Intel Core i9-12900HK: 14 core and 20 thread processor pulls no punches with the Apple M1 Max in leaked Geekbench listing

[Redaktion]

The Core i9-12900HK has landed on Geekbench, albeit in a closed benchmark. The early signs are good for Intel's premier laptop processor, with it outscoring all contemporary processors in its wake, including the Apple M1 Max.

https://www.notebookcheck.net/Intel-Core-i9-12900HK-14-core-and-20-thread-processor-pulls-no-punches-with-the-Apple-M1-Max-in-leaked-Geekbench-listing.574868.0.html

[JasonG]

Thank is indeed impressive. Competition is great for consumers.

[Joe]

So it only takes Intel 4 more cores, and 10 more threads, to barely beat the M1 Max and the M1 Max hasn't even shown benchmarks for high powered mode. 

Also, the Intel might be claiming 95 or 125 watts, but we all know it will likely be closer to 250, where the M1 Max is at 30 peak. 

Sorry, but Intel is outright getting embarrassed by the Apple engineers here.

[MikeA]

Challenge: see if you guys can spot the apple fanboy lmfao

[duh]

3% Lead lol

[Barebooh]

13,000 marks in GB5 multi...
This is an absurd amount. What could possibly require this much computational power (that isn't already run on a supercomputer)?

[Anom]

Joe this is a laptop CPU... it will be 45W in most laptops up to ~70ishW in the big chungus laptops... Yes apple M1 is awesome but no need to go full stupid... Wouldn't mind seeing a laptop running a 200W CPU that you are imagining

[schembfs]

i9-12900HK = 10nm Intel (although Intel calls it Intel 7, it's actually 10nm)
M1Max = 5nm TSMC
Imagine what the same Intel architecture built on 5nm TSMC or a future 5nm Intel process could do. Folks... 5nm is a TSMC innovation.  An M1Max built on an older 7nm or 10nm process won't be able to do what it can do now. M1Max is only possible because of TSMC 5nm process. Without it, it would be on a much bigger die and running a lot hotter.

[Jan Onderwater]

Sooo
A CPU that has a unverified benchmark and is not a product that you can buy, may outperform the M1Max in some tasks.
It is not that Intel has done something like producing a benchmark with a CPU that was connected to an industrial cooler. Nooo, Intel would never do something like that.
I will believe it when its on the market.

[dsadasdsa]

No matter what the haters say, this is a good improvement gen on gen (think it is 40-45% improvement in MT which is massive). And that is what matters.

Now, in regards with Apple M1 Max, the fanboys forget a few things:
- Apple is using 5nm TSMC which is at least node better than the Intel 7 that it is used for ADL. That on its own brings an absolute huge improvement in efficiency; and Apple doesn't have any merit on that; at least Intel is engineering their own process;
- Apple is using ARM ISA which as we know it is more streamlined vs x86, plus it doesn't support ANY legacy app and more so it is a tweaked and improved version of the original ARM v8 ISA, so that on its own is one big reason why M1 max is more efficient; Sure, it is more efficient but you use only new software which is....................; plus you are stuck with Mac OS ecosystem;
- Apple has integrated everything into a very tight package, memory is on the same substrate and everything customized in the OS for this; Compare this to Windows which has to support a million configurations of hardware, old/new software, etc, etc, etc;

What I am trying to say here is that Apple efficiency is explainable, but for it you lose more (in terms of usability) than what you gain (in efficiency); for some cases this is fine, but the majority of people will not like the limitation of the apple platform.

[abc]

I like when people are objectives defined by the therm "haters" from an outside developer of this or that chip company . 1 because need a lot of hours of working 2 because x86 is old and can't compete against scaling vector 3 because humans is so stupid to continue milking "sylicon" nowadays elegant therm to appropriate their hardware product of an xyz company. Intel sylicon, Apple sylicon... wait for AMD sylicon  ;D

[Walkop]

Read about and talked to a few experts in the field of CPU architecture; there is nothing inherently slow about x86 when compared to ARM. When AMD was making their big comeback they were trying to figure out how to approach things in the most efficient way possible, and they knew x86 had massive potential.

This is a very common misconception and is straight up wrong in my understanding.

Now, Intel has been sitting on their laurels hardcore for nearly a decade. That's why Apple and AMD are crushing x86 performance records; not because of x86.

I'm curious for Zen 4. Not really Intel's offering. Memory bandwidth is supposed to be a massive limiting factor that AMD supposedly came up with a great solution for, since the scheduler was always starved for bandwidth (why even in the past Zen systems need good memory). Similar to what Apple has done with M1, ironically; I heard about AMD's new memory solution over a year ago and it sounds like Apple has a very similar approach overall. At least, similar results.

[Fanboy]

So, a laptop, possibly running on battery is 3% slower???
Hm...

[Arturo9]

So it only takes Intel 4 more cores, and 10 more threads, to barely beat the M1 Max and the M1 Max hasn't even shown benchmarks for high powered mode. 

Also, the Intel might be claiming 95 or 125 watts, but we all know it will likely be closer to 250, where the M1 Max is at 30 peak. 

Sorry, but Intel is outright getting embarrassed by the Apple engineers here.

Both Intel and Apple engineers are doing just fine. It's you who's embarrassing yourself with absurd comment.

[Lins]

It turns out that Intel just cannot make their chip design working as well as AMD or Apple on 7nm or 5nm. As a longtime Intel Laptop user, I highly doubt about the long-term performance instead of just a peak performance for the upcoming 12th i9 mobile chip. My current machine is a i9 9980hk, it can only stay at around 3.2 GHz with the fan spinning like a helicopter. While doing benchmarks, it is possible to peak above 4.2 GHz for one minute and that is enough to offer a decent score for advertisement. Though everyone nowadays knows that 5nm, 7nm chips can reduce the heat and raise the performance per watt much higher and keep the high-power mode just much longer, Intel just won't or can't do it. Do you still remember how long Intel has been using the 14nm chip designs? The most embarrassing thing of Intel is that, while they are leading the market, they slow down and stop pushing the performance up. What they care is only making money, but 0% of customer satisfactions. Now they are losing the market and people's trust, they start to be worried and rushes into 10nm, because they are too late to design the 5 or 7nms.