Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
1. Physics, with a SOC like the M series the ram is built into the CPU and GPU so data has to travel much less distance significantly reducing latency and signal degradation

2. Wider memory buses for higher bandwidth, versus narrower interfaces in socked/upgradable setups limited by pin counts and traces.

3. Optimized, proprietary interfaces without socket overhead like inductance or standardized protocols that constrain speeds in upgradeable RAM setups.

Its like a normal engine that is designed for public use and therefore has to support a wide variety of fuels vs a highly tuned engine optimized to a specific fuel (race gas, E85, etc) it can squeeze out way more power but then you are stuck with whatever it is tuned for.

The speed differences are why GPUs have never been able to use the socketed RAM. Its just way too slow for GPU purposes, but with a SOC like the M series the GPU can use the same RAM as the CPU allowing the studio for example to have access to 512 GB of video ram (- some RAM for running the system etc) allowing it to run huge AI models that would require extremely expensive and power hungry GPUs running together to even compete (like 10s of thousands of dollars worth of GPUs)
Yeah, right, I’m sure there’s just no possible way to do it…
 
Same - I got my M4 11" iPP a few months after it came out. I even got AppleCare+ (and still have it for this) even when I usually don't for my iPads. Because of the newness of the tandem OLED, I wanted to be covered if some issues popped up down the road (eg. like the butterfly kbs of the 2016-2020 MBP lineup). So far, zero issues. So that's good news that Apple hopefully has a good handle on the technology.

True longevity of the screen will be 3-4 years from now - but if it can keep this up, will be good. And hopefully speaks well for it coming to the next-redesigned MBP.
 
As H.d mentioned, upgradable memory is not likely to be an option, but upgradable storage would get my attention as well.

I doubt we see either as Apple moves to even more integration on the SOC. Making storgae upgradeable would be counter to Apple's approach of thinner and more integration.

Now, Apple could make memeory software upgradable by making the installed storage say 2 TB, and selling say 256/512/1TB/2TB options where you could later upgrade via sofatware unlocking more memory. I doubt that would happen, however because:
  • It would be a complicated solution, probably requiring returning the device to Apple to do the upgrade
  • Hackers would try to find a way to do it for free and if they do, Apple loses a lot of revenue
  • People would complain how "They have 2TB but evil Apple only lets them use 512..."
It looks like Apple got the hint on batteries, with battery replacement in the most recent MacBooks being much easier.

I suspect that was a combination of making it cheaper and easier for Apple to do repairs and compliance with right to repair regulations. Cutting the time for techs to do a repair, and possibly doing more in store, frees up time to do more repairs and increase productivity. Making it easier for 3rd parties is just a byproduct and satisfies regulators where needed.
 
I was a MacBook Pro user with M1 but I don't think that I would be with M5 or M6. I need at least 32 GB of RAM which wasn't an option with M1 and the base M1 wasn't enough but the base M5 would be so I'm not longer a MacBook Pro customer for performance reasons. The speakers, screen and ports are a factor but I could also just go for the M5 Air if the portability outweighs the other stuff.
 
After having mbp 16 m1 pro for all these years,
I was planning to purchase mbp 16 m6 pro, in 2027.
Just to have something 5 generations ahead.

But today I read an article that m6 generation possibly will not come,
and the new ones will be m7 family?

See for yourself
 
After having mbp 16 m1 pro for all these years,
I was planning to purchase mbp 16 m6 pro, in 2027.
Just to have something 5 generations ahead.

But today I read an article that m6 generation possibly will not come,
and the new ones will be m7 family?

See for yourself

That seems to be the case, with an emphasis on running AI applications. I'm interested to see how well Ollama runs a local model on the M7 Max, it runs reasonably well on my base M3 Max.
 
  • Like
Reactions: cool11
1. Physics, with a SOC like the M series the ram is built into the CPU and GPU so data has to travel much less distance significantly reducing latency and signal degradation

2. Wider memory buses for higher bandwidth, versus narrower interfaces in socked/upgradable setups limited by pin counts and traces.

3. Optimized, proprietary interfaces without socket overhead like inductance or standardized protocols that constrain speeds in upgradeable RAM setups.

Its like a normal engine that is designed for public use and therefore has to support a wide variety of fuels vs a highly tuned engine optimized to a specific fuel (race gas, E85, etc) it can squeeze out way more power but then you are stuck with whatever it is tuned for.

The speed differences are why GPUs have never been able to use the socketed RAM. Its just way too slow for GPU purposes, but with a SOC like the M series the GPU can use the same RAM as the CPU allowing the studio for example to have access to 512 GB of video ram (- some RAM for running the system etc) allowing it to run huge AI models that would require extremely expensive and power hungry GPUs running together to even compete (like 10s of thousands of dollars worth of GPUs)
BIG THX for these explanations ! (really...)
 
The speed differences are why GPUs have never been able to use the socketed RAM. Its just way too slow for GPU purposes, but with a SOC like the M series the GPU can use the same RAM as the CPU allowing the studio for example to have access to 512 GB of video ram (- some RAM for running the system etc) allowing it to run huge AI models that would require extremely expensive and power hungry GPUs running together to even compete (like 10s of thousands of dollars worth of GPUs)

There is also an advantage, IMHO, at the lower memory levels as well - instead of running huge models, they'll run smaller focused ones very efficiently, allowing Apple to tailor AI to consumer machines as well. Small, task based models can be loaded and unloaded as needed, providing variety of capabilities with smaller amounts of RAM than say the Studio. Models optimized for specific tasks or sets of would be built into th OS, something Apple is well positioned for since tehy control both sides of the equation - hardware and software.
 
After having mbp 16 m1 pro for all these years,
I was planning to purchase mbp 16 m6 pro, in 2027.
Just to have something 5 generations ahead.

But today I read an article that m6 generation possibly will not come,
and the new ones will be m7 family?
I also have been waiting for redesigned MacBook Pro/Ultra assuming it would be released later this year but with recent crazy price hikes I've been pondering whether I should just get current Pro or even Air because there are still some left at old price to replace my current MBP 14" M1 Pro. I just don't like the idea of getting soon obsolete design because I plan to use it for next 4-5 years like my current MBP. Also with AI boom it seems likely that upcoming model prices likely will be around 4000EUR.
 
  • Like
Reactions: cool11
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.