You said it my friend, how the economy currently is in the country; well i hope the people buy wisely.
Current users with M1 (i myself) still happy with the machines no reason to upgrade, others are finding whatever excuse to just buy it.
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.
iPad Pro's M4 Chip Outperforms M3 by Up to 25%
- Thread starter MacRumors
- Start date
- Sort by reaction score
Current users with M1 (i myself) still happy with the machines no reason to upgrade, others are finding whatever excuse to just buy it.
You know, for those like me who still rock a 2018 iPad Pro, this would be a significant upgrade. Also, it’s a pretty impressive performance bump over the m2 and m3, if the Geekbench scores are real
The folks who would upgrade sooner are generally the folks who either need the extra power anyway, or are enthusiasts who tend to upgrade a little faster than the average market. Nothing wrong with that, it's just not what the everyday average customer is as likely to be doing.
Most of the people I know pretty much upgrade on a four or five year cadence, sometimes longer for the laptops.
I just hope by the time it gets put into the Mac, there are better options with more memory.
If Apple are serious about an AI push and doing things on device, it needs more memory and lots of it. Baseline needs to be at least 16gb with a lot of 32gb versions. AI stuff loves ram.
If Apple are serious about an AI push and doing things on device, it needs more memory and lots of it. Baseline needs to be at least 16gb with a lot of 32gb versions. AI stuff loves ram.
I’ve been hearing a lot about the single core still not “revamped” with the M4, i see the M4 is really just a light boost but nothing revamped. Go For it Mate, definitely a huge jump for ya.
HOLY. Thank goodness I decided to skip the 16" M3 Max MacBook Pro and wait to max out the 16" M4/M5 Max MacBook Pro. That's insane. Good things really do come to those who wait! Super excited, especially since with the start of the M3, MacBooks get hardware based ray tracing
May 20th, MS is holding an event, in partnership with Qualcomm & Asus, to showcase new Windows devices powered by Snapdragon X Elite. Guess we'll find out then.
Fascinating. Thank you very much for sharing this, crazy dave.
Welcome! just to reiterate though the exact numbers are subject to run to run variation and looking at the data on GB's website I may have picked a particularly good M4 iPad result.
But's it's more about the overall relationships rather than the absolute numbers.They were confident about beating M2 and plan was to provide Windows Laptop makers chips by end of 2022, and launch windows ARM devices in early 2023.
Qualcomm Promises to Outplay Apple's M2 With Its Upcoming Arm Chips
Crediting ex-Apple engineers in a white-gloved wave at Cupertino.
Just think of all those Qualcomm Snapdragons that haven't even shipped in laptops yet!! 😂🤣
Higher single core perf on the cores (and there is a big jump here) may make it easier to bin that way without huge perf loss, you realize that right?
Different gens of chips are going to be binned differently depending on yield and what tradeoffs need to be made, that’s how the industry works, I’m not sure what you’re not getting that this is typical of CPUs/GPUs/SoCs/etc
I have my doubts the limiting factor on the lower end pros is going to be the loss of a CPU core anyway, the RAM is more likely to be a problem before the CPU is. And for most apps the iPad doesnt peg those CPUs anyway. *And* thermal throttling actually may mean those lower end SKUs can peg the CPU longer before throttling down (remember, no active cooling, and one less core means a corresponding drop in TDP) which may even things out even more. Oh *and* the boost in performance to the efficiency cores and scheduling on behalf of the OS will probably smooth that out even more
Not to mention the performance jump from even the previous gen (which isnt the upgrade target since by far most people dont upgrade every gen), even with a lower core count on the lower end SKUs compared to higher end ones, let alone from earlier devices (the upgrader targets) kinda negates this as an issue anyway.
Tl;dr you’re making a mountain out of a molehill on this
Last edited:
They were confident about beating M2 and plan was to provide Windows Laptop makers chips by end of 2022, and launch windows ARM devices in early 2023.
Qualcomm Promises to Outplay Apple's M2 With Its Upcoming Arm Chips
Crediting ex-Apple engineers in a white-gloved wave at Cupertino.www.tomshardware.com
I actually wonder if Apple launched the M4 on ipads first to give Qualcomm a launch window to try and compare to M3 before Apple smacks them with the M4 in Macs just afterwards
Edit: Also really really hoping that M4 or M5 macs bring back bootcamp, Microsoft’s exclusivity deal with Qualcomm should end this year, it’d be nice to have that back
Also it would certainly be a feather in Apple’s cap for Apple to beat Qualcomm at an ARM windows laptop
Last edited:
Dude, calm down, it's not a mountain. You're making out like I said it was a catastrophe or something.
If those numbers are at all accurate, that performance jump is pretty amazing. Maintaining that kind of significant, steady performance jumps every year for four years straight is not a small feat.
Putting it in broader context, the few-month-old, consumer-top-of-line desktop Intel Core i9-14900K is averaging 3100 single core and 21,000 multi-core on the Geekbench Browser. That's with a 125W TDP. The newer enthusiast-grade i9-14800KF with a 150W TDP doesn't have enough listings in the browser to get an average, but looks to average maybe 3350 single and 23,000 multi.
On the mobile side, the top-of-line i9-14900HX manages about 2800 single and 16,300 multi with a 55W TDP.
There's also the oddball massively-underclocked 35W desktop i9-14900T, which is actually the closest to the M4's presumed 20W TDP, but has very few tests logged and they're extremely erratic, probably because it can boost to >100W, so the length of the test and heat sink are both going to have a huge impact on performance. Look like it's probably around 2700 single and 12,000 multi, but it's possible to get about 3000/19,000 out of it with under some circumstances and/or the right thermal management.
Leaving aside sustained performance in a fanless, ultra-thin tablet (which it will be very interesting to see measured, since the previous fanless Airs could go maybe 10 minutes before needing to throttle), all that is to say that a single peformance core of Apple's new but lowest-end laptop/ultralight CPU is currently outperforming the absolute top-of-line, top-binned, power-hungry best Intel has to offer by a good 10% and their "normal" high-end consumer desktop chip by 20%. And because Intel's thermals (and maybe CPU design?) limit multi-core performance, on that end things are closer than you'd expect against CPUs with double the performance cores and 2.5x more efficiency cores.
Moreover, compared to actual mobile processors, Apple's new low-power base model is not far behind Intel's high-power top-of-line--a full third faster single core and 12% slower multi-core in around a third the power consumption, and again doing that with less than half the cores.
If the M4 Pro, M4 Max, and M4 Ultra all end up existing, and all have a similar performance ratio to their respective current iterations, we can probably expect 19,000, 26,000, and 32,000 multi-core scores out of them. If a theoretical M4 Ultra actually scales better than the previous Ultras, that would be something else entirely.
Putting it in broader context, the few-month-old, consumer-top-of-line desktop Intel Core i9-14900K is averaging 3100 single core and 21,000 multi-core on the Geekbench Browser. That's with a 125W TDP. The newer enthusiast-grade i9-14800KF with a 150W TDP doesn't have enough listings in the browser to get an average, but looks to average maybe 3350 single and 23,000 multi.
On the mobile side, the top-of-line i9-14900HX manages about 2800 single and 16,300 multi with a 55W TDP.
There's also the oddball massively-underclocked 35W desktop i9-14900T, which is actually the closest to the M4's presumed 20W TDP, but has very few tests logged and they're extremely erratic, probably because it can boost to >100W, so the length of the test and heat sink are both going to have a huge impact on performance. Look like it's probably around 2700 single and 12,000 multi, but it's possible to get about 3000/19,000 out of it with under some circumstances and/or the right thermal management.
Leaving aside sustained performance in a fanless, ultra-thin tablet (which it will be very interesting to see measured, since the previous fanless Airs could go maybe 10 minutes before needing to throttle), all that is to say that a single peformance core of Apple's new but lowest-end laptop/ultralight CPU is currently outperforming the absolute top-of-line, top-binned, power-hungry best Intel has to offer by a good 10% and their "normal" high-end consumer desktop chip by 20%. And because Intel's thermals (and maybe CPU design?) limit multi-core performance, on that end things are closer than you'd expect against CPUs with double the performance cores and 2.5x more efficiency cores.
Moreover, compared to actual mobile processors, Apple's new low-power base model is not far behind Intel's high-power top-of-line--a full third faster single core and 12% slower multi-core in around a third the power consumption, and again doing that with less than half the cores.
If the M4 Pro, M4 Max, and M4 Ultra all end up existing, and all have a similar performance ratio to their respective current iterations, we can probably expect 19,000, 26,000, and 32,000 multi-core scores out of them. If a theoretical M4 Ultra actually scales better than the previous Ultras, that would be something else entirely.
I do not understand this kind of comment. I personally also feel impressed by these numbers but just because I picked a Macbook Pro over a Macbook Air and I now realise that for my workflow the MBA would have been enough in terms of performance (with 16 or 24 GB of RAM). And I like the lighter weight of the Macbook Air (it's a laptop after all, we don't only sit on the same desk everyday with it).
But for the actual performance of the chip, I don't even know how to find its limits. Yes training a ML model used 5-6 cores on the M2 Pro. But that's not something I do everyday. I don't do video editing so I don't know how limited the M chips are for this kind of thing. For everyone else (so 98-99% of posters here), I suppose even the M1 Pro is enough for 99% of their daily use in the next 5 (10?) years.
These machines are just overly powerful and a Macbook Air nowadays offers so much power for such a small weight that it's hard to go beyond that. My MBP is better than any computer (including a desktop) I could ever dream of (I can just plug my SD card full of photos and there I go, I have one of the beautiful screen I can dream of), and I can still carry it in my backpack everyday. As my girlfriend puts it, the weight difference is just a couple chocolate bars
More chocolate or more power, that might be the real dilemma
Sadly I can only offer you $300 for such an old machine. But you really should upgrade. Trust me... please!
Nice that the M4 can run faster without using more power. Impressive results.
It shows a nice jump in the Antutu benchmark too!
It shows a nice jump in the Antutu benchmark too!
Very impressive year-on-year gains. Looks like the M5 will be double the M1.
What are the odds we get M4 Pro and M4 Max before Christmas?
What are the odds we get M4 Pro and M4 Max before Christmas?
On multi core, yeah, while on single core it should be M6, if progress remains consistent.
Hopefully we see Pro and Max make an appearance at WWDC...
Strange release timing. I wonder if an AI boost was needed earlier than planned, and the M3 series wasn't up to the task?
WWDC and the Fall events will be exciting!