Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

mactinkerlover

macrumors regular
Original poster
Sep 20, 2020
174
113
So, title. Why did the original M1 Pro/Max only have 2 efficiency cores? Was it because apple thought they didn't need more than 2? Or was it a die space issue? Or was it simply a design decision?

Also, why did they add 2 more with M2 Pro/Max? Was it just to boost the cpu performance a bit more? Was it because they knew m2 was going to be less efficient so they added 2 more efficiency cores?

Thoughts? Why couldn't they just have 4 efficiency cores on M1 Pro/Max to start with?
 
So, title. Why did the original M1 Pro/Max only have 2 efficiency cores? Was it because apple thought they didn't need more than 2? Or was it a die space issue? Or was it simply a design decision?

Also, why did they add 2 more with M2 Pro/Max? Was it just to boost the cpu performance a bit more? Was it because they knew m2 was going to be less efficient so they added 2 more efficiency cores?

Thoughts? Why couldn't they just have 4 efficiency cores on M1 Pro/Max to start with?
They probably had a transistor/area budget and fit in as much as they could to get the best performance. The efficiency cores are small but they connect to things like cache which takes up some space. (I don't know enough to actually know the tradeoffs.) I assume that adding 2 E cores didn't provide enough boost to be worth it. Don't forget the M2 E cores are more powerful and are probably worth it since they take up more or less the same amount of space as the M1 E cores.
 
So, title. Why did the original M1 Pro/Max only have 2 efficiency cores? Was it because apple thought they didn't need more than 2? Or was it a die space issue? Or was it simply a design decision?

Also, why did they add 2 more with M2 Pro/Max? Was it just to boost the cpu performance a bit more? Was it because they knew m2 was going to be less efficient so they added 2 more efficiency cores?

Thoughts? Why couldn't they just have 4 efficiency cores on M1 Pro/Max to start with?
Design targets?
 
Almost certainly transistor budget related. Engineering is making tradeoffs. Cutting efficiency cores was likely the best way to meet speed goals without going over the transistor budget. It made sense to me at least. Efficiency cores are less important to users with heavy workloads, which depend on the power cores. Using efficiency cores also matter less with bigger batteries of the MBPs and no batteries at all for the Studio.

Ultimately, I think 8/4 ratio is well balanced ratio and I wouldn't be surprised if the M3 kept the same number of cores - just make them even faster and even more efficient.
 
  • Like
Reactions: Tagbert
So, title. Why did the original M1 Pro/Max only have 2 efficiency cores?
My take is they made decisions to balance performance, and battery life. They opted for 2 E cores, because they wanted to make sure that their inaugural M series processor out performed Intel. So much was at stake for apple, and they opted for performance minded processor with great battery life.
 
My take is they made decisions to balance performance, and battery life. They opted for 2 E cores, because they wanted to make sure that their inaugural M series processor out performed Intel. So much was at stake for apple, and they opted for performance minded processor with great battery life.
Wouldn't 2 more E cores improves multi-core performance? And since E cores consume so little energy, it probably wouldn't hit battery life much.

My take is on dies size budget, since when the M1 Pro/Max came out, the 5nm process might still be relatively expensive. Come M2, the cost probably came down enough to add in 2 more E cores and more circuitry.
 
They were going to add 1 P core and 2 E cores but the jocks of Silicon Valley started saying
”Tim PEE Cores! Tiiiim PEE Coores!” so they cancelled that.
 
  • Haha
Reactions: Scarrus
I think there are a few reasons why the M1 SoCs were set up that way. As the first of the M-Series, part of the process from the manufacturing side is trying to determine a balance between production capability and customer needs/demands. Another consideration to be made is what the yield rates for each design is. While Intel and AMD often simply bin a CPU up or down based on factors such as stable speed, number of working cores, etc., I'm not sure how much binning can be done with Apple Silicon (Specifically M1 Pro/Max/Ultra) given what we know about the SoC designs. Given these reasons, I still think the main reason is much simpler in nature. By keeping core counts down with the first generation Ml-series, Apple can demonstrate performance advantages over competing CPUs with even the base M1 then build upon that with the Pro, Max, and Ultra SKUs, all while maximizing the reliability of a first-generation part.
 
I am inclined to believe that their initial modelling showed that 2 E-cores would be plenty for desktop applications. The goal of E-core was really just to run mon-critical background tasks (save resources and power, improve system responsiveness). After the hardware was out maybe they determined that E-cores were more useful (or more used) than they initially thought. Plus, improving the performance of E-cores also helped in justifying to use more of them, as they can offer a measurable improvement in multicore thought put.
 
  • Like
Reactions: andrewmarich
A15 efficiency cores performance improved by 23% according to Anandtech. A14/M1 E-cores just weren’t that good. Adding more E cores wouldn’t have added enough performance compared to the higher power consumption.

If you look at Apple deign, they alternate improvements each year on the P cores followed by E cores.
 
The two E cores on my M1 MBP max out periodically throughout the day. When this happens, parts of the system feel slow. Two E cores are really not enough in my experience, so the M2 brings a needed improvement.
 
  • Like
Reactions: rmadsen3
Anymore evidence of this?

Apple has been using this kind of cyclical improvement strategy since at least A8 if I remember correctly. On average they take about 2 years per subsystems redesign - can be P-core every year, caches the other year etc. A16 is the first design in a while where this rhythm is broken (it did improve E-cores however If I remember correctly).

You can look up the Anandtech reviews of A-series chips, Andrei F. did an excellent job tracking these things.
 
  • Like
Reactions: T'hain Esh Kelch
Apple has been using this kind of cyclical improvement strategy since at least A8 if I remember correctly. On average they take about 2 years per subsystems redesign - can be P-core every year, caches the other year etc. A16 is the first design in a while where this rhythm is broken (it did improve E-cores however If I remember correctly).

You can look up the Anandtech reviews of A-series chips, Andrei F. did an excellent job tracking these things.
Didn't the CPU of the A series improve significantly every single year until the last 2 years or so?

Anyways, I'm hoping for a big P core improvement in M3. That's when I'll upgrade.
 
A15 efficiency cores performance improved by 23% according to Anandtech. A14/M1 E-cores just weren’t that good. Adding more E cores wouldn’t have added enough performance compared to the higher power consumption.

If you look at Apple deign, they alternate improvements each year on the P cores followed by E cores.
What do you mean weren’t that good? Apple has by far the most powerful e cores in the industry, and at shockingly low wattage to boot.
 
The two E cores on my M1 MBP max out periodically throughout the day. When this happens, parts of the system feel slow. Two E cores are really not enough in my experience, so the M2 brings a needed improvement.
Would you please explain how one can monitor this? Is it an iStat Menus thing? Thanks!
 
What do you mean weren’t that good? Apple has by far the most powerful e cores in the industry, and at shockingly low wattage to boot.

Apple is most likely benchmarking themselves with die area and thermal targets, not against others.
 
Last edited:
62F450FF-8DE8-4733-91CE-6091DDD7EF0A.png
Apple is most likely benchmarking themselves with due area and thermal targets, not against others.
But that doesn’t matter what Apple is doing in terms of E core performance, they don’t even talk about it publicly. Our resident chip expert @cmaier explained how far ahead Apple’s E cores are from both a performance AND extreme low power consumption prospective a while back. I’ll see if I can dig that up after work.

Edit: Apologies, I mixed up users, it was actually @lemans (another hugely knowledgeable member in chip design).
 
Last edited:
Edit: Apologies, I mixed up users, it was actually @lemans (another hugely knowledgeable member in chip design).

You mean an “opinionated amateur who never has anything to do with chip design” 😅 I do think I have some knowledge regarding the software side. Even though still amateur.
 
  • Like
Reactions: rmadsen3
You mean an “opinionated amateur who never has anything to do with chip design” 😅 I do think I have some knowledge regarding the software side. Even though still amateur.
Yes but you take the time to actually understand and parse the reports you look into, rather than the armchair warriors who just repeat verbatim claims made by technological illiterates in tech “journalism” circles. You’ve contributed greatly to the more technical discussions on this site. So much so that I often mistake you and CMaier (whom I now realize is no longer here).
 
But that doesn’t matter what Apple is doing in terms of E core performance, they don’t even talk about it publicly. Our resident chip expert @cmaier explained how far ahead Apple’s E cores are from both a performance AND extreme low power consumption prospective a while back. I’ll see if I can dig that up after work.

It’s also important to keep in mind that Intel’s E-cores and Apple’s E-cores have very different design targets. Intel is about performance per area - they want to maximize multi-core throughout without making the chip bigger. Apples E-cores are about minimizing energy per task. Two very different goals, two very different approaches.
 
It’s also important to keep in mind that Intel’s E-cores and Apple’s E-cores have very different design targets. Intel is about performance per area - they want to maximize multi-core throughout without making the chip bigger. Apples E-cores are about minimizing energy per task. Two very different goals, two very different approaches.
Very true, Apples done a ton of work in the logic of how and when the e-cores get used. It’s what they deliver at the insanely lower power draw that I still can’t wrap my mind around.
 
  • Like
Reactions: Scarrus
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.