Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Hmm, $400 extra for an extra performance core and twice the ram. But I don't need 1tb, 512 is plenty.... meh, it's an iPad not a Mac. I'm happy! If I need P cores I'll use all 10 on my Mac! Maybe one less P core will help with battery and heat anyway 🤔
 
  • Like
Reactions: Nozgog
Yes. M4 is a M3 with faster NPU performance.

Samsung is all over the place saying that S24 has an IA hardware. Now, Apple improved its NPU engine to say "mine are faster than yours".

It's a refined M3, with better memory bandwidth, better per core optimizations (we have to see how this will translate in benchmarks) and a couple more efficiency cores.

The NPU performance is more than twice faster.
 
Releasing M4 ahead of the perceived schedule makes sense. Lower demand for iPad gives them time to get yields up for release in Mac. If it’s a chip related to iPhone there’s benefit there too
I just about to say the same thing. With a new processor it takes a while to get yields where they need to be for a larger product category like laptops. Yield is key for profitability in this business.

I’d have to go back and check but hasn’t the iPad often been the first with new apple silicon and things like new WiFi standards
 
  • Like
Reactions: IllegitimateValor
If that happens, the M2 and M3 really were delayed and now they are back on track for a 12 month release every WWDC, which makes perfect sense.

I also got the feeling the iPad will get macOS support with the next OS. Why else do the new mac-like keyboard?
Will be interesting to see - I would already be happy with an improved iPadOS …
 
Looks like the biggest functional changes over the M3 are in the display controllers, and moving to adding one or two efficiency cores.

The efficiency cores take up much less space the the performance cores, which is why I expect we won't see big jumps in performances cores for M4 (save some possible M4 Ultra.)
Maybe they will make a Logic Pro for Mac used the efficiency cores, currently it only uses the performance cores. Same with Ableton Live it only uses the performance cores.
 
But the other poster stated they made the switch to a cheaper node from a more expensive node? Are you agreeing or disagreeing with them? We know it's expensive to design chips, but the discussion is about saving money on build costs.
Switching nodes requires redesign and retesting. The node that Apple was using was too expensive in production because of yield problems, but it was the best one when M3 was designed. With M4, they switched to a newer node that was ready for M4 and has better yield and is an improvement.
 
Put the PC crowd to bed? Lol. What does this even mean?
Thats the usual stuff I come here to see. Some comment about how a product they can’t quite get yet (m4) is superior to another product they don’t have (snapdragon arm pc), and being absolutely confident in the apple Mx always being better.
 
  • Like
Reactions: beermode
Hmm. No direct comparison with M3, but 50% faster than M2 should mean it’s about 20% faster than M3 if memory serves me.

But a 50% jump in 2 CPU generations is quite impressive. Intel would be rubbing their hands with glee if they could manage that.
 
  • Like
Reactions: midkay
Hmm. No direct comparison with M3, but 50% faster than M2 should mean it’s about 20% faster than M3 if memory serves me.

But a 50% jump in 2 CPU generations is quite impressive. Intel would be rubbing their hands with glee if they could manage that.
The M3 were 50% faster than M1 as well.
 
It takes years to roll out new silicon. I'd be surprised if Apple has an internal cadence lower than 2 years.

Was this known a year ago? Then they might have had the time to do something. If not, then no. (Also note that the exploit being published isn't the same thing as manufacturers getting to know about it, most likely they get that information beforehand).
Apple received the information about the goFetch exploit from the security researchers on 5 Dec 2023. There is also an older, albeit less sophisticated but related exploit of the same component in the SoC that was disclosed back in 2022.
 
  • Like
Reactions: NetMage
Switching nodes requires redesign and retesting. The node that Apple was using was too expensive in production because of yield problems, but it was the best one when M3 was designed. With M4, they switched to a newer node that was ready for M4 and has better yield and is an improvement.
I agree with all of that. Did you think I didn't? I'm still confused. 😅
 
Some sources say 33-50%. $10 might be a bit low, but still would add up to quite a lot of money.
Wow. Maybe they brought the cost of M4 down below the cost of M1 and 2? Impressive. I'm guessing they had production problems with M4, though, if the vast majority of chips will be binned to only 3 cores. Unless of course they're just using the crap up on iPad Pro and will save the good chips for Mac...
 
Hopefully the AI capabilities will be more than just hype.
Uh, they DEMO'd the AI functionality (at least some of it). Yes, it's more than just hype.
Go look at the video, especially the Final Cut and Pro Logic sections; also the camera section.
 
  • Like
Reactions: Bodhitree
Assuming same clock: single-core will be the same, and multi-core will only go up <5% if that.

Apple touted better per core optimization, so single-core should be improved as well. Anyway, will have to see if the clock rate in the iPads has been lowered for thermal reasons.
 
  • Like
Reactions: SomeSoftwareGuy
So, M4 vs M3, the M4 has:

A better display engine.

Better GPU, the M3 marketing slides mentions everything from 20% to 2.5x up from M2, so I'm assuming the 2.5x is related to ray trancing. Which begs the question if the 4x is "just" better ray tracing HW or something else?

More E cores, and probably some other CPU related improvements. Since the M3 was 20% better than the M2, again according to marketing materials, and while 10 cores is 25 % more than 8, and 1.2*1.25 = 1.5. I really, really doubt that's what they mean.

Better NPU, they probably started supporting lower precision ops, which is where a lot of that boost comes from. Which is fine, as inference doesn't need anymore anyhow. If a bit sneaky.

Better memory BW. This may or may not translate to more usable BW (as per the M1 Pro/Max), but I'd be surprised if it wasn't (usable).

Less power usage.

This omits very technical improvements that are possible, but difficult to explain to the public.
The most interesting of these would be the ability to tie together different processing blocks so that latency between them is much much lower.
Right now obviously you CAN move work from the CPU to the GPU to the ANE and back to CPU, but the cost for each transition is disappointingly high (so that, for example, if you have less than about a million points to process, it's almost never worth shifting a task from CPU to GPU).

A combination of technologies (including, eg, faster context switching; and having the OS synchronize scheduling of CPU, GPU, and ANE) could dramatically improve performance, beyond the obvious numbers.
If you could, eg train an NPU using the CPUs, GPUs, and ANE all simultaneously (something that right now you usually don't want to do because of the costs of data transfer and delays while you wait for the new target to finish what it was doing and switch to your task) this would be a very nice speed boost.
Similarly if it were much cheaper to move code from CPU to GPU and back, using each for what it does best.

Does M4 have such improvements? Who knows, but if I were inside Apple, that's one of the "company wide" projects I'd be pushing, needing involvement from the GPU and ANE teams, and on the OS side.
 
It’s a very minor update overall (from m3; for MacBook/studio folks) and a substantial one from m2.

They basically took the M3 redesigned CPU/GPU architecture and slapped on a new display engine and the A17 Pro’s neural engine which runs already at 35 Tops. People forget Apple has had “AI” logic in their chips for ages. The only real new “AI”
For them will be generative AI. They’re already well positioned to support that on existing hardware.

It also seems they increase the core counts or the change in the the increase in CPU speed, including the increase in core counts over M2 is about the same as M3 was considering the same amount of core increases. They may have been some slight speed improvements.

So what you’re looking at for an M3 with a better neural engine.
so basically, M3 with AI, not really an "M4" if you really expect a revolutionary new architecture based on the naming scheme?
 
If that happens, the M2 and M3 really were delayed and now they are back on track for a 12 month release every WWDC, which makes perfect sense.

I also got the feeling the iPad will get macOS support with the next OS. Why else do the new mac-like keyboard?
macOS support for iPadOS is a BIG call. That would introduce sideloading, which would cut out Apples app store cash cow rort. I can't see it happening, well at least without macOS-for-iPad being heavily modified to prevent sideloading, which would mean there would be no terminal.
 
careful with the storage...you get binned M4 with the lowest storages , this is a first for the M ipads
I’m betting the higher storage are binned M4 binned as well, compared to what go into the MacBook Air, and those are binned M4 pros, and those are just binned M4 Max.
 
  • Like
Reactions: FriendlyMackle
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.