Intel's Alder Lake mobile flagship CPU is faster than the Apple M1 Pro and M1 Max

nanoguy

Posts: 1,203   +20
Staff member
In context: Ever since Apple started its transition away from Intel processors, every incarnation of Apple Silicon has been pitted against both mobile and desktop processors from Intel and AMD. Now that the M1 Pro and M1 Max have been out, reviewers are looking into how they compare against Intel’s 12th Gen mobile offerings.

This week, PCWorld took a look at how the 14-core Intel Core i9-12900HK compares against the previous Rocket Lake flagship CPU and the AMD Ryzen 9 5900HX. The analysis revealed that at least in the realm of Windows laptops, Intel’s mobile Alder Lake flagship offers the best performance in a variety of tasks and also affords plenty of battery life thanks to its eight efficiency cores.

Macworld took the opportunity to compare Geekbench and Cinebench results between the Core i9-12900HK and Apple’s M1 Pro and M1 Max chipsets, equipped with 10-core CPUs. It found the Intel part was slightly faster than both pieces of Apple silicon, which is a different story than that told by early engineering samples of the Alder Lake CPU.

Apple silicon usually shines bright in Geekbench 5, but the Core i9-12900HK managed to score 1,838 points in the single-core test and 13,235 points in the multi-core. Those scores are only a few percent better than the M1 Pro and M1 Max, but it’s a lead nonetheless. When it comes to the Geekbench 5 OpenCL test, the Intel Iris Xe stands no chance against the 16-core GPU in the Apple M1 Pro, let alone the 32-core GPU in the M1 Max, which is almost three times faster.

The Core i9-12900HK is also faster in Cinebench R23, where it manages 1895 points in the single-core and 15,981 points in the multi-core test. The 10-core CPU in the Apple M1 Pro isn’t slow by any means, but it doesn’t come too close at 1,531 points and 12,381 points, respectively.

It’s worth noting that while the Alder Lake advantage is evident in these benchmarks, it comes at the cost of higher power consumption. We already know that energy efficiency is what elevates the M1 Pro and M1 Max into a class of their own, but the Core i9-12900HK can’t compete when it draws over two times the power, and at times it will even spike to 140 watts. The MSI GE76 Raider that PCWorld and Macworld used in their testing is also double the price at $3,999, so take that as you may.

Apple expects to complete the transition to first-party silicon this year by introducing a new Mac Pro that could be powered by a monster chip with up to 40 CPU cores and 128 GPU cores. Intel CEO Pat Gelsinger wants to earn back Apple’s business by outcompeting it, so it looks he has his work cut out for him. Taiwan Semiconductor Manufacturing Company’s founder says time is against Gelsinger and his ambitions, but we’ll have to wait and see.

Permalink to story.

 

VitalyT

Posts: 6,225   +6,757
That's like comparing a muscle car to an EV, while ignoring the efficiency.

And a 140W monstrosity is as mobile as Randy Marsh's balz...

randy-gets-cancer.png
 
Last edited:

Theinsanegamer

Posts: 3,364   +5,590
That's like comparing a muscle car to an EV, while ignoring the efficiency.
How apt, given the "efficient" ARM chips are limited to only $3000+ macbook laptops that can only run macOS and have soldered storage and RAM. They are severely limited in scope, just like EVs!
If the Core i9-12900HK and drawing 140W, there’s nothing “mobile” about this chip. The “laptop” it’s inside will have to be plugged into an outlet.
And how long do you think the macbook can ram it's m1 max on battery? 140W is "spike" power, not sustained power.
 

fps4ever

Posts: 905   +1,336
Intel's flagship mobile cpu power usage comparison. I'm no Apple fan but yikes. Especially where power, efficiency and cooling are key factors in mobile.
 

brucek

Posts: 1,126   +1,668
Welcome back Intel.

Now if we could just get efficiency + power + gaming compatibility instead of having to choose...

 

Sausagemeat

Posts: 1,597   +1,421
This actually amusing. Intel failed big time here, 140 watts is loads, that will need a fan, it will drain your battery. Your device will get hot.

X86 is out of date. Intel (And AMD for that matter) dont have a hope in hell of competing with RISC based architecture for personal computing. RISC is like a textbook case of a disruptive technology.
 

psycros

Posts: 4,147   +5,776
This actually amusing. Intel failed big time here, 140 watts is loads, that will need a fan, it will drain your battery. Your device will get hot.

X86 is out of date. Intel (And AMD for that matter) dont have a hope in hell of competing with RISC based architecture for personal computing. RISC is like a textbook case of a disruptive technology.

Gee, where have I heard that before...and somehow I think Apple was involved.
 

moon982

Posts: 100   +26
The benchmark is comparing Apples to oranges.

One is ultra thin netbook and other is big think gaming computer.

One is using mobile CPU and other is using desktop CPU.
 

DZillaXx

Posts: 527   +676
This actually amusing. Intel failed big time here, 140 watts is loads, that will need a fan, it will drain your battery. Your device will get hot.

X86 is out of date. Intel (And AMD for that matter) dont have a hope in hell of competing with RISC based architecture for personal computing. RISC is like a textbook case of a disruptive technology.

Current X86 chips are RISC and have been for ages.

Intel's Pentium Pro arch was their first x86 RISC chip, and AMD's K5 was their first x86 RISC design.

The real issue with X86 is the large amount of bloat from years of usage. ARM is not free from this as well.


You'd be surprised how close AMD's Ryzen is compared to Apple's M1 chips when it comes to performance per watt, given Apple's large Fab advantage and very integrated design. There simply isn't a X86 chip designed on a very integrated package like the M1, nor is any X86 chip using top of the line Fabing techniques.
 

moon982

Posts: 100   +26

Intel's flagship mobile cpu power usage comparison. I'm no Apple fan but yikes. Especially where power, efficiency and cooling are key factors in mobile.

x86 was designed for desktop computers not smartphones, tablets and netbooks.

Windows laptops always needed a fan and heat sink and ran really hot.

And all day battery life on some windows high in laptop now is only a new thing and never was thing before. Go back 10 years ago and you where lucky to get windows laptop battery to last 4 hours and enjoy a space heater and noisy fans.

The Microsoft surface tablets can never compete with the iPad using Intel and never will. Unless you want to enjoy the space heater and noisy fans.

The Intel Celeron , Pentium and i3 on less than $600 laptops and desktop computers are horrible compared to Apple A12 used on the $300 iPad.



 

Sausagemeat

Posts: 1,597   +1,421
Current X86 chips are RISC and have been for ages.

Intel's Pentium Pro arch was their first x86 RISC chip, and AMD's K5 was their first x86 RISC design.

The real issue with X86 is the large amount of bloat from years of usage. ARM is not free from this as well.


You'd be surprised how close AMD's Ryzen is compared to Apple's M1 chips when it comes to performance per watt, given Apple's large Fab advantage and very integrated design. There simply isn't a X86 chip designed on a very integrated package like the M1, nor is any X86 chip using top of the line Fabing techniques.
Nope, X86 is not RISC. It’s “CISC”.

Also Ryzen 5000 is hopeless compared to the M1 in terms of performance per watt. But Ryzen 5000 and Intel 12th gen are both faster architectures overall (some apps run faster on them). But in terms of efficiency they both get humiliated by the M1. Look it up lol.
 

umbala

Posts: 603   +1,007
It has become obvious that Intel can no longer compete not only against AMD but even Apple now. Yes, they can win benchmarks by tiny percentage points so the Intel fanboys can brag about it, but they can only do it at the cost of pushing power requirements to ridiculous levels. Their flagship desktop chip that was just released can barely beat AMD's best and only while using something crazy like 240+ watts!
 

dangh

Posts: 576   +912
Nope, X86 is not RISC. It’s “CISC”.

Also Ryzen 5000 is hopeless compared to the M1 in terms of performance per watt. But Ryzen 5000 and Intel 12th gen are both faster architectures overall (some apps run faster on them). But in terms of efficiency they both get humiliated by the M1. Look it up lol.
X86 is cisc. And cpus from 586 up are more risc than cisc. No need to look up for it, maybe you should actually uderstand that x86 instructions are extremely small set in modern cpus and are mostly emulated.
Please read this do you not stuck in terminology from last millennium:
https://www.extremetech.com/computi...the-wrong-lens-to-compare-modern-x86-arm-cpus
 

HardReset

Posts: 1,665   +1,321
Also Ryzen 5000 is hopeless compared to the M1 in terms of performance per watt. But Ryzen 5000 and Intel 12th gen are both faster architectures overall (some apps run faster on them). But in terms of efficiency they both get humiliated by the M1. Look it up lol.
Efficiency comparison is pointless because M1 is not CPU but SOC.
It seems that only AMD's next gen can truly compete in terms of efficiency with the M1 chips.
Probably not but still: Apple M1 is not CPU, it's SOC. AMD's next gen CPUs are CPUs, not SOCs. Apples to oranges comparison.

Tbh, there is nothing special about M1 SOC. Basically Apple just want to show everyone they can afford to waste tons of wafer area that makes chip shortage much worse.
 

DZillaXx

Posts: 527   +676
Nope, X86 is not RISC. It’s “CISC”.

Also Ryzen 5000 is hopeless compared to the M1 in terms of performance per watt. But Ryzen 5000 and Intel 12th gen are both faster architectures overall (some apps run faster on them). But in terms of efficiency they both get humiliated by the M1. Look it up lol.
Yes modern X86 are RISC. How about do some research before looking like a fool. Like I said for Intel Starting with the Pentium Pro (later known as Pentium 2) and AMD K5, the Chips are RISC. There is a decode/encode engine that converts X86 CISC commands to more basic RISC opt code. And Modern X86 compilers take full advantage of the fact that modern X86 chips are RISC, and the decode/encode engine works damn well. The Encode/Decode engine does take up die space and some power, but not very much. Biggest downside is the wasted die space TBH.

I don't know of any modern CISC chip that is actually CISC in function. The Converting of cisc to risc opt code was perfected two decades ago.

For Chip Architecture we come to the conclusion that CISC vs RISC vs VLIW, RISC is the way to go. VLIW can do quite well, but it is a pain to get right. Intel knows this, and AMD did a great job with it. But AMD spent a lot of time on the driver side keeping up because of it.


There is a lot of backwards compatibility to X86 and X86 compilers. Even if that backwards compatibility isn't needed.
 

Loadedaxe

Posts: 72   +106
Current X86 chips are RISC and have been for ages.

Intel's Pentium Pro arch was their first x86 RISC chip, and AMD's K5 was their first x86 RISC design.

The real issue with X86 is the large amount of bloat from years of usage. ARM is not free from this as well.


You'd be surprised how close AMD's Ryzen is compared to Apple's M1 chips when it comes to performance per watt, given Apple's large Fab advantage and very integrated design. There simply isn't a X86 chip designed on a very integrated package like the M1, nor is any X86 chip using top of the line Fabing techniques.
Ryzen, although very close is not the badass at power efficiency as you think. Its getting there but AMD has alot to do to get there especially if your counting performance per watt.
 
Man of these comments are like so many "arguments" today - more political or tribal then having to do with the real issues. RISC/CISC chips aren't apple or Intel. What apple has done is incredible, but of course with out Intel/Windows Apple would have died in 1997.
Desktop vs portable chips are different, and Intel has never had this competition before - it will be interesting to see where they go, how they respond. Better results will benefit all of us.
 

Sausagemeat

Posts: 1,597   +1,421
X86 is cisc. And cpus from 586 up are more risc than cisc. No need to look up for it, maybe you should actually uderstand that x86 instructions are extremely small set in modern cpus and are mostly emulated.
Please read this do you not stuck in terminology from last millennium:
https://www.extremetech.com/computi...the-wrong-lens-to-compare-modern-x86-arm-cpus
If you say that X86 is RISC I’m going to call you out for that bullshit. They are emphatically not RISC CPUs and cannot run the same software without emulation.

If you would like further schooling then DM me. I’d be more than happy to school you.
 

UncleMikeRetro

Posts: 17   +31
If the Core i9-12900HK is drawing 140W, there’s nothing “mobile” about this chip.

Reminds me of when the 40lb IBM Portable PC was considered “portable” because the screen and keyboard were built in and it even had a handle to “carry” it around!
Remember they even tagged them as "Luggables" because they were more like a piece of luggage you had to haul all over the place. Some had wheels!
 

rmcrys

Posts: 155   +133
Man of these comments are like so many "arguments" today - more political or tribal then having to do with the real issues. RISC/CISC chips aren't apple or Intel. What apple has done is incredible, but of course with out Intel/Windows Apple would have died in 1997.
Desktop vs portable chips are different, and Intel has never had this competition before - it will be interesting to see where they go, how they respond. Better results will benefit all of us.

You're right.

What is important to us as consumers it is: ​"what car runs faster consuming less". What people are discussing is if it runs on unloaded 95, 98, 100...

The important thing is not how internally it works but what they achieve. If Apple M1 chips achieve a great performance and consume little, fantastic. Sadly it doesn't run windows neither most games....