OMG so tempting to get the M4 IPad Pro with all the trimmings. I am holding out though to understand if the 8GB vs 16GB is of material importance or not for the future.
M4, that’s an interesting move. Especially considering the M3 isn’t even fully out yet with the Ultra chips (and hopefully the Extreme variant which the Mac Pro desperately needs to justify its existence).
Wonder why they’ve jumped so far forwards, it will simply put people off buying M3 Apple Macs, no one wants to buy last years tech at this years pricing.
M3 is much more expensive to produce and is less power efficient. Makes sense to make the move sooner rather than later. Have a feeling the Mac Studio and Mac Pro will skip the M3 and go straight to M4 this summer.
Woah.. NPU in M4 seems to be 2x of TOPS than M3... Memory bandwidth is also up 20%. GenAI here we go.
Can't wait to see an Ultra or Extreme version of this baby.
I'm really surprised that Apple decided to launch M4. M3 with support Tandem OLED could have worked. M1 to M2 and M2 to M3 were both exactly +25% in transistor count. M4 is a +12% bump and feels like a M3.5.
M4 is delivering A17 Pro (2023) level NPU performance. Is it enough for 2024/2025 outside iPad Pro? Intel Lunar Lake is aiming for 45 TOPS. Will M4 Pro and Max deliver more? With M4 on TSMC N3E the power consumption and performance could be worse than M3. It is however fantastic to see it this soon. Clearly this in not the time to buy a MBP, Studio, or Mac Pro.
So disappointing. Apple had to make it so thin. That means it has a much smaller battery than it could have had. OLED screens typically have higher power demands. In reality it will probably only get 6 hours or so.
Actually the Battery is larger on the new model and the Runtime specs are exactly the same as the previous model.
That’s a lot of power — I can’t help think that something else with be unveiled at WWDC to take advantage of this. Maybe I’m completely off,
Hello? Did you see the new Live Multicam capabilities with four iPhones?! I'm sure more is coming at WWDC, but that alone is INSANE.
HOWEVER... Mark Gurman really nailed the crux of the issue in his most recent Power Up column: Apple has intentionally held back the iPad from becoming a true laptop replacement by hobbling it in iPadOS. How else to explain that they can pull off a near miracle like Live Multicam but the simple task of multitasking on an iPad remains a kludge? This much is clear: the new iPad Pros, at first blush, might be the best hardware Apple has ever produced--they are at least in the running for that title. But will Apple FINALLY provide an iPadOS as good as the hardware? We'll soon see... If they do, I'd trade in my MBA M2 immediately.
Thermal ceiling is much lower in a device as thin as the new iPad Pro… Limiting the OS in how much it can do at once, will still provide the perceived amount of performance users expect. They actually mentioned how they’ve attempted to increase the thermal ceiling to enable being able to make it thinner. If you attempted to run a completely unrestricted OS, like macOS on this device, you would see a major hit to performance as the SoC got hotter and hotter.
OMG so tempting to get the M4 IPad Pro with all the trimmings. I am holding out though to understand if the 8GB vs 16GB is of material importance or not for the future.
Yep, waiting for WWDC announcements before spec'ing my new iPad Pro.
That’s a lot of power — I can’t help think that something else with be unveiled at WWDC to take advantage of this. Maybe I’m completely off,
Hello? Did you see the new Live Multicam capabilities with four iPhones?! I'm sure more is coming at WWDC, but that alone is INSANE.
HOWEVER... Mark Gurman really nailed the crux of the issue in his most recent Power Up column: Apple has intentionally held back the iPad from becoming a true laptop replacement by hobbling it in iPadOS. How else to explain that they can pull off a near miracle like Live Multicam but the simple task of multitasking on an iPad remains a kludge? This much is clear: the new iPad Pros, at first blush, might be the best hardware Apple has ever produced--they are at least in the running for that title. But will Apple FINALLY provide an iPadOS as good as the hardware? We'll soon see... If they do, I'd trade in my MBA M2 immediately.
Apple is in league of their own they don't need to double back and do what Microsoft is doing with the Surface, new software and upgraded software will be introduced to take advantage of another step forward for Apple who has again cemented the iPads place as the best tablet in the world.
Note: Samsung tried to talk Apple out of tandem OLED but LG stepped up and said yes we can and they are building the screen for the 13" model.
Yep. It's the best tablet in the world. And the largest-selling. And the iPad (as well as the Mac) have been struggling with growth. All those things can be true at the same time. And since Apple stock is priced as a growth company, lack of growth in major product lines is a big problem. Apple doesn't need to replicate what Surface does--but it basic computing functions like multi-tasking shouldn't be a kludge. I'm very anxious to see what awaits in the iPadOS unveiled at WWDC.
Maybe a combination of M4 efficiency and new thermal structure was required before giving the iPad more Mac-like capabilities? Really looking forward to WWDC!
So disappointing. Apple had to make it so thin. That means it has a much smaller battery than it could have had. OLED screens typically have higher power demands. In reality it will probably only get 6 hours or so.
Actually the Battery is larger on the new model and the Runtime specs are exactly the same as the previous model.
2024 iPP13 battery is 39 WHr 2022 iPP13 battery is 41 WHr
2024 iPP11 battery is 31 WHr 2022 iPP11 battery is 29 WHr
2024 iPA13 battery is 37 WHr 2024 iPA11 battery is 29 WHr
But like you said, runtime is the real metric, and Apple changes battery sizes to try to hit 10 hrs of light web browsing as well as they can.
Well, as usual I spent a lot of money I didn’t have to. So I bought the whole thing, 13”, keyboard and Pencil and also the nano glass. I hope the glass isn’t too difficult to clean as my iPad screen gets dirtier than my Apple monitor screen. ... anyway, I’m pretty excited to see this new one. It should be here between the 16th and the 20th. My wife’s, which doesn’t have the nano screen and 512GB instead of 1TB, will be here the 15th.
Can you give us a report on the crispness, clarity and contrast of the two, nano-texture versus smooth? And the difference in feel with the Pencil?
M4, that’s an interesting move. Especially considering the M3 isn’t even fully out yet with the Ultra chips (and hopefully the Extreme variant which the Mac Pro desperately needs to justify its existence).
Wonder why they’ve jumped so far forwards, it will simply put people off buying M3 Apple Macs, no one wants to buy last years tech at this years pricing.
WWDC2024 should clarify. I wouldn’t buy an M3 Mac between now and then.
Just the ability to do nano texture on a touchscreen is very interesting. I'm drooling a bit, but not upgrading anytime soon.
Yeah, how is that going to work? What if it starts to wear off unevenly? Maybe they have refined it so it’s not so fragile as on the displays? That would be amazing!
OMG so tempting to get the M4 IPad Pro with all the trimmings. I am holding out though to understand if the 8GB vs 16GB is of material importance or not for the future.
I’ve never once thought “gee, I wish I had less RAM”
I have a 2018 Intel MacBook Pro, and I might be fine with a MacBook Air M4 rather than a Pro laptop. Although Mr. Cook probably will push me on his spec ladder, with his main weapon: storage and memory.
The recent announcement makes it more likely that Mac Studio will have M4 chip. I think they wanted M4 for Mac Studio for a few reasons - it is expensive to produce Mx Ultra and the improved 3nm process makes it more economical to scale for bigger chips and at same time, improve performance. Also, I suspect they also went M4 to patch security vulnerabilities in M1-M3 but of course, Apple could somehow patch it with a chipset to block attacks in real time for older M chips in new devices.
I wonder if they are going to launch the M4 Macs soon. Given that it is available in the iPad Pro I wouldn’t be surprised if some Mac purchasers are going to want to wait for the M4.
I just ordered one at close to the base price for the 13" model with my trade-in but including Apple Pencil and Magic Keyboard. I would have upgraded to Nano Glass and 16GB of RAM if I could have gotten both for a total of $200 more. $700 more is too much. I don't need an extra core or an extra 744GB of storage.
There's a lot to be said for AI on the edge, in this case in the device in your hand. Great for photo editing and transcription, for example, and for optimising the performance of the device itself. NPUs (Neural Processing Units) will soon become mainstream components across every type of device from every major manufacturer.But the M4 announcement also highlights a major weakness of Apple. They still remain very much a device company. They have not matched the cloud computing power of Google, Azure and AWS. Powerful edge computing is interesting, but it needs more behind it.It's why Apple's maps offer is so poor, and it's why generative AI came first to Google and Microsoft productivity apps. Apple doesn't have the datasets either, which is why Siri depends on Bing for search.Apple has extraory product engineering that delivers lovely phones with beautiful interfaces, but it offers a shallow stack whose weaknesses are becoming more evident.
I was super surprised by Mark Gurman being right about the M4! Damn, that’s a big move for Apple. This then makes a lot of sense why the M3 ULTRA is missing, I bet Apple is just going to jump over it directly to an M4 MAX, ULTRA, and maybe EXTREME to be revealed at WWDC. Can’t wait!
Woah.. NPU in M4 seems to be 2x of TOPS than M3... Memory bandwidth is also up 20%. GenAI here we go.
Can't wait to see an Ultra or Extreme version of this baby.
M4 is delivering A17 Pro (2023) level NPU performance. Is it enough for 2024/2025 outside iPad Pro? Intel Lunar Lake is aiming for 45 TOPS. Will M4 Pro and Max deliver more?
The TOPs numbers are not quoted the same way by different manufacturers. If they use INT8, they can double the number over FP16, if they use INT4, they can quadruple the number.
Nvidia claims their 4090 is 1321 TOPs but in practise is only 2-3x faster than M3 Max on AI tasks.
Some manufacturers are counting the whole device CPU + GPU + NPU, others just the NPU.
Apple's Neural Engine page lists iPhone numbers as FP16:
Further confusing the issue is there are numbers that include the use of sparsity, which is a technique for compressing matrices with unused values so they can fit more in a process at the same time and they double the numbers again:
For the Nvidia 3090, it's described as 35TFLOPs FP32, 71TFLOPs FP16, 284TOPs INT8, 568TOPs INT8 with sparsity, 568TOPs INT4, 1136TOPs INT4 with sparsity.
If Apple did the same, they'd go 144TOPs INT8 -> 288TOPs INT4 -> 576TOPs INT4 w/sparsity. They can put it on a slide saying 576 AI TOPs but it's not a meaningful number.
For marketing, companies like Nvidia promote the highest numbers because they are trying to convince big data center contractors to go with their product over AMD/Intel and the latter do the same.
Intel could easily be quoting INT4 with sparsity numbers, in which case divide them by 8 when comparing numbers.
It's not entirely wrong for them to use lower precision numbers because an AI model might allow using them but you can't compare different precision values to make assessments about new hardware, you have to compare the performance at the same precision.
Comments
M3 is much more expensive to produce and is less power efficient. Makes sense to make the move sooner rather than later. Have a feeling the Mac Studio and Mac Pro will skip the M3 and go straight to M4 this summer.
M1 to M2 and M2 to M3 were both exactly +25% in transistor count. M4 is a +12% bump and feels like a M3.5.
M4 is delivering A17 Pro (2023) level NPU performance. Is it enough for 2024/2025 outside iPad Pro? Intel Lunar Lake is aiming for 45 TOPS. Will M4 Pro and Max deliver more? With M4 on TSMC N3E the power consumption and performance could be worse than M3. It is however fantastic to see it this soon. Clearly this in not the time to buy a MBP, Studio, or Mac Pro.
Thermal ceiling is much lower in a device as thin as the new iPad Pro… Limiting the OS in how much it can do at once, will still provide the perceived amount of performance users expect. They actually mentioned how they’ve attempted to increase the thermal ceiling to enable being able to make it thinner. If you attempted to run a completely unrestricted OS, like macOS on this device, you would see a major hit to performance as the SoC got hotter and hotter.
2022 iPP13 battery is 41 WHr
2024 iPP11 battery is 31 WHr
2022 iPP11 battery is 29 WHr
2024 iPA13 battery is 37 WHr
2024 iPA11 battery is 29 WHr
But like you said, runtime is the real metric, and Apple changes battery sizes to try to hit 10 hrs of light web browsing as well as they can.
I have a 2018 Intel MacBook Pro, and I might be fine with a MacBook Air M4 rather than a Pro laptop. Although Mr. Cook probably will push me on his spec ladder, with his main weapon: storage and memory.
Nvidia claims their 4090 is 1321 TOPs but in practise is only 2-3x faster than M3 Max on AI tasks.
Some manufacturers are counting the whole device CPU + GPU + NPU, others just the NPU.
Apple's Neural Engine page lists iPhone numbers as FP16:
https://machinelearning.apple.com/research/neural-engine-transformers
M3 Max has a 14TFLOPs FP32 GPU, which is theoretically 28TFLOPs FP16 and 56TOPs INT8.
If the M4 Neural Engine is 38TOPs FP16, then it's 76TOPs INT8 and an M4 Max chip would be 17TFLOPs FP32 = 68TOPs INT8 so M4 Max total = 144TOPs INT8.
Nvidia's claim suggests the 4090 is nearly 10x faster but clearly isn't and the following page shows numbers with INT4:
https://wccftech.com/roundup/nvidia-geforce-rtx-4090/
Further confusing the issue is there are numbers that include the use of sparsity, which is a technique for compressing matrices with unused values so they can fit more in a process at the same time and they double the numbers again:
https://blogs.nvidia.com/blog/sparsity-ai-inference/
For the Nvidia 3090, it's described as 35TFLOPs FP32, 71TFLOPs FP16, 284TOPs INT8, 568TOPs INT8 with sparsity, 568TOPs INT4, 1136TOPs INT4 with sparsity.
If Apple did the same, they'd go 144TOPs INT8 -> 288TOPs INT4 -> 576TOPs INT4 w/sparsity. They can put it on a slide saying 576 AI TOPs but it's not a meaningful number.
For marketing, companies like Nvidia promote the highest numbers because they are trying to convince big data center contractors to go with their product over AMD/Intel and the latter do the same.
Intel could easily be quoting INT4 with sparsity numbers, in which case divide them by 8 when comparing numbers.
It's not entirely wrong for them to use lower precision numbers because an AI model might allow using them but you can't compare different precision values to make assessments about new hardware, you have to compare the performance at the same precision.