Although my main Mac at home is a 2018 mini, I also use a 2011 Mac Book Air (entry-level model - 11" screen, 4G RAM, 128G SSD). Which works great for what I use it for - web browsing, streaming video and light use of office apps when I don’t want to be in my office.
The last version of macOS that can run on it is High Sierra, and I’m actually still running Sierra on it because I never thought Apple fixed some its APFS growing-pains bugs before they moved to the next version.
MS Office hasn’t received updates for several years, but I don’t care because I only open documents that I create.
But Firefox is dropping support some time next year. And a web browser needs to be kept up to date, both for security and because web sites will become incompatible. So I’ll be forced to upgrade this 12 year old computer next year, even though, as far as I’m concerned, it’s still working great.
(Yes, I could probably install Linux on it, but if I’m going to switch to a Linux laptop, I’d rather do it on PC hardware. And I don’t want to use a Linux laptop when traveling.)
Can someone please remind us (me) what Apple Silicon is actually all about? It’s just that every time I’m at the point of thinking about how I’ll upgrade my perfectly good, perfectly fast, top-end iMac '20, I suddenly realise the sharp divide that exists between need and desire. It also helps that I bought this iMac just as Apple Silicon was being earnestly speculated about, and I didn’t want to be left without a Windows box if it all went wrong. (Even now, sadly, my need for Windows can’t be entirely met by a VM, but it’s getting closer as screen readers adapt.)
So I guess no, you’re probably not. But it’s interesting to step back and take a fresh look at the question every so often. In truth, all the speed you need is actually the correct assessment of what you have. That doesn’t mean opposition to innovation, or refusal to recognise legitimate needs. It just means being ready to be content with what one already has. I just moved from a 1G/150M cable connection to a 3G/3G FTTP connection; I would rather (read: strongly prefer/desire to) not go back, but I’m comfortable that I could, if the need arose. The FTTP connection is about so much more than just speed–congestion free, symmetrical speed, exceptionally low latency, and these things together clearly represent future progress. But these things though desirable aren’t essential either, and the most that one can say of them is that every good citizen ought in principle to be able to enjoy them today, while future potential uses for them for the great unwashed are dreamed up tomorrow.
Based on a recent Eclectic Light article, apparently using benchmarks like Geekbench to evaluate M3 performance may be misleading
Although we’re going to hear a lot of results from benchmarking apps like Geekbench, remember that the tests they run don’t simulate real-world CPU usage. For instance, they’re designed to run the same processes on each core when being used to measure multicore performance. In reality, macOS should manage distribution of the very different threads running in real-world use, to make best use of the cores available. Benchmark results are but part of the evaluation of performance.
When you hear anyone making claims that the 6P + 6E design of the M3 Pro is merely 50% more than a regular M3 chip with its 4P + 4E, or slightly over half an M3 Max at 12P + 4E, get them to show you their evidence. Measuring and comparing the performance of Apple’s new M3 chips has become much more complicated, and that’s before we’ve even considered the GPU.
Finally, be very wary of what you see in Activity Monitor’s CPU History window. While it does show broad trends in the distribution of workload across different cores, it doesn’t take account of frequency. There’s a world of difference between an E core running at 100% and a frequency of 1 GHz and a P core running at 100% and well over 3 GHz. If you want the full picture, then you have to resort to tools like powermetrics .
His tests focus on tight code loops of basic arithmetic. That’s very useful to gauge the performance of a P or E core, learn how they’re clustered, how their clocks behave, compare cores to each other, and learn how macOS distributes load among them. But I also like seeing more integrated benchmarking besides such tight testing because it does attempt to replicate a more common workflow such that we learn something about interfaces to other components (eg. memory bandwidth) or things like load balancing between CPU and GPU cores. In fact, Howard has never claimed to benchmark GPU core performance, which quite obviously, is also an important part of the Apple Silicon sauce.
IMHO he’s right to caution people against judging these new Mac solely by a Geekbench score. I also believe that Geekbench scores, just like his FPO/s tables do give us valuable clues as to how these new Macs perform and why they perform the way they do.
Count me in as another person that has managed to keep older computers going. I have a early 2009 24" iMac running El Capitan as a Daily Driver. I installed an additional 4GB of RAM to bring it up to 8GB, added a FireWire enclosure with an SSD as a boot drive, spinning hard drives for time machine. It still has its original HD, but I run the machine with it unmounted. Performance is decent for such an old machine and I rather not upgrade the OS with OpenCore as it would prevent some perfectly good apps from running on it, and the performance would degrade. I use Chromium as a current Web Browser as support for all of the popular ones have ceased. I can also remote into my M1 MacBook Pro if needed.
I think you’re both right, and nevertheless, I feel with Apple Silicon it’s the reason to upgrade that changes, not the urge.
I would argue for most people (myself included) the M1-M3 we’ve so far seen offer ample performance for 99% of our tasks. So I doubt CPU core improvements or clock adjustments or mem b/w increases or any other direct CPU tweak will be an immediate reason to upgrade. In that sense I completely get why a lot of people are now finding themselves feeling “why should I get a new M3 Mac, my M1 is still awesome at everything”.
But I also know that beyond just CPU performance there’s a lot I could see improve and when it does I know I’ll want to jump on a new Mac at once. I’ll admit, much of it pertains more to mobile computing than desktop. (That said, these days my desktop reality is mobile computing because rather than a Mac Studio plus say MBA, I just use a beefed up MBP with a nice docking station and fancy screen for my desk work).
Working in bright sunshine. The screen can never be too bright. Once we get even brighter displays or perhaps micro LED that to me would be a very good reason to upgrade my MBP.
Non-tethered mobile. Tethering works but it’s always a bit finicky and it tends to run down the iPhone battery fast. One day I suppose MBPs will come with an eSIM and native cellular capability. That is an upgrade I’d jump on. Not because of dire need (there is tethering), but because of convenience.
80 Gbps Thunderbolt 5: To this day the fastest external storage you can get is ~2.8 GB/s read and ~2.1 GB/s write. That’s fast, but it’s also 2-3x slower than the Mac’s internal flash. The reason for that bottleneck is the “40 Gbps TB4” that effectively offers 32 Gbps peak data transfer which then in reality translates to sub-3 GB/s. So when the next MBP with TB5 comes out (which should finally get external storage b/w at least into the vicinity of today’s internal flash), you can bet I’ll want to upgrade along with getting new TB5 docking stations.
Battery life. My 14" makes it through a full work day on one charge, sure. But if I’m spending lots of it on wasteful apps like Zoom and I perhaps also have to crank up my screen real high because I’m sitting for hours at an airport with lots of sun shining into the waiting area and onto my screen, I’ll be having to keep an eye on battery nevertheless. And I’m not inclined to carry around battery packs forever or go chase after outlets. So the moment we get MBPs with 24+ hr spec’ed battery life, I’ll definitely want to upgrade to that.
And that’s just my list. I’m sure others would have other points (10G Ethernet perhaps? Or touchscreens? Faster/better mem card slot? Colors?). Either way, even if performance is overkill and we’re already sufficiently happy with speed and responsiveness, I think there will eventually be plenty things left that will get people to upgrade their Apple Silicon Macs outside of Mx upgrades. I have a suspicion Apple will try to entice folks to get new Macs by marketing the heck out of Mx renewals, but that it will be other things like above that will get Mx owners to actually get a new Mx Mac.
Thanks, I follow Howard’s articles religiously and agree that the multicore Geekbench results can be misleading. The single core results are probably more reliable and show Apple’s claimed improvements of 15%.
The Geekbench tests also reveal the clock speed which is 4 GHz for the M3 chips - this is partly responsible for the speed increases and I would think is one of the chief benefits of the 3 nm process. (M2 chips are clocked at 3.7 GHz). Of course another benefit of the process is lower power consumption.
On the one hand, I don’t do anything that stresses my current hardware that much. So its safe to say (as I have said) that everything sold today by every manufacturer (not counting companies selling complete junk) would be able to do what I need. I could probably meet all my personal computing needs with a Raspberry Pi and enough USB storage to hold my media collection.)
But that’s very different from “can imagine ever needing”. I can imagine quite a bit. I think I would really like a voice assistant that’s completely self-contained, only accessing the Internet services that I explicitly authorize. Sort of like an off-line Siri. We may be getting close the the time what that becomes possible. I can also imagine needing much faster GPUs in order to handle future streaming media programming (holographic immersive programs? Integration with VR/AR hardware?). I can easily imagine Star Trek level computing tech (just watch the show). Not counting future requirements (like being able to fly a starship), a lot of what they show is stuff I could imagine wanting and maybe even needing, should it become reality.
So I don’t think computing will ever get as fast as I can ever imagine needing.
But then we can fly to the complete opposite drection. What do I actually need? For my employer, I obviously need enough computing to do my job. Which means a modern office suite, software developer tools for a wide variety of embedded platforms, fast enough to meet our customers’ requirements.
For my personal life, however, I don’t need computing at all. I could track everything I need tracking using paper notebooks. I could (and until about 12 years ago did) pay all my bills by mailing paper checks. Although I certainly like my iPhone, I could get by just fine with a land-line and no other capabilities. We’re not yet at the point where it’s impossible for function without computers, although we might be getting there in the next 5-10 years.
But that’s just like most things in life. We don’t need streaming TV services. We don’t need TV at all.
But the fact that I don’t need something doesn’t mean I don’t want it. The fact that I can get all of my entertainment from books, board games and the radio doesn’t mean I don’t want my home theater system, Apple TV and PlayStation game console.
I’m still using a 2010 MacPro5,1 running Mojave. My wife has a 2012 13" MacBook Pro, also running Mojave. Both computers are more than enough for our needs. However, with web browsers now dropping support for Mojave it may be time to move on.
I just ordered an M3 24" iMac (16Gb, 1TB) to replace the MBP. The M3 will seriously outperform the 2012 MBP and the nuances discussed here will be mostly irrelevant. We’ll evaluate how that goes and then I’ll decide if I want to replace my MacPro.
Opinion: The MacPro up through 2012 may have been the high point of Mac desktop computers with expandability, capability, and price. The 2013 expandability was limited and the 2019 was priced too high.
Yes, the ports and the bigger screen speak for the M3 MB Pro, but the Air is more than a pound lighter. That weight difference make me tend towards the Air.
I wish when a presentation was aimed at an English speaking audience that all presenters would have English as their native language. I struggle when not.
My solution to the single external display support on the lower end MacBooks is to use a DisplayLink-supported dock/hub. That will happily drive two external displays at up to 4K each.
Here’s an odd one. The presently shipping M3 MBPs (regular M3, not Pro) ship with Ventura that claims 13.5 is the latest version it can update to. That they ship with Ventura is no surprise since Apple has likely been stockpiling for a while already, but that these MBPs don’t suggest the Sonoma update is a bit odd. Fortunately, folks can download the installer straight from Apple and then get 14.1 that way.
Although I have no interest in getting any of these devices, I’ve been following this conversation with interest. For those who haven’t seen it, there’s now this posting on MacRumors.
Faster/better mem card slot? Colors?). Either way, even if performance is overkill and we’re already sufficiently happy with speed and responsiveness, I think there will eventually be plenty things left that will get people to upgrade their Apple Silicon Macs outside of Mx upgrades. I have a suspicion Apple will try to entice folks to get new Macs by marketing the heck out of Mx renewals, but that it will be other things like above that will get Mx owners to actually get a new Mx Mac.
We’ll evaluate how that goes and then I’ll decide if I want to replace my MacPro.