M5 Pro and Max unveiled

Extremely unlikely
They've literally said it's 3rd gen 3nm. I don't even understand how you can say it's "extremely unlikely." It's literally impossible unless Apple directly lied in their marketing. This thread is utterly confusing to me

This brings together two third-generation 3-nanometer dies with high bandwidth and low latency using advanced packaging.
 
Despite my earlier statements, there is one way where the new SOC design might make a difference and that's if Apple shoves the streaming/web browsing thread onto the M-"Performance" cores instead of the P-"Super" cores on battery - Intel does this to save battery life. I don't think Apple will do that given their P-cores are extremely efficient compared to Intel's P-core and Apple tends to run full throttle on battery vs plugged-in while Intel very much doesn't do so, but Apple might do this. That could potentially result in the increased battery life and it be from the change in core design.
I'm sticking to what I said because it's verifiably proven. Explain why otherwise if it's the N1 chip doing the heavy lifting for battery improvements that the Air with an M5 chip and N1 chip doesn't get any official increase?
 
There are 3 types of cores. This isn't confusing.
And yet we keep seeing people (including developers who have been writing software for Apple’s platforms for 30 years) saying they are confused, news reports that show that reporters are confused, etc.
 
And yet we keep seeing people (including developers who have been writing software for Apple’s platforms for 30 years) saying they are confused, news reports that show that reporters are confused, etc.
And yet I with none of that am confused lol. I don't understand what you're confused by. This is an invitation to explain what you feel
 
And yet I with none of that am confused lol. I don't understand what you're confused by. This is an invitation to explain what you feel
I’ve explained it repeatedly. Calling the same thing by two different names depending on what the date on the calendar is is confusing. Changing the names of things to be different than what everyone in the industry calls them is confusing. Using a name that used to mean one thing to now mean something else is confusing.

It’s like if you had batman and robin, and decided to change batman’s name to superman and robin’s name to batman.

So if I say “batman” who am I talking about? New batman? Old batman?
 
I’ve explained it repeatedly. Calling the same thing by two different names depending on what the date on the calendar is is confusing. Changing the names of things to be different than what everyone in the industry calls them is confusing. Using a name that used to mean one thing to now mean something else is confusing.

It’s like if you had batman and robin, and decided to change batman’s name to superman and robin’s name to batman.

So if I say “batman” who am I talking about? New batman? Old batman?
Unless this is a constant thing, I don't see why you can't adapt to a single name change. If they constantly in changed it every time they updated a chip, then I can see the reasoning for it be confusing.

They have three tiers of core across the entire company lineup: super, performance, and efficiency. Performance is the new one. It is in the middle between super and efficiency. It offers most of the capability of the super core without as much power. It is faster than the E core.

Depending on what product, the cores will change. E cores NEED to be extremely low energy because they're used in Apple Watch. If they keep forcing it to use more wattage over time, this is an issue. This move allows them to plot forward a new path: high performance products like Mac desktops and some notebooks will use super cores and performance cores. High efficiency products like iPhone and Apple Watch will use super cores and efficiency cores.

Apple is not changing the strategy for the chip using 2 tiers of cores: high performance and high efficiency. But what the product needs will change what core designs they have.
 
Alongside the other reasons I've stated in comments in previous pages, I think this is telling about the highest end chip they will make. Before it was 2 Max chips, but now it will be 1 chip itself using the brand new tech. I think the introduction of this third core will skyrocket overall performance of the chip CPU wise
 
They have three tiers of core across the entire company lineup: super, performance, and efficiency.
No it’s not! The *name* “performance” already existed and already referred to a different kind of core. Hence the confusion.

If you come up with a new thing and keep selling the old thing, name the new thing something different.

Call them Performance, Balanced, and Efficiency. Whatever.
 
They've literally said it's 3rd gen 3nm. I don't even understand how you can say it's "extremely unlikely." It's literally impossible unless Apple directly lied in their marketing. This thread is utterly confusing to me
Simply because I forgot they said that explicitly ...
I'm sticking to what I said because it's verifiably proven. Explain why otherwise if it's the N1 chip doing the heavy lifting for battery improvements that the Air with an M5 chip and N1 chip doesn't get any official increase?
That's a fair point - however, I think that was @leman's hypothesis - edit: actually rereading he also included the un-core too. My hypothesis is that it's the un-core (which would include the N1 but is more than that). That said, I've added an addendum that if Apple has updated how it does its load balancing for ST and lightly multi-threaded workloads then that could indeed produce this effect if these types of threads are simply no longer running on the S-cores by default, especially when on battery.
 
We shall see, when we have benchmarks to look at. FWIW, the best x86 SC score in GB6 is about 86% of the best Mac SC score when adjusted for clock speed (much worse for the Ryzen 9 in raw numbers). If the M5 Max shows a leap here, it will be obvious that "S" is not just a different name for "P".
 
We shall see, when we have benchmarks to look at. FWIW, the best x86 SC score in GB6 is about 86% of the best Mac SC score when adjusted for clock speed (much worse for the Ryzen 9 in raw numbers). If the M5 Max shows a leap here, it will be obvious that "S" is not just a different name for "P".
The “S" core is definitely just a different name for the previously known “P” core. Apple said so.
 
No it’s not! The *name* “performance” already existed and already referred to a different kind of core. Hence the confusion.

If you come up with a new thing and keep selling the old thing, name the new thing something different.

Call them Performance, Balanced, and Efficiency. Whatever.
My dude the names have changed. It's just how it is. Until we actually see what the P core does, we won't know how performant it is. But I'm saying now that it's going to be very performant, more than the E core, and that was already good to begin with. Hence, I support calling it performance and labeling the fastest core "super."

I guarantee this isn't just a marketing change and for the sake of it. There is an engineering reason behind it and it will become very clear at release, and it will affect how chips are made from here on out. The highest end M chip will have way more performance overall. Apple Watch will be able to increase its battery life. Apple will no longer need to force the E core to draw more wattage because they need to offload most of the tasks to the E cores to save on battery.
We shall see, when we have benchmarks to look at. The best x86 SC score in GB6 is about 86% of the best Mac SC score when adjusted for clock speed (worse for the Ryzen 9 in raw numbers). If the M5 Max shows a leap here, it will be obvious that "S" is not just a different name for "P".
It literally is a just a different name though. Apple said that. The actual new core is the P core, performance core, the thing in M5 Pro etc, as in M5 Pro/Max have 6 super cores and 12 performance cores. Super is the high performance core and performance is what they chose for high efficiency. In the regular M5 they chose super core and efficiency cores.

I'm calling it now: it will be 4500 when it's all said and done. That's an estimation of mine.
 
Simply because I forgot they said that explicitly ...

That's a fair point - however, I think that was @leman's hypothesis - edit: actually rereading he also included the un-core too. My hypothesis is that it's the un-core (which would include the N1 but is more than that). That said, I've added an addendum that if Apple has updated how it does its load balancing for ST and lightly multi-threaded workloads then that could indeed produce this effect if these types of threads are simply no longer running on the S-cores by default, especially when on battery.
Fair, sorry. But yes, it's 3rd gen 3nm.

As to whatever the "un-core" is, I'm going to say as I've said the whole time. It's the new performance core that allows Apple to boost overall performance, boost core counts, boost battery life, and maintain thermal efficiency.

Pretend they only had 2 cores still ("performance and efficiency"). If they reduced M5 Max to 6 performance cores and added 12 efficiency, that would increase battery life but it wouldn't increase performance that much . It would be similar to what happened with M3 Pro. M4 Max has TWELVE of the highest performing cores, which draw the most power. Halving that will have a major effect. M5 base has 6 super cores itself with 4 efficiency, yet it's far away from M4 Max let alone M5 Max.

It's the P cores that has allowed Apple to achieve the objective results they did: 20% increase in code compilation ,2 hours extra battery life, maintain thermal. The new fusion tech also will affect it, correct
 
I guarantee this isn't just a marketing change and for the sake of it. There is an engineering reason behind it and it will become very clear at release, and it will affect how chips are made from here on out. The highest end M chip will have way more performance overall. Apple Watch will be able to increase its battery life. Apple will no longer need to force the E core to draw more wattage because they need to offload most of the tasks to the E cores to save on battery.

It literally is a just a different name though. Apple said that. The actual new core is the P core, performance core, the thing in M5 Pro etc, as in M5 Pro/Max have 6 super cores and 12 performance cores. Super is the high performance core and performance is what they chose for high efficiency. In the regular M5 they chose super core and efficiency cores.
I can't speak on behalf of others, but to me, it's not that I don't get it. It's not that I don't see the engineering reasoning. It's that what we call a thing, and what that thing is, are two orthogonal concepts.

The engineering decided to create a core between E and P and how to lay out the chip. Marketting decided to name the new core a P core, a name that was already in use, and reshuffle the labels, causing a name change + a new name, instead of only a new name.

So yes, it's a matter of engineering that we have a new core. What it is named, is a matter of marketting.

A rose by any other name would still smell as sweet, as Shakespeare wrote. The performance characteristics would not be any different had the exact same cores been labelled Performance, Scaling and efficiency cores or whatever else instead
 
I can't speak on behalf of others, but to me, it's not that I don't get it. It's not that I don't see the engineering reasoning. It's that what we call a thing, and what that thing is, are two orthogonal concepts.

The engineering decided to create a core between E and P and how to lay out the chip. Marketting decided to name the new core a P core, a name that was already in use, and reshuffle the labels, causing a name change + a new name, instead of only a new name.
as I said, it's probably because the new P core will be far faster than "balanced" would suggest.
A rose by any other name would still smell as sweet, as Shakespeare wrote. The performance characteristics would not be any different had the exact same cores been labelled Performance, Scaling and efficiency cores or whatever else instead
I don't care about the names. I care about the engineering and the implications of it. I've said repeatedly what I think the effect will be. I might be wrong, but I think talking about what it might unlock is more substantial.
 
Fair, sorry. But yes, it's 3rd gen 3nm.

As to whatever the "un-core" is, I'm going to say as I've said the whole time. It's the new performance core that allows Apple to boost overall performance, boost core counts, boost battery life, and maintain thermal efficiency.

Pretend they only had 2 cores still ("performance and efficiency"). If they reduced M5 Max to 6 performance cores and added 12 efficiency, that would increase battery life but it wouldn't increase performance that much . It would be similar to what happened with M3 Pro. M4 Max has TWELVE of the highest performing cores, which draw the most power. Halving that will have a major effect. M5 base has 6 super cores itself with 4 efficiency, yet it's far away from M4 Max let alone M5 Max.

It's the P cores that has allowed Apple to achieve the objective results they did: 20% increase in code compilation ,2 hours extra battery life, maintain thermal. The new fusion tech also will affect it, correct
That depends. Depending on how you structure your die, it's entirely possible to power gate an entire set of cores, such that they only use any power at all while they are actively in use, and their idle power is 0. That's why we talk about the un-core. All the stuff surrounding the cores themselves that you can't just power gate, or even necessarily clock gate.

Power gating also isn't necessarily per-core. Sometimes you can only power gate a whole cluster, so what you're saying could still play a part, but for idle or near-idle power consumption, the power draw of clusters of high performance cores may be reduced to 0 or ~0 while some blocks surrounding it cannot be disabled at all.
 
My dude the names have changed. It's just how it is.
I know. And as much as it upsets you, I am entitled to be annoyed by the name change, and countless people are confused by it.


Until we actually see what the P core does, we won't know how performant it is. But I'm saying now that it's going to be very performant, more than the E core, and that was already good to begin with.
Right. So? That means it needs to be named “P” instead of, say, “M” or “B”? It’s not as performant as the “S” core, for sure.


I guarantee this isn't just a marketing change and for the sake of it. There is an engineering reason behind it and it will become very clear at release, and it will affect how chips are made from here on out.
What? It’s clearly a marketing change. Internally, on the engineering team, they don’t use these names. These are literally names created by the marketing department.

Sure, there is an engineering change behind the architectural and design changes. But those architectural and design changes do not necessitate renaming what used to be “P” *just a few months ago* to “S” and then naming the new thing “P”
 
Back
Top