More performance is a good problem to have.
There might be apps outside your use case that need 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 M4 Geekbench scores wow…
- Thread starter SpitUK
- Start date
- Sort by reaction score
More performance is a good problem to have.
There might be apps outside your use case that need it.
Mores Law says nothing about performance but rather numbers of transistors per area.
The performance/watt-ratio of Apple's Mx chips are nothing but astounding. The M4 seems to be on par with the latest and greatest CPUs from Intel and AMD, and while they are in the "hundreds of watts" range the Apple M4 is probably around 20W. How on earth is that "pretty disappointing"?
It would of course be nice if the performance doubled every two years but that isn't realistic nowadays, and hasn't been for a decade or more. If your're disappointed by Apple's Mx SoC, prepare yourself for the future because we'll probably never again see such improvements that we did going from Intel's x86 processors to Apple Silicon. And it is not Apple's fault.
Maybe someone will come up with some new cool material that will enable a substantial raise in frequency.
That analogy would make sense if they were putting an Intel i9 in it and had to make it larger and heavier to support active cooling and a much bigger battery.
What most people are missing is the fact that this SoC is extremely efficient for what it can do. As stated it consumes half the power of the M2 at the same performance levels. This means it will run cooler and require a lot less battery to be much more performant during normal use.
The major advantage with the M4 in a device like the touch-based iPad though, is the high single core performance - user interaction and UX is extremely important and that single core performance will make it much more fluid and responsive.
Also need to consider that Apple needs to manufacture these SoCs in scale to make it worth the development and this SoC will also end up in the MB Air, MB Pro, iMac and Mac mini.
Plus, what about future-proofing? Get a lot of comments how the M1 is still a viable SoC and this is 3+ years on. This SoC ought to last a long time before you’d want to upgrade. (I’ve gone 6 years between iPad upgrades; original iPad - 2010, original iPad Pro - 2016, and now I have the M2 iPad Pro - 2023.)
P.S. Also find it humorous that we get comments like “way overpowered for the ipad”, and then others who complain that the two low-storage models “only” have 9 CPU cores (3p+6e). Unless you’re doing a lot of sustained high-end work loads, you’ll rarely ever see the CPU cores fully top out… and even then, the thermal ceiling will force the SoC to throttle rather quickly.
Moore's law has nothing to do about computer speed, but instead transistor amount. See https://en.wikipedia.org/wiki/Moore's_law
That’s flawed logic. Even if a task is not multithreaded, a system with multiple compute units can run multiple tasks at the same time. So while that single-threaded task can only run on a single core, it is not constantly interrupted by the system to run other tasks.
Your response is correct, I agree with you. My reply was to the poster about why we don’t have 24c processors in our iPads.
It’s very common that most applications are geared toward and take advantage of faster single core performance more so than multi core.
That might have more to do with core design than process node.
There's ample evidence that the M3 and A17 recycled performance and efficiency core designs from the A16 and added the ray-tracing GPU that didn't make the cut for the A16, hence the modest increase for both chips. The A16's "Everest" performance core codename is H15 whereas the A17's codename is 'H15 Coll' (and doesn't get a spiffy mountain or weather-related nickname). Prior generations were sequential; A15/M2's "Avalanche" cores were H14, A14/M1's "Firestorm" cores were H13, etc. This explains why the M3 is something of a hybrid of the A16 and A17 (and also has no core nicknames).
Apple likely introduced a new microarchitecture with this generation that runs significantly faster. We'll know once the M4 lands in a Mac that can pull the core codenames via Terminal.
I don’t think you understand what improvements Apple actually wanted from Intel; has little to do with performance, which Intel in fact had at the time. What Apple wanted is performance per watt, i.e. more efficient chip designs. No one could argue that Apple’s silicon team didn’t deliver. The M4 uses half the power of the M2 at the same performance level.
Side note: I believe Moore’s Law is actually about chip density (transistor count), that it should double every 18 months. And it’s not an actual law of physics was speculation based on how things were progressing at the time… also, it was said 50 some years ago I doubt he assumed it would go on forever.
(and… percentages!? You realize the larger a number gets the likelihood of sustaining percentage jumps naturally decreases. If you walk 100ft one day, then 200ft the next, you’ve increased your distance by 100%… but if you walk a mile one day then 1.5 miles the next you’ve “only” increased your distance by 50%. Which is actually a bigger increase in distance walked in a day? 100ft or over 2500ft? A8 -> A10 is 330 point increase. M2 -> M4 is 1200 point increase)
Last edited:
Conversations like this just make me look back fondly at the single CPU single core era and marvel how much they got done on such little computing power. Incredible stuff. Our software today hasn't improved 1000X even though our processing power has. Some bugs in some apps are worse than they were back then.
Today we have incredible amount of compute and people cry in ways very few people did back then. The more people have the more stuff they have to complain about. We were amazed just to get a 100Mhz bump on a 1Ghz machine.
What does it matter? There's no cooling so none of theses scores mean much as they won't be sustained for long due to thermal constraints. Fine for short bursts of general use, but for anything else like gaming when performance really does matter...
Normal people don’t wait for the next “Big” thing, because most people don’t care. Most people buy what they need, when they need it. If their current system works they don’t upgrade, there’s no reason to, until there is. This worrying about something better coming along is futile because something better will always come along. I just replaced my A9 iPad Pro last summer and bought the M2 iPad Pro - mainly because the display was cracked.
If I were you, I would probably buy the previous iPad Pro with an M2. You get the features you want and the M2 is going to be a huge improvement over your current iPad.
good bless the Celeron 300A.....overclock that to 450mhz by moving a jumper. 50% MHz gains!
Only with very basic applications. Most operating systems today are smart enough to break off code into separate tasks when possible (even able to determine which compute unit is more efficient to run the task). So even though an application may not have been designed to be multithreaded, they may actually make use of multiple cores.
Biggest example of that, is the fact the all user interface interaction runs on its own thread, separate from the application code.
I think the other poster is suggesting those short bursts might be less impressive on the cut down chip.
Ooops, my bad, I have a M1 Max, not an Ultra.
Anyway, this is a bare bones M4 here, and my M1 Max has 32 GPUs I think, so it makes sense that it kind of doubles the M4 Metal benchmark.
So I correct my original analysis:
Core for core, the M4 is still faster (45% of the performance using just 30% of the cores). This makes a little more sense now!
I don’t mean applying effects to thousands of images in batch. I mean editing as in making changes and adding effects per image. With a desktop os you can simply import as many images as your system can handle and switch between images during editing for real work being done. Try on an iPad and fail miserably. The file system is terrible trying to work off an external drive especially when u can’t eject the drive. iPad can do 90% of most workloads but you have to change your workflow habits to suit the iPadOS. I just use my MacBook and be done quickly. No real multitasking either. Can’t route audio sources like you can on a real macOS. I’m tired of all that power and the iPad is actually much slower than a MacBook that’s lower spec when actually using it for real workflows.
That's not the fault of iPadOS which is a very fast and efficient OS built on the same kernel.
It's the fault of apps, like the Files app you mention and also the ridiculously bad file launcher screen on the iPad version of Photoshop. File managers could be much better on iPadOS.
The apps, including File Managers, could do what you need but developers are more interested in touch friendly interfaces and working with one file at a time. They prefer you to do the batch processing and more complex tasks using their desktop apps.
See Adobe Bridge for example. No version on iPadOS.
I have many iPad pros and MacBooks. I'm not saying the iPad just sucks in general. It has its use but for that bit of work that needs to be done efficiently, it just doesn't work because IT IS DUE TO iPadOS. The apps are limited BECAUSE of this iPadOS. Apple just trying to make you buy two devices and they have succeeded many times over. The iPad has its own way to do work on but you have to change the way your workflow is just to suit the iPad. It's just not as efficient as with a laptop. Another CEO would probably merge the two devices one day.
Yes, but all this conversation has been done and dusted for a decade so it is better to stop repeating. An iPad workflow is an iPad workflow. A Mac workflow is a Mac workflow. Hell even on the macOS side the way people work on different Macs can differ greatly.
Does Moore's Law even apply to performance increase? I thought it simply said the number of transistors will double every two years or so.
Sir I could only OC the Celeron from 333 to 409Mhz before meltdown occurred but boy that was a good value meltdown.
"Moore's law is the observation that the number of transistors in an integrated circuit (IC) doubles about every two years."