yes, i still think M2 Macbook air is worth just for the base config, when you start to add up you are better with the 14" Mbp until the M2 Macbook Air gets down to the usually $999 probably next year when they will remove the M1 Mba
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.
M2 Pro and M2 Max will be disappointing?
- Thread starter Zest28
- Start date
- Sort by reaction score
While many of us have been stressing over leaked Geekbench results for apparently imminent Macs, I thought it may be useful to hear the perspective of someone who is close to the industry. Rene Ritchie has taken some time away from his day job of shilling for Google, to conduct an interview with analyst Ben Bajarin, and has posted two parts of that interview on Youtube.
In the first section, Bajarin covers the challenges facing the semiconductor industry in general.
The entire clip is worth watching, but if you want to hear what he has to say about Apple specifically, then skip ahead to the 6:25 minute mark.
In the second part, Bajarin addresses the challenges facing individual companies.
Again, if you only want to see what he has to say about Apple, then skip ahead to the 12:35 mark.
He's bullish about Apple Silicon's future, lauding Apple's ability to scale their designs while dealing with global supply chain issues, which he believes will be resolved soon. As many knowledgeable folks here have pointed out, packaging is important to Apple's designs going forward, while continuing to take advantage of cutting edge process technologies, contributing to maintaining a performance per watt advantage.
For those of us who have been demoralized by the leaked Geekbench results, long-term he stated "everybody that competes in the silicon industry that competes with Apple...what keeps them up at night is Apple" and "almost every designer worth their salt recognizes how good [the] products [are] that Apple is putting out on a consistent basis" as well as "some people can shrink that gap, but not close it". Finally, he notes that "this philosophy of design positions them well for any category that they want to go in consumer electronics".
So, if industry analyst Ben Bajarin is correct, Apple's future designs have a bright future, one which their competitors may have great difficulty catching up with, if they ever do. Hence, we need not suffer a serious case of the vapors over a handful of leaked benchmarks which may or may not represent a final product. Even if Apple's next iteration of the M2 fails to impress, much of that can be attributed to setbacks from global issues that we are all aware of and which will likely be resolved relatively soon.
This is, of course, one man's opinion, take it for what it's worth, but I find his analysis encouraging for the future of Apple Silicon. I think this is a good reminder that the immediacy of the latest rumor does not necessarily represent a long-term downward trend. Most of us here are concerned about the Mac and our favorite fruit company, but it's vastly premature to count out Johny Srouji and his team inside Apple's secretive skunkworks.
In the first section, Bajarin covers the challenges facing the semiconductor industry in general.
The entire clip is worth watching, but if you want to hear what he has to say about Apple specifically, then skip ahead to the 6:25 minute mark.
In the second part, Bajarin addresses the challenges facing individual companies.
Again, if you only want to see what he has to say about Apple, then skip ahead to the 12:35 mark.
He's bullish about Apple Silicon's future, lauding Apple's ability to scale their designs while dealing with global supply chain issues, which he believes will be resolved soon. As many knowledgeable folks here have pointed out, packaging is important to Apple's designs going forward, while continuing to take advantage of cutting edge process technologies, contributing to maintaining a performance per watt advantage.
For those of us who have been demoralized by the leaked Geekbench results, long-term he stated "everybody that competes in the silicon industry that competes with Apple...what keeps them up at night is Apple" and "almost every designer worth their salt recognizes how good [the] products [are] that Apple is putting out on a consistent basis" as well as "some people can shrink that gap, but not close it". Finally, he notes that "this philosophy of design positions them well for any category that they want to go in consumer electronics".
So, if industry analyst Ben Bajarin is correct, Apple's future designs have a bright future, one which their competitors may have great difficulty catching up with, if they ever do. Hence, we need not suffer a serious case of the vapors over a handful of leaked benchmarks which may or may not represent a final product. Even if Apple's next iteration of the M2 fails to impress, much of that can be attributed to setbacks from global issues that we are all aware of and which will likely be resolved relatively soon.
This is, of course, one man's opinion, take it for what it's worth, but I find his analysis encouraging for the future of Apple Silicon. I think this is a good reminder that the immediacy of the latest rumor does not necessarily represent a long-term downward trend. Most of us here are concerned about the Mac and our favorite fruit company, but it's vastly premature to count out Johny Srouji and his team inside Apple's secretive skunkworks.
Oh damn, we have a surviving M2 thread! Shame about the title tho.
Rene Ritchie doesn’t deserve this ☹️
This doesn't appear to be true. Sure the M2 uses a bit more power than the M1 but it also gets more done using the equivalent power. Apple added a measly 2.7 W of additional battery in the M2 MacBook Air over the M1 MBA yet they both get about the same battery life according to Apple. I get about an 18% multi-core GB 5 performance improvement with the M2 MacBook Air but I haven't noticed any difference in battery life with only a 5% increase in battery capacity.
"Disappointing" in what regard? The OP doesn't provide a baseline for what's impressive or not, and I'm talking specifically about how your software runs.
These are just numbers that being thrown around here, and whilst yes they highlight some of the overall compute power that we can expect, what really matters is actual real world usage. Believing that you can judge how "good" or not an SoC is purely on numbers only tells half a story today. What matters is the overall hardware and software integration.
There are not many people (at least on these forums) complaining about the M1 Pro and Max being poor performers, and frankly no-one with a rational mind would believe that we're going to see substantial improvements year over year. And even if we did, would those professionals actually invest in new hardware regularly? The cost-benefit likely wouldn't add up. I can edit multiple streams of 4K video, Logic Pro projects, Photoshop documents and run a business from a device that barely gets warm and never runs its fan. Why would I be "disappointed" by the alleged M2 specs?
Also keep in mind that when we're seeing Intel Geekbench results, we're usually seeing them in a 'best case scenario', which is to say that if you ran the same chip in a Windows laptop that was as compact as a 14/16 MBP, it would rarely if ever hit peak performance - especially on battery power (which would last, what, 30 minutes?).
These are just numbers that being thrown around here, and whilst yes they highlight some of the overall compute power that we can expect, what really matters is actual real world usage. Believing that you can judge how "good" or not an SoC is purely on numbers only tells half a story today. What matters is the overall hardware and software integration.
There are not many people (at least on these forums) complaining about the M1 Pro and Max being poor performers, and frankly no-one with a rational mind would believe that we're going to see substantial improvements year over year. And even if we did, would those professionals actually invest in new hardware regularly? The cost-benefit likely wouldn't add up. I can edit multiple streams of 4K video, Logic Pro projects, Photoshop documents and run a business from a device that barely gets warm and never runs its fan. Why would I be "disappointed" by the alleged M2 specs?
Also keep in mind that when we're seeing Intel Geekbench results, we're usually seeing them in a 'best case scenario', which is to say that if you ran the same chip in a Windows laptop that was as compact as a 14/16 MBP, it would rarely if ever hit peak performance - especially on battery power (which would last, what, 30 minutes?).
Has it been mentioned the number of gpu cores the pro and max will have?
As for as I can tell the pro will have 19 cores and the max will have 38 cores. And are these the same cores as the M1 Pro/max or will they be updated cores?
I was planning on getting the M2 max 16" but after watching a few videos the M1 max already throttles at high uses so im guessing the M2 max could be worse in the same mbp body.
Im now thinking I should just get the m2 pro or maybe the binned m2 max if there is one for less heat and fan noise.
As for as I can tell the pro will have 19 cores and the max will have 38 cores. And are these the same cores as the M1 Pro/max or will they be updated cores?
I was planning on getting the M2 max 16" but after watching a few videos the M1 max already throttles at high uses so im guessing the M2 max could be worse in the same mbp body.
Im now thinking I should just get the m2 pro or maybe the binned m2 max if there is one for less heat and fan noise.
Last edited:
Revolutionary better cores? Probably not. 100% exactly the same as the M1 GPU cores ? Probably not. Minimally there are likely some quirky features that have incremental feature fixes for. What they are implemented in would likely have more impact on thermals than in structural changes in baseline design.
Geekbench and most of the tech press industry porn benchmarks are singularly focused on 'bigger numbers' (or faster (smaller) time numbers). they are mainly interested in the magnitudes as 'better'. They tend not to directly measure efficiency at all.
"Throttling" videos generate more ad views. What is more pertinent is if the workload you need to use would generate that kind of workload. It is a laptop. if throw tons of heavy workload at it then at some point the fans will kick in and it will run hot.
If Apple threw TSMC N3 process at M2 Max and didn't try to maximally goose the Geekbench numbers and mainly just take the power gains their the 'throttling" issue would mostly subside even if there were relatively modest increases in GPU core count. Apple would have a smaller die (more afforable than a same sized one) and better overall system battery life.
Instead of 8 core clusters the M2 seems to have 10. So 19 means Apple is holding one GPU core back (Pro would have two clusters so 20 cores present). Apple did same thing with A14X ( and later winked on the extra GPU core in A14Z ) . If they scaled the Max the same way then 38 (instead of 40 ). If had a thing for odd numbers that could be 37.
[ could be leaving these off for slightly better yield management. ] The CPU core count goes up to 12 but pretty good chance that is just a full sized E core cluster of 4. (8 P 4 E ).
Two more E cores isn't going to blow out the thermals either. Two more on TSMC N3 even less so.
The 12 core should have been the first iteration of M1 Pro. Why they only used 2 efficiency cores baffled me in the M1 Pro.
The battery life on the M1 Pro/Max MBP would have been a lot better with a quad efficiency cores setup.
I think the new M2 Pro/Max will be a significant upgrade to the M1. Better efficiency cores and more of them. Better GPU cores and probably more of them. Those two things alone will add a lot and about a 15% lift in CPU with better battery life is all very good.
Sure it may not blow away benchmarks but for what people actually use these laptops for it should be very good.
Sustainable performance under heavy load is a lot better than crazy high boost benchmarks for Pro level devices.
M3 will most likely be based on a16 which is a 4nm node and then M4 will be 3nm. All the rumors about 3nm don't follow precedent to me
The battery life on the M1 Pro/Max MBP would have been a lot better with a quad efficiency cores setup.
I think the new M2 Pro/Max will be a significant upgrade to the M1. Better efficiency cores and more of them. Better GPU cores and probably more of them. Those two things alone will add a lot and about a 15% lift in CPU with better battery life is all very good.
Sure it may not blow away benchmarks but for what people actually use these laptops for it should be very good.
Sustainable performance under heavy load is a lot better than crazy high boost benchmarks for Pro level devices.
M3 will most likely be based on a16 which is a 4nm node and then M4 will be 3nm. All the rumors about 3nm don't follow precedent to me
We don't know. What I can tell you that some alleged insiders claimed that Apple has been working on a new GPU, and that there are patents published in 2022 that describe more efficient shader execution as well as hardware raytracing. Given the fact that Apple tends to tie publishing of patents to actual hardware releases (e.g. patents describing some features of A15/M2 GPUs were published just few month before A15 was out), this would suggest that they have a new GPU core ready to go. A recent report claimed that A16 should have featured a new GPU but these plans had to be scrapped in the last minute due to power consumption concerns (as some have speculated this is because the new GPU as designed for 3nm and A16 ended up using 14nm).
It is possible that the M2 Pro/Max family will use this new GPU, as power consumption will be less of a concern in a larger laptop. As to the amount of cores or the likelihood that this will happen, I do not dare to make any predictions. Until now Apple's execution has been very conservative. If they go with the "safer" option of A15/M2 GPU cores we will get 20 cores on M2 Pro and 40 on M2 Max.
IMHO because the Max was too chunky. The CPU cores of the Pro and Max matched. The top 'half' of the Max was used for the Pro. Apple was so 'hell bent' to crank the GPU core count as high as possible on the Ultra that they ran out of room for a full 4 E cluster.
[ The Ultra using TSMC Info-LSI packaging which is limited by the 1x reticle size. So two Max had to fit inside of a fixed window size. If the Max grows too big then have to trim something. Two E cores isn't humungous but it is something if looking for incremental cuts. ]
If look at the Pro die all by itself... yeah it is goofy. Even the Max all by itself. The hole Apple dug for themselves is trying to press the laptop optimized Max die into a big , fat , too chunky chiplet role.
Secondarily I think they were trying to take some bandwidth workload of the internal bus also. Again the main objective was to feed the GPU system bandwidth monster first and sprinkle the other stuff around that.
only if the applications being run queue that up and most of the GPU is asleep. If the GPU is awake and extremely power hungry E cores aren't going to offset all that well.
Apple could skip the A16. The M3 could be mirroring A17 ( and a M2 Pro/Max/Ultra N3 port ). [ If Apple really has volume production problems with N3 then perhaps a 'clean up' of those initial forays onto N3B onto N3E. ]
The A16 is in part a 'stop gap' before N3. And likely a long term 'cheap' wafer solution for a bunch of 'hand me down' products below the iPhone that Apple will distribute to over the next couple of years. (plain iPad , AppleTV ,etc) It wouldn't something to pivot the M-series entry model on.
The A-series dogmatically always getting the most advance first doesn't make any sense over the long term. New nodes come out on 18-21 month like cycles. Not a rigid 12 month cycle. The entirely non-technological, artificial cycle that the iPhone is locked into is whacked in terms of staying in synch with fab process node updates. And there is little sane reason to put the M-series on a marketing department driven rigid 12 month cycle also.
M1 -> M2 was not 12 months. M1 Pro/Max -> M2 Pro/Max was not 12 months. It has already happened.
The A12X / A14X didn't before that ( iPad Pro SoC tended to skip the odd numbered A-series SoC).