But you can easily notice the single core speed increase just going back and using an iPad or iPhone from 3-4 years ago. And that's just interacting with the system as a normal person.
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 Silicon is Overpriced and over valued in technology market
- Thread starter trevpimp
- Start date
- Sort by reaction score
I have a Mac Mini M1 and a MacPro 5,1. A M1 without the special hardware mp4 Decoder/Encoder is a bit slower than my MacPro 5,1 Dual 6-Core 3,33 GHz from 2010.
Cinebench is faster on my MacPro
Handbrake without Videobox is faster on my MacPro
Topaz Video AI ist faster on Mac Mini M1
But the big PLUS: My MacPro needs more than 300 Watt to do the work, and the M1 around 30 Watt for the same work in nearly the same time.
Last edited:
Yeah the m1 is incredible and the dedicated hardware in the pro and later variants really overshadows just how impressive it is even without that.
And how doe the speeds compare on more mundane tasks like coding, browsing, photo editing, word processing? Tasks that are not massively parallel?
Indeed. And that's what determined the architecture of A and now M chips. However, the situation on the computers is very different. Here, the CPUs have had adequate ST performance for quite awhile.
I'm not sure what you mean. How is the situation on computers "very different"? Single threaded performance has been pretty stellar compared to desktop CPUs on Apple Silicon for quite some time.
To steal from ahurst: "at the time of the M1's release, the top-of-the-line 27" Intel iMac with an Intel i7-10700K managed a solid 1250 single-core score in GeekBench". The A13 released before that computer had a single-core score of 1322, which is higher.
And you can feel the difference between an A13 based device and an A15 based device, for example. They're both very good experiences, but you can feel the difference.
Last edited:
I remember being absolutely blown away by the A14. It was blisteringly fast compared to any desktop CPU at the time and it was low power enough to fit in a cell phone? Absolutely stunning.
Now Apple is pushing 1900+ SC scores on the M2 in Geekbench, and that's at less than 5 watts per core. Unbelievable to this day.
I still remember before the M1, all the claims of geekbench scores being somehow cheated on iPhone because there's no way it should be even close, nevermind faster than Intel.
We don't hear those anymore lmao...
I went from from a maxed-out-everything 2017 5K iMac to a 14” M1 Max 32GB with a Studio Display, and I believe I can feel an increase in general snappiness. There is still room for improvement, apps do take noticeable time to start, and searches could be faster.
M1 was incredible, but Intel and AMD have both already caught up and will surpass Apple due to how long they go between updates.
M2 was mostly a lateral move, and the main product it’s in saw a giant price hike, killing the value proposition.
caught up? If you look at performance alone, but in performance for watt? Not even close. And in performance for fan noise? Again, not even close.
The added value of M2 is not in performance, it's in smaller bezels, magsafe, redesign in general. Having said that M1 value still has more value per money, especially as it's more often discounted
There’s no actual value in smaller bezels or a redesign. Those are simply cosmetic changes.
There is no value in what? In your mind? I believe that not everyone shares your opinion on this.... 😉
One might argue that cause and effect are backwards here: because Intel chips made such minor single-core gains between 2012 and 2020 (and base RAM sizes also plateaued around 8 GB), programs that were designed for Intel's latest CPUs would also perform pretty well on their 8-year-old ones as well, hence CPUs having "adequate ST performance" for a long time.
Honestly, as annoying as it was to see such minor benefit to upgrading for so long, Intel's years of struggles with die shrinks was an unintentional-but-incredibly-effective e-waste reduction program. Thanks to the slowing of the pace of improvement, my Late 2013 27" iMac is still perfectly usable for modern work in a way that a 2004 iMac in 2013 wasn't, and my 2011 ThinkPad X220 held me off nicely until late 2021 when they released the 14" MBP.
My understanding is that the startup may be because that is when recent Mac OS runs it’s malware check on apps. I know it is an easy metric to count bounces but I’ve never found that it amounted to a significant amount of time compared to actual run time of apps. The only apps where it is really noticeable is the Affinity apps when loaded from the App Store. The versions purchased directly from Affinity open quickly. Something is going on there.
It's nice to hear a real-world take on the practical day-to-day improvements.
That's pretty much how I'd describe the change in going from my max-CPU 2014 15" MBP to my max-CPU 2019 27" iMac: Noticeable and useful, but modest—nothing that blew me away.* [Here I was going from Intel's fastest non-extreme 2014 mobile CPU to their fastest non-extreme 2019 desktop CPU.]
Since I'd like to see even more improvement when I upgrade my iMac to Apple Silicon, it seems I'm going to need to wait a few more generations (if I upgraded my 2019 i9 iMac to an M1 Max, I'd see less improvement than you did in upgrading your 2017 i7).
*I suspect the change in OS—I'm comparing the MBP runnning High Sierra to the iMac running Monterey—sucks up some of the speed improvement.
Last edited:
Smaller bezels do have value in a portable-it means a smaller form factor for the same screensize.
And while I value function above form, aesthetics do have value.
Last edited:
Yeah, cosmetics have no actual value. That's why the world cosmetics market is only $380 billion.
This is just messing around in Finder and Safari though.
If I export or transcode video, compile code or compress big archives the difference is massive, especially considering the iMac would sound like a jet taking off after 20 seconds of such an activity.
However, I spend a lot more time on stuff that benefits more from single-thread performance than I do on these tasks.
This is what my CPU has been up to the past 30 days.
Plausible. I am a compulsive app quitter, for no good reason really, other than 35 years of habit. I may need some retraining.
Yep. $1,200 M1 Mac Mini traded blows with my $4,500 i9 iMac. I sold my iMac shortly after.
It would cost me $5200 + $1600 = $6800 to replace my 2019 27" i9 iMac with an equivalently-equipped (128 GB RAM + 2 TB SSD) Studio plus Studio Display. And the Mini wouldn't work for me in any case, since I use three monitors.
Just not worth it at this point, since the main program I use that needs more speed is Mathematica, and a detailed benchmark corresponding to my typical workload showed only a 10%–20% improvement in going from my i9 to a Studio (except for image processing in Mathematica, where the Studio is slower).
I'll revisit this again when the 3nm M3 Studio comes out. Maybe by then Mathematica will be better optimized to run on AS, and the price may drop significantly if I don't need to spring for the Ultra to get enough RAM (a maxed-out Max should offer 64 x 12/8 = 96 GB, which may be enough, depending on the work I'm doing—and anticipate doing—at the time).
Last edited:
If you're talking Mac Studio (didn't think Mac mini went to 128G) - then it can drive 3 monitors or more.
The studio monitor is also a step up from the iMac display.
Honestly, I've found that the Intel Macs don't really feel much slower in normal day-to-day use. On Apple Silicon, Mac OS is trying to use the efficiency cores when it can. These cores are nowhere near as fast as the P-cores, so you aren't necessarily always getting the full performance of the processor when Mac OS doesn't think you need it.
Work sent me a 2017 Macbook Pro with a dual core 7th gen i7, and it geekbenchs about 980 SC and 2400 multicore. It honestly feels just as snappy on most day-to-day tasks as my M1 does (1700 SC, 7600 multicore). It's to the point where I sometimes forget which machine I'm on, and that's despite a pretty massive difference on the benchmarks.
Apple Silicon of course clearly wins on power efficiency, and it also obviously wins massively on raw performance for pretty much anything that is even remotely demanding (as soon as I start running XCode builds for iOS apps, I feel it immediately). But I wouldn't say that everyone is in for an earth shattering difference on the speed of their systems either, some of the more recent Intel ones are still quite snappy and are perfectly fine for a lot of people. For real-world "everyday tasks," the difference sometimes hardly even feels noticeable.
Work sent me a 2017 Macbook Pro with a dual core 7th gen i7, and it geekbenchs about 980 SC and 2400 multicore. It honestly feels just as snappy on most day-to-day tasks as my M1 does (1700 SC, 7600 multicore). It's to the point where I sometimes forget which machine I'm on, and that's despite a pretty massive difference on the benchmarks.
Apple Silicon of course clearly wins on power efficiency, and it also obviously wins massively on raw performance for pretty much anything that is even remotely demanding (as soon as I start running XCode builds for iOS apps, I feel it immediately). But I wouldn't say that everyone is in for an earth shattering difference on the speed of their systems either, some of the more recent Intel ones are still quite snappy and are perfectly fine for a lot of people. For real-world "everyday tasks," the difference sometimes hardly even feels noticeable.
Last edited: