes varying from 50MB to 600MB, but I've seen over a GB in processes running Jira tags with X-Ray plugin active. Maybe Jira mem usage there is higher because uBlock Origin stops their trackers then they try loading again and again? I've seen well over 100 in the "blocked" count on a long-lived tab, but in a fresh tab it's about 20.
- Call it another 3GB
- 3GB on PyCharm with ~4 projects open, none particularly complicated IMO but one has a lot of YAML files and the others are multi-microservice monorepo projects so quite a few venvs although they're not indexed.
- Excel is almost a GB even though I've only got a couple of hundred rows across two oworkbooks!
- Outlook about 750MB.
- Miro app with one really big diagram open taking about a GB.
- Slack's renderer is half a GB.
I just checked my new M4 MBP and it uses 7.75GB before any apps open, according to the bottom panel of Activity Monitor's Memory tab. No wonder Apple upped the base spec to 16GB, if the OS is using almost 8GB there's not a lot left for AI anything!
That's a surprising amount of memory eaten by some apps for sure.
Unused RAM = wasted RAM -> MacOS uses more RAM if you have more available, it's about 3-4GB on an 8GB system after a clean reboot - though that can easily be 13GB on a 32GB system.
I guess they're really looking to eat another 3-5 gigs with Apple Intelligence, which would push 8GB systems into heavy swap usage even in the lightest of workloads...
Sentinel I think? Whichever it is, it doesn't use enough resources that I notice it much.