Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
How your computer will be worse because m2 exist?

I just bought the 14 pro, I don’t care if they release the 14 m2 pro tomorrow. Mine will still be as fast as yesterday and will gives me many many years of usage.

And for 95% of buyers the M2 will be no different in real world use than the M1 pro. Mark my words.
 
No, that’s not the same as having more cores or purpose-designed multi-processor systems, and needs specially written software - but things like ”compiler farms” and “render farms” are already common practice, and Apple used to have a product called XGrid to manage that sort of thing.

Sure, but they're never the preferred choice, and the low bandwidth does limit your use cases to:

  • requires massive computing power (the M1 Ultra's cores are not enough)
  • does not require much synchronization during the processing (otherwise, your limited bandwidth is gonna kill any advantages of the computing power)
  • is frequent enough that not simply renting a few nodes at AWS or Azure isn't the better alternative
 
Username checks out!

No, they're right. You buy something when you need it.

Yes, it's painful when you buy something and something better comes along, but that will inevitably happen sooner or later. If you can wait for the better thing, wait.
 
  • Like
Reactions: SirKeldon
I think Apple will just have a two prong approach to new hardware.

Consumer products (Mini, MB Air, iMac) will always be the first hardware to get the newest Apple silicon. They will keep these at 8-10 CPU cores and limit the SSD and memory capacities (16gb ram and 2TB SSDs) on these to steer pros towards their more expensive products. Apple's main focus for these will be battery life & thermal/power efficiency. These products will get very nice speed increases with each generation as Apple silicon gets better, but not enough to overtake the previous pro products that have much higher CPU/GPU core counts, increased memory capacity/bandwidth, and faster SSDs.

Pro products (Studio, MB Pro, Mac Pro) will have much higher CPU and GPU core counts and will be the last products to get each generation of Apple silicon. Because of the scaling of Apple silicon with more CPU/GPU cores, these will always be faster than the next generation or two of consumer hardware.

Updating their products in this way means everybody wins. Consumers get a lot better performance every generation, and the pros buying the more expensive hardware don't have it invalidated as soon as there is a new generation of Apple silicon. On Apple's side, they are making the smaller/simpler consumer chips while the new architecture/fab process is still young, and then making the larger/more complex chips when its more mature, which should increase yields across the board.
 
I think Apple will just have a two prong approach to new hardware.

Pro products (Studio, MB Pro, Mac Pro) will have much higher CPU and GPU core counts and will be the last products to get each generation of Apple silicon. Because of the scaling of Apple silicon with more CPU/GPU cores, these will always be faster than the next generation or two of consumer hardware.
I wonder how close M2 8-core/10-core will be to M1 Pro 8-core/14-core in terms of CPU speed.
 
I wonder how close M2 8-core/10-core will be to M1 Pro 8-core/14-core in terms of CPU speed.


M2 4P+4E CPU cores versus M1 6P+2E CPU ?

In multithreaded is that even a question?

If single threaded drag racing is the only thing that matters then the 8 cores are really immaterial. It is really 1P versus 1P.

But holistically... probably a loss if have workload worthy of buying a M1 Pro for. The 4E's in M2 probably gets a 1P worth of MT and so have 5 vs 6+ ( and probably not much of a frequency gap once get into extended computation and the M1 Pro has better cooling solution attached. )
It is a 16% drop from 5 to 6.

If M2 has matching LPDDR5 even more of a level playing field. ( m1 would have higher nominal aggregate bandwidth to offset any improvements in gen 2 architecture. ).

The M2 beating a binned down -25% Pro isn't earth shattering even if it does eek out a win. It won't touch the non binned down version.
 
Last edited:
M2 4P+4E CPU cores versus M1 6P+2E CPU ?

In multithreaded is that even a question?

If single threaded drag racing is the only thing that matters then the 8 cores are really immaterial. It is really 1P versus 1P.

But holistically... probably a loss if have workload worthy of buying a M1 Pro for. The 4E's in M2 probably gets a 1P worth of MT and so have 5 vs 6+ ( and probably not much of a frequency gap once get into extended computation and the M1 Pro has better cooling solution attached. )
It is a 16% drop from 5 to 6.

If M2 has matching LPDDR5 even more of a level playing field. ( m1 would have higher nominal aggregate bandwidth to offset any improvements in gen 2 architecture. ).

The M2 beating a binned down -25% Pro isn't earth shattering even if it does eek out a win. It won't touch the non binned down version.
It wasn’t really about if M2 would eek out a win over binned M1 Pro (multi-core). The question was more about how close M2 would get.

Would binned M1 Pro be 15-20% faster than M2? Or 0-5% faster? If the latter was true, then the statement made in the post I was responding to — “Because of the scaling of Apple silicon with more CPU/GPU cores, these will always be faster than the next generation or two of consumer hardware” — doesn’t really ring true.

PS. At this point I’m guessing M2 non-Pro will have LPDDR4X, just because.
 
I think 2 years is a more likely cycle length. That allows them to stagger M2/Pro/Max/Ultra releases and prevent consumer overload from refreshing the whole Mac line at once.
I don't know what cycle length Apple will choose but, when it comes to consumer overload, I'd say the opposite. Having a regular, annual update cycle for each device (like they have for the iPhone) provides the kind of stability and predictability that simplifies things for consumers.

The only type of consumer that might be "overloaded" by annual refreshes of every Mac device would be someone who owns every Mac device and feels obliged to replace them every time there's a refresh. Such consumers, if they exist at all, are at the extreme margins.

The most typical consumer, I'd guess, owns one or two Mac devices, and doesn't update them at the same time. Thus suppose it's June 2025 and I need to update my MacBook Pro. Knowing the MacBook Pro is refreshed, say, Nov. of each year simplifies my planning (compared to a 2-year cycle where I have to determine whether it's a 6-month or 18-month wait and, in the latter case, whether I want to wait that long). And what's going on with refreshes of the other devices is irrelevant to me.

If anyone's life is made easier by a 2-year staggered cycle, it's that of Apple itself--having to update products on a 2-year cycle is easier than a 1-year cycle. Though with Apple's resources, it seems offering a yearly update of at least the chip should be reasonable (as opposed to redesiging the product as a whole) (assuming a new generation of architecture is in fact available each year, which seems to be the case now, as evidenced by the iPhone's history).
 
Last edited:
  • Like
Reactions: Houpla and ric22
With the supply chain issues, they can announce M2 variations all they want, as long as they clear out current back orders first. At the rate they’re going, you’d be lucky to get an M2 in 2024.
 
With the M2, I'd like to see the maximum RAM options on the M/Pro/Max/Ultra increased from their current 16/32/64/128 GB to 32/64/128/256 GB.

Right now, if you're buying the Studio, and want 128 GB RAM, you have to spend the extra $1400 to upgrade to the Ultra even if you don't need the extra CPU/GPU cores (that's in addition to the extra $800 for the RAM itself). At least it's not so bad with the Pro/Max—there the upcharge to go to the Max (needed if you want 64 GB RAM) is only $200.
 
Last edited:
It wasn’t really about if M2 would eek out a win over binned M1 Pro (multi-core). The question was more about how close M2 would get.

Would binned M1 Pro be 15-20% faster than M2? Or 0-5% faster? If the latter was true, then the statement made in the post I was responding to — “Because of the scaling of Apple silicon with more CPU/GPU cores, these will always be faster than the next generation or two of consumer hardware” — doesn’t really ring true.

PS. At this point I’m guessing M2 non-Pro will have LPDDR4X, just because.
If we just pick something like Geekbench as our benchmark we can pretty easily see where you could slot the m2.


M1, m1 pro, and m1 max all have similar single core scores, around 1750ish. The M1 in multicore performance is around 7800, and the M1 pro is around 12400. Put the M2 base at around 10000 and you have a chip that’s 50% better than the M1 but still doesn’t eclipse the M1 pro.
 
With the M2, like to see the maximum RAM options on the M/Pro/Max/Ultra increased from their current 16/32/64/128 GB to 32/64/128/256 GB. Right now, if you're buying the Studio, and want 128 GB RAM, you have to spend the extra $1400 to upgrade to the Ultra in even if you don't need the extra CPU/GPU cores (that's in addition to the extra $800 for the RAM itself). At least it's not so bad with the Pro/Max—there the upcharge to go to the Max (needed if you want 64 GB RAM) is only $200.
I doubt there is any real market for people who want a ton of ram but don’t want the better CPU too.
 
I doubt there is any real market for people who want a ton of ram but don’t want the better CPU too.
Some of the calculations done by engineers and scientists are difficult to parallelize (and thus run on one core), but need lots of RAM.
 
  • Like
Reactions: Zorori
It wasn’t really about if M2 would eek out a win over binned M1 Pro (multi-core). The question was more about how close M2 would get.

Would binned M1 Pro be 15-20% faster than M2? Or 0-5% faster? If the latter was true, then the statement made in the post I was responding to — “Because of the scaling of Apple silicon with more CPU/GPU cores, these will always be faster than the next generation or two of consumer hardware” — doesn’t really ring true.

I mean, it depends on your workload? Given the gains in the A15 over A14, single-threaded performance will probably go up something like 9%, if even that, assuming they bump the clock to 3.5 GHz. If they leave it the same, there will be almost no gain at all — in the p cores, that is.

The e cores are better, and the GPU cores are way better.

But also, if this chip does have more p cores, multi-core performance would be up a fair bit.

PS. At this point I’m guessing M2 non-Pro will have LPDDR4X, just because.

A15 has LPDDR4X, so, yeah, probably.

With the M2, I'd like to see the maximum RAM options on the M/Pro/Max/Ultra increased from their current 16/32/64/128 GB to 32/64/128/256 GB.

Right now, if you're buying the Studio, and want 128 GB RAM, you have to spend the extra $1400 to upgrade to the Ultra even if you don't need the extra CPU/GPU cores (that's in addition to the extra $800 for the RAM itself).

Probably not happening until RAM density gets increased. The M1 Pro/Max already have LPDDR5.

I doubt there is any real market for people who want a ton of ram but don’t want the better CPU too.

I'm in that market. Lots of software development doesn't benefit that much from many cores. RAM and disk are the bottlenecks.

Some of the calculations done by engineers and scientists are difficult to parallelize (and thus run on one core), but need lots of RAM.

Yep.
 
  • Like
Reactions: theorist9
I doubt there is any real market for people who want a ton of ram but don’t want the better CPU too.
I'm one of those people. Base CPU is enough, but I do require lots of ram running docker containers and other virtualized stuff.
 
Apple is completely in the driver seat. No longer waiting around on Intel for gains.
Apple has been dragging its feet 'in the driver seat' for a year and a half now. The releases have been ridiculously slow, slower than in Intel era. They could learn a thing from AMD and Intel, for example to release high-end CPUs first and slower CPUs later. Somehow AMD & Intel get that, Apple doesn't.
 
If we just pick something like Geekbench as our benchmark we can pretty easily see where you could slot the m2.


M1, m1 pro, and m1 max all have similar single core scores, around 1750ish. The M1 in multicore performance is around 7800, and the M1 pro is around 12400. Put the M2 base at around 10000 and you have a chip that’s 50% better than the M1 but still doesn’t eclipse the M1 pro.
I was talking specifically about binned M1 Pro, and I did mention that in my posts.

Entry-level M1 Pro is the binned 8/14 variant. If we use your Geekbench 5 multi-core benchmark for comparison, it's right about that 10000 score you listed for M2:


Screen Shot 2022-04-19 at 8.17.54 AM.png



Truthfully I think a 10000 score for M2 is optimistic, but it may be in the same ballpark. If true then that means the entry level Pro CPU is only as fast as the next generation's entry level non-Pro CPU.
 
Last edited:
They could learn a thing from AMD and Intel, for example to release high-end CPUs first and slower CPUs later. Somehow AMD & Intel get that, Apple doesn't.

That's something Intel only does when their laptop releases are mediocre, such as with Alder Lake. It's absolutely not what they did with Broadwell, Ice Lake, or Tiger Lake.

And even when they do go desktop-first, it's not "high end" — their workstation/server CPUs tend to arrive years later.
 
Apple has been dragging its feet 'in the driver seat' for a year and a half now. The releases have been ridiculously slow, slower than in Intel era. They could learn a thing from AMD and Intel, for example to release high-end CPUs first and slower CPUs later. Somehow AMD & Intel get that, Apple doesn't.
You realize its a transition period right? Sure I'd like to see their most performant parts, who wouldn't, but Apple placed a heavy emphasis on efficiency. While you may be waiting for the fastest chip, Apple is selling incredible amounts of the base m1.
 
It wasn’t really about if M2 would eek out a win over binned M1 Pro (multi-core). The question was more about how close M2 would get.

Would binned M1 Pro be 15-20% faster than M2? Or 0-5% faster? If the latter was true, then the statement made in the post I was responding to — “Because of the scaling of Apple silicon with more CPU/GPU cores, these will always be faster than the next generation or two of consumer hardware” — doesn’t really ring true.

PS. At this point I’m guessing M2 non-Pro will have LPDDR4X, just because.
If M2 uses A15 cores, then they would be at most 15% faster than the M1 cores.

With M2 retaining the 4 performance cores compared to the 8 performance cores if the M1 Pro, then the M2 should performance at about 60% of the M1 Pro, assuming both were using all performance cores. (I don't know if efficiency cores come into play here though the difference probably wouldn't be large.)
 
Apple has been dragging its feet 'in the driver seat' for a year and a half now. The releases have been ridiculously slow, slower than in Intel era. They could learn a thing from AMD and Intel, for example to release high-end CPUs first and slower CPUs later. Somehow AMD & Intel get that, Apple doesn't.
Intel usually releases the laptop level chips first and the higher performance chips later.

Apple has released 4 chips in the last 18 months and it looks like a new generation will launch in the second half of this year. That is some impressive foot dragging, there. :)
 
  • Like
Reactions: Zdigital2015
If M2 uses A15 cores, then they would be at most 15% faster than the M1 cores.

With M2 retaining the 4 performance cores compared to the 8 performance cores if the M1 Pro, then the M2 should performance at about 60% of the M1 Pro, assuming both were using all performance cores. (I don't know if efficiency cores come into play here though the difference probably wouldn't be large.)
You touched upon what I was getting at indirectly, although you missed the part where I was talking about binned M1 Pro.

8/14 binned M1 Pro CPU performance is roughly 30% faster than M1. If M2 is 15% faster than M1, that places it squarely in between M1 and M1 Pro. If M2 is say 25% faster than M1, then that gets it very close to M1 Pro CPU performance.

Not that it really matters for me personally, since I mainly just want more ports. I'm sounding like a broken record now, but I would love to see a middle ground with 6 USB ports. Right now you have the option of 4 with the Mac mini, or 8 with the much more expensive Mac Studio.

Why not just use a USB hub or dock? Because in my experience the reliability with USB devices plugged into hubs is nowhere near as consistent as it is with USB devices plugged directly into Macs. Some hubs/docks are better than others, but this still holds true, and the better hubs/docks are often 3-digit $ anyway. I'd rather just spend that money up front and get the extra ports in the Mac itself. Those who don't want to spend the money can get the entry level models.

As for performance, anything would be a major improvement for me. Right now my Mac mini's multi-core speed is slower than M1's single-core speed. :D
 
  • Like
Reactions: ric22 and Tagbert


Apple is internally testing several variants of the next-generation M2 chip and the updated Macs that will be equipped with them, reports Bloomberg, citing developer logs. There are "at least" nine new Macs in development that use four different M2 chips that are successors to the current M1 chips.

13-inch-macbook-pro-m2-mock-feature-2.jpg

Apple is working on devices with standard M2 chips, the M2 Pro, the M2 Max, and a successor to the M1 Ultra, with the following machines in the works:
  • A MacBook Air with an M2 chip that features an 8-core CPU and 10-core GPU.
  • A Mac mini with the M2 chip and a variant with the M2 Pro chip.
  • An entry-level 13-inch MacBook Pro with M2 chip.
  • A 14 and 16-inch MacBook Pro models with M2 Pro and M2 Max chips. The M2 Max chip features a 12-core CPU and 38-core GPU, along with 64GB Memory.
  • A Mac Pro that will include a successor to the M1 Ultra used in the Mac Studio.
Apple has also tested an M1 Max version of the Mac mini, but the release of the Mac Studio may make such a machine redundant, so Apple could stick with M2 and M2 Pro chips when the Mac mini eventually sees a refresh.

According to Bloomberg, the internal testing is a "key step" in the development process, and it suggests that the machines could be released in the coming months. We have heard multiple rumors about a new MacBook Air, an updated 13-inch MacBook Pro, a Mac Pro, and a new Mac mini, but this is the first we are hearing of a possible 14 and 16-inch MacBook Pro refresh this year.

Prior rumors have suggested that we can expect to see the MacBook Air, low-end MacBook Pro, and Mac mini come out in 2022, and Bloomberg has previously said that at least two Macs will launch mid-year, perhaps at WWDC.

Article Link: Apple Testing at Least Nine New Macs With Four Different M2 Chip Variants
What's important for me is whether the M2 will finally support HDMI 2.1 and 8k at 60Hz. If it does, I can see a Mac mini with M2 Pro in my future replacing my 2018 Intel.
 
What's important for me is whether the M2 will finally support HDMI 2.1 and 8k at 60Hz. If it does, I can see a Mac mini with M2 Pro in my future replacing my 2018 Intel.

Absolutely time for HDMI 2.1 on all new Macs

Should have been there on the M1s honestly
 
  • Like
Reactions: ric22
[Increased RAM] Probably not happening until RAM density gets increased. The M1 Pro/Max already have LPDDR5.
They could also increase RAM by increasing the number of RAM I/O lanes, but are you thinking that would take too much redesign?

As far as increased density goes, it looks like Samsung will be mass-producting 24 GB LPDDR5x RAM chips* in 2022; 32 GB chips* won't be available until 2024-2025 (see this article, published April 14, 2022: https://www.tomshardware.com/news/amd-and-samsung-webinar-talks-raphael-overclocking-ddr5-future ).

[*More technically, from what I understand, what I'm calling a chip is actually a stack: The chips will be 24 Gb; with 8 chips to a stack, this gives 24 GB.]

Those should be available for the 2023 MacBook Pro and Studio and, if used, would increase maximum memory for the Pro/Max/Ultra from 32/64/128 GB to 48/96/192 GB. [I think the Pro/Max/Ultra use 2/4/8 RAM chips respectively, and that the chip sizes are currently 8 and 16 GB, which explains why the Pro offers 16 or 32 GB, the Max 32 or 64, and the Ultra 64 or 128.]

1650418182759.png
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.