Apple M2 SoC Analysis - Worse CPU efficiency compared to the M1

[Love-hate 🍏 relationship]

Geekerwan also compared the M1 and M2 to other Windows based systems and shows that the M2 MacBook Pro actually achieves higher battery life than the M1 MacBook Pro on their script based battery life test. Their battery test is based on 1 hour cycle that streams music constantly, 10 minutes each to Microsoft Office (Excel, PowerPoint and Word), 10 minutes web browsning, 10 minutes online video streaming in 4K and 10 minutes of instant messaging.

View attachment 2117506

that makes sense. whilst the new SOC is overall less efficient, the new E cores are , from what i gather, much more powerful than those of the m1

thus, for rather light tasks (or perhaps for all tasks actually, unless they are super heavy 4k editing ) , the processes are likely to be given to the E cores, while the m1's E arent powerful enough to sustain such load, and will give the processes to the P cores . and m1 P are obviously consuming more than m2 E .

 

[Love-hate 🍏 relationship]

I disagree.

not to bash on you but i'm afraid you don't know how chips work and neither do you understand what power efficiency , or other technical terms , mean .

i dont mean to offend you by saying that mind you, but your claims just dont hold

 

[Pressure]

that makes sense. whilst the new SOC is overall less efficient, the new E cores are , from what i gather, much more powerful than those of the m1

thus, for rather light tasks (or perhaps for all tasks actually, unless they are super heavy 4k editing ) , the processes are likely to be given to the E cores, while the m1's E arent powerful enough to sustain such load, and will give the processes to the P cores . and m1 P are obviously consuming more than m2 E .

That's the correct assessment.

The Efficiency cores received the largest upgrade going from the Icestorm core design to the Blizzard core design making them much more capable and receiving the biggest generational upgrade this cycle, while still being much more power-efficient than the High-Performance cores.

Overall giving the M2 better battery life with normal workloads. Benchmarks obviously skews this by running heavy loads.

 

[PauloSera]

not to bash on you but i'm afraid you don't know how chips work and neither do you understand what power efficiency , or other technical terms , mean .

i dont mean to offend you by saying that mind you, but your claims just dont hold

I know exactly what I'm talking about, thanks for chiming in. You can show yourself out.

 

[Xiao_Xi]

that makes sense. whilst the new SOC is overall less efficient, the new E cores are , from what i gather, much more powerful than those of the m1

Has anyone tried to run benchmarks only on the efficient cores?

 

[Love-hate 🍏 relationship]

Has anyone tried to run benchmarks only on the efficient cores?

max tech have i think

unless i'm mistaken , i remember that enabling low power mode on monterey will lower the clock (at least on the m2) and only use the little cores .

oh and to answer your question ,the m2 fares much better in that test , than the m1 does

 

[Love-hate 🍏 relationship]

overall , the m2 is kind of a disapointment, particularly in the mba new design, which has a worse heatsink (if at all ?) , due to its thinner design.

the GPU seems rather efficient that said,and we get 8c base and 10c unbinned which is cool (tho m2 is much more expensive so it makes sense we get more for the price)

but the CPU improvements are honestly marginal , whilst the power consumption is much higher as shown in the figures.

if not for the the new little cores , i actually prefer the m1 design.

still 5nm, armv8, wifi 6 w/ 2*2 MIMO ....sad

 

[Love-hate 🍏 relationship]

Geekerwan also compared the M1 and M2 to other Windows based systems and shows that the M2 MacBook Pro actually achieves higher battery life than the M1 MacBook Pro on their script based battery life test. Their battery test is based on 1 hour cycle that streams music constantly, 10 minutes each to Microsoft Office (Excel, PowerPoint and Word), 10 minutes web browsning, 10 minutes online video streaming in 4K and 10 minutes of instant messaging.

View attachment 2117506

that's too bad he didnt test the mba m1. i suppose that given the figures shown here, it should be around 10h or something ? since the mbp m1 and m2 have a near 2h difference , and the same battery size, while the mba m2's battery is bigger than the m1 (albeit not by a lot , less than 3Whr lol)

i have the mba m1 and the battery is already unbelievably good honestly

 

[neinjohn]

I disagree that this is the wrong direction. In fact, I welcome this direction as it gives the user more performance without any meaningful drawbacks. I mean, it would be definitely worrying if this trend continues but Apple can easily go up to 8-10W on the P-core and still maintain 2-3x efficiency headroom over the competitors. The actually worrying thing about M1 for me is the lack of any vertical performance capability on the P-cores and M2 is a proof that Apple is not committed to run their laptops at the same power level as a smartphone.

I thought the same from reading this forum and others with the second release of M1 products and as the clock was maintained. Meanwhile, Apple gave a clock boost to the M2 and released 3 generations on the A-series with the same P-cores. It's weird.

I can see the need for a new SoC, with new, much better E-cores, on their fan-less and mainly portable minded machines but not on everything else.

 

[leman]

Meanwhile, Apple gave a clock boost to the M2 and released 3 generations on the A-series with the same P-cores. It's weird.

Doesn't have to be weird. Could be that their main CPU team is busy with the next-gen P-core.

I can see the need for a new SoC, with new, much better E-cores, on their fan-less and mainly portable minded machines but not on everything else.

E-cores are also important on performance-oriented machines as they a) remove the contention from the P-cores and b) provide an extra boost of sustained performance. Of course, one doesn't have to take it as far as Intel with their 8+ E-cores, but when doing e..g matrix operations, and extra AMX cluster or two are always welcome.

 

[scottrichardson]

Article clearly states that the GPU is more powerful than the M1, while using LESS power. So that also needs to be factored into the equation since this is a chipSET with GPU on the die. Need to look at total package power consumption while using BOTH the cpu cores AND GPU cores at the same time.

 

[fromgophonetoiphone]

Seriously, the title is missing the point entirely. Do you want a 25% faster GPU? there you go, bigger chip.

-d

We've been getting faster chips generation after generation. If a bigger chip and more power hungry chip was needed, we'd have CPUs the size of your house already and power consumption of a Bitcoin mining farm. This is exactly why efficiency matters.

 

[jdb8167]

We've been getting faster chips generation after generation. If a bigger chip and more power hungry chip was needed, we'd have CPUs the size of your house already and power consumption of a Bitcoin mining farm. This is exactly why efficiency matters.

Except by all accounts, the M2 GPU is more efficient than the M1 GPU. It's only the M2 performance CPU cores that are less efficient. The M2 efficiency cores are also more efficient. Personally, having faster single-core performance is definitely worth having slightly less efficient performance cores. Because of the additional efficiency of the efficiency cores, I get the same battery life from my M2 MacBook Air as I did with my previous M1 MacBook Air. I doubt the extra 2.7 Watt-hour battery in the M2 MBA is helping much.

 

[Unregistered 4U]

Seriously where the hell does this article get less efficiency from?

Oh COME now, this is today’s internet! The article doesn’t need to get “less efficiency” from anywhere. They could post that M2 has infinitely less crealm and they’d still get lots of links, clicks, and ad views.

 

[Gudi]

Seriously, the title is missing the point entirely. Do you want a 25% faster GPU?

Absolutely not. The M1 is more than fast enough for me. Any hotter, louder, thicker, heavier for more performance is the opposite of what I want. Gimme a 3 nanometer TSMC chip with the same performance as the M1 and use the efficiency gains to make the MacBook Air thinner and lighter.

 

[Queen6]

Article pretty says at the end "Verdict - The M2 is just a small upgrade". From a user's perspective it's more a matter of battery life little else. Some sites show the M1 slightly ahead and some show the M2.

I've a 13" M1 and see little value in upgrading to M2. If I needed a more powerful MBP I'd opt for the 14" From the more detailed comparisons the M2 is around 11% faster and around 5% worse on battery and the numbers shift depending on the benchmarks hardly much to get concerned about and TBH expected.

Lot of comparison here, but nothing to lose your mind over...

Apple MacBook Pro 13 2022 M2 Laptop Review – Debut for the new Apple M2

Notebookcheck reviews the updates 2022 Apple MacBook Pro 13 with the brand-new Apple M2 SoC and 16 GB RAM.

www.notebookcheck.net

Q-6

 

[falainber]

Is this the first time Apple has replaced an SoC with a less efficient one?

If it is, the big reason for this could be the slowdown in TSMC process improvements. In the past, every new TSMC process brought big performance improvements but they somewhat choked on a 3nm process. It's been delayed and Apple had to use a slightly better version of 5nm process instead. But they couldn't have waited another year for releasing a new chip.

 

[Xiao_Xi]

Is this the first time Apple has replaced an SoC with a less efficient one?

Can anyone make sense of this chart with the third party benchmark test results we have?

At the M2 unveiling, Apple claimed that the M2 is 18% more efficient than the M1.

 

[Gudi]

Can anyone make sense of this chart with the third party benchmark test results we have?

View attachment 2119806

At the M2 unveiling, Apple claimed that the M2 is 18% more efficient than the M1.

Okay, in theory the M2 should have similar CPU performance to the M1, with the same number and kind of cores. It runs at a higher clock speed which increases both performance and power consumption, but it's also build with a slightly improved 5nm process which slightly decreases power consumption. And the CPU is only one part of the package, RAM, SSD and display also contribute to this 15W system power. So overall at roughly (but not really) the same total power consumption the M2 is a little bit faster.

 

[Luis Ortega]

Apple M2 SoC Analysis - Worse CPU efficiency compared to the M1

The Apple M2 is here and after our reviews of the MacBook Pro 13 and the MacBook Air M2, we want to have a closer look at the new M2 chip. What changed, which improvements can you expect, and how does the increased power consumption affect the efficiency, especially compared to the previous M1 chip?

www.notebookcheck.net

Review of the M2 and how it stacks up against the M1, M1 Pro, M1 Max and latest offerings by AMD and Intel. Very interesting read.

I think that this game is getting ridiculous. As a consumer, I would expect that any m2 would be better than any m1 variant. If not, what the hell are they improving with a new generation? It sounds more like marketing gimmicks to drive sales.

 

[leman]

I think that this game is getting ridiculous. As a consumer, I would expect that any m2 would be better than any m1 variant. If not, what the hell are they improving with a new generation? It sounds more like marketing gimmicks to drive sales.

It is better. That’s all that is relevant to a consumer.

 

[D*I*S_Frontman]

I just bought an M2 MBA (24GB/2TB). It doesn't matter to me that the M2 is slightly less efficient in some cherry-picked tests. I wanted an Air's dead-quiet, fanless design, but also more memory than M1 MBAs can be configured with, better peak performance when plugged in, and longer battery life when portable. In the real world, where one is already paying a premium for portability, the M2 MBA lasts longer than an M1 MBA on a charge while still having better performance and the option of more unified memory. M2 FTW, in my case.

It may not be a "quantum leap" in technology, but the M2 does still move the ball forward. It is certainly not a regression. If one measures "CPU efficiency" alone, under very specific criteria, the M1 can come out on top. But in practical terms, I would view "system efficiency" as a combination of factors, the chief of which being battery life, where the M2 wins, and peak performance, where the M2 also wins.

I think of the M2 MBA as a "semi-pro" unit: better than the M1 for CPU-intensive tasks, but not as good as any other offerings up the product line. For content creators, esp. musicians with project studios, this machine makes sense, esp. if one plans to record remotely. If not, the same money I bought this M2 MBA could have snagged me a base Studio.

Compared to my aging 8-core 2.8ghz MP tower (2008) which has served as the centerpiece of my project studio for years, this insanely thin, fanless laptop has 4x the raw processing power, 20x the memory bandwidth, and draws 13x less power at peak performance to do meaningful work than its predecessor while weighing 15x less. It's a compelling option for music producers, for whom silent operation is a critical factor.

 

[Xiao_Xi]

in theory the M2 should have similar CPU performance to the M1

What performance does CPU performance refer to: single-core/multi-core/P-core/E-core?

 

[Gudi]

What performance does CPU performance refer to: single-core/multi-core/P-core/E-core?

Integer or Floating Point operations? You can compare a dozen different CPU benchmarks and the best marketable number says 18% more performance. On average over many benchmarks it's about 11% better, but depending on the individual use case it can be anything in-between. The point is, both chips have very comparable performance and power consumption. You might as well think of them as (almost) the same chip.

 

[lcubed]

I think that this game is getting ridiculous. As a consumer, I would expect that any m2 would be better than any m1 variant.

this is a ridiculous expectation. i've benchmarked several pieces of AI driven photo software where the M1 Ultra is simply faster than the current M2 since the software takes advantage of the additional ANE and GPU's available.