Intel 12900HK is faster than M1 Max for both single and multi cores.

[leman]

And given what you say about the Air and 2/3 of maximum frequency, I think you agree with me more than not. All my other laptop experience is on laptops that have active cooling, and I've never seen throttling that severe (a full 3rd) on any other laptop. I never owned an i9 laptop btw, not many people have, they sell far more lesser CPU's and I buy thin and light, which you can't even get an i9 in.

Again, what you call "throttling" the industry call normal operation with turbo boost. You are not supposed to be able to run boosted all the time. I mean, it's great for yo if you can, but that is not the official spec. There is no laptop on the market that can sustain full turbo on Intel CPUs. The most bulky gaming machines can maybe manage around 70-75% (still 10% over the spec), but that's it.

I do have an i9 desktop and it doesn't throttle. (it's got cooling that can handle it.)

Then you are running it way above the rated spec. How much power does it draw under full load? 300W?

I do not understand why your customized setup that runs the machine above the specification offered by the manufacturer should be the benchmark for the average application.

 

[Fomalhaut]

Again, what you call "throttling" the industry call normal operation with turbo boost. You are not supposed to be able to run boosted all the time. I mean, it's great for yo if you can, but that is not the official spec. There is no laptop on the market that can sustain full turbo on Intel CPUs. The most bulky gaming machines can maybe manage around 70-75% (still 10% over the spec), but that's it.



Then you are running it way above the rated spec. How much power does it draw under full load? 300W?

I do not understand why your customized setup that runs the machine above the specification offered by the manufacturer should be the benchmark for the average application.

We need to define what throttling means on the M1 chips. On Intel chips with turbo boost I understand it to mean that the CPU can no longer support its base frequency on all cores simultaneously.

Apple doesn’t specify the base frequency for the M1 as far as I know, but it clearly runs at lower frequencies when it makes sense to do so. I think Max Tech demonstrated M1 Air downclocking with repeated runs of Cinebench.

 

[bombardier10]

Really Intel's star will be 12 gen Intel i5-12600K - "budget" CPU (260$) faster than
i9-11900K so near M1 Max .

Intel Core i5-12600K mid-range CPU shines again in new benchmarks

Intel's new mid-range Core i5-12600K 'Alder Lake' CPU is up to 46% faster than the AMD Ryzen 5 5600X, more impressive in-person.

www.tweaktown.com

 

[Lihp8270]

I love these threads.

Apple destroy completion : “omg! Look at these leaks. So much power! So good!”

Intel perform better than apple : “got to wait for real world tests”

M1 Pro/Max's GPU performance for gaming is quite poor.

https://news.mydrivers.com/1/791/791715.htm 32 core Max is similar to mobile RTX 3060 for gaming which is quite disappointing. Also there isn't any difference between Roshetta 2 and native binary for the performance btw. At this point, M1 Max's gaming performance sucks after all.

forums.macrumors.com

Apple: “The apple claims with synthetic benchmarks are indicative of what it can do with optimisation”

Intel: “Synthetic benchmarks are meaningless. Needs real world applications”

 

[leman]

M1 Pro/Max's GPU performance for gaming is quite poor.

https://news.mydrivers.com/1/791/791715.htm 32 core Max is similar to mobile RTX 3060 for gaming which is quite disappointing. Also there isn't any difference between Roshetta 2 and native binary for the performance btw. At this point, M1 Max's gaming performance sucks after all.

forums.macrumors.com

Apple: “The apple claims with synthetic benchmarks are indicative of what it can do with optimisation”

Intel: “Synthetic benchmarks are meaningless. Needs real world applications”

Also Intel: "we will run a bunch of synthetic benchmarks with weird configurations and settings, cherry pick some results where our CPU is marginally faster, present them in a very stretched graph and tout it as a resounding victory" 😩

Both Intel and AMD are guilty of massive benchmark machinations. AMD was recently caught enabling broken CPU features in selected games and benchmarks to get higher scores at the increased risk of crashes.

 

[Lihp8270]

Also Intel: "we will run a bunch of synthetic benchmarks with weird configurations and settings, cherry pick some results where our CPU is marginally faster, present them in a very stretched graph and tout it as a resounding victory" 😩

Both Intel and AMD are guilty of massive benchmark machinations. AMD was recently caught enabling broken CPU features in selected games and benchmarks to get higher scores at the increased risk of crashes.

No doubt.

All companies are the same, all marketing departments are the same. Their only goal is to show the product in its best light and sell it.

But my commentary is on the posting trends of this forum rather than any of the products the posts represent.

 

[bobcomer]

Maybe you have a defective unit then, as that doesn’t track with almost every other user I’ve seen.

No, it works fine, I just have a different working set. (VM's, Java apps, browsers, office apps.) It was my mistake in buying something passively cooled.

 

[bobcomer]

Again, what you call "throttling" the industry call normal operation with turbo boost. You are not supposed to be able to run boosted all the time.

No, that's not what I call throttling, I'm actually talking about throttling when the MBA gets well over 100C.

Then you are running it way above the rated spec. How much power does it draw under full load? 300W?

No, I'm not talking about overclocking and turbo at all, just normal operation. When I talk about throttling is when it slows down massively because of heat. Right now I'm typing on my i9 desktop, and it's running slightly over 1Ghz, that's not throttling, it's running that way because the only thing it's doing is running the browser right now (Chrome).

I do not understand why your customized setup that runs the machine above the specification offered by the manufacturer should be the benchmark for the average application.

Mainly because your making assumptions about what I mean by throttling and I'm not running benchmarks, just normal work. I do not overclock, I like stability, and the only thing I've done to this desktop is add more RAM, otherwise it's off the shelf. (It's a Lenovo)

 

[m1maverick]

Like it or not, SPEC (and therefore GB5, if you trust Nuvia's data) is a highly relevant benchmark in that CPU architects use it an awful lot. Say you have an idea for a new branch predictor, and you want to figure out if you should actually use it. Architects evaluating questions like this will run trade studies which predict the changes in area, power, frequency, and performance. Performance predictions are often done by running code traces from SPEC (and other benchmarks, but SPEC is common) in a cycle accurate simulator modified to implement the new idea.

Perhaps, in your desire to see yourself as a wise man above the fray, you have led yourself into distrusting things you should trust.

Or perhaps you're a certain other kind of poster I remember from my own days reading CSMA.

This may be a use for GB but it's being used in a very specific way for a very specific purpose. In your example you're using it in a controlled way by altering one thing (branch predictor) and one thing only (at least I hope that would be the case).

The way GB is generally used by the public is to present one GB score from a disparate number of configurations and declare the system with the highest score the winner. Unfortunately the real world isn't so simple. GB is interesting but I wouldn't make a buying decision on it.

I'd also like to point out that I am not singling out GB, the same can be said of most any benchmark. A benchmark is only representative of how well a system can run that benchmark. I would no more consider Cinebench to be representative of a system which is heavily I/O dependent.

The best benchmark is to run those benchmarks with the applications / workload you need and see how they perform. Or at least benchmarks which are as closely aligned with that workload.

 

[leman]

No, that's not what I call throttling, I'm actually talking about throttling when the MBA gets well over 100C.

Well, yes, that's why it throttles. The sustained power dissipation of M1 Air is around 7-10W. Once it reaches high temperatures it will reduce it's clock to bring down the power consumption to those levels Third-party reviewers observed throttling of around 20% under heavy load (https://www.notebookcheck.net/Apple...le-M1-CPU-humbles-Intel-and-AMD.508057.0.html)

You mention that your M1 gets laggy and sluggish. That is definitely not normal behavior. I have no reason to doubt you, but lagging under load is simply not common experience with these laptops. There must be some third factor that causes this.

No, I'm not talking about overclocking and turbo at all, just normal operation. When I talk about throttling is when it slows down massively because of heat. Right now I'm typing on my i9 desktop, and it's running slightly over 1Ghz, that's not throttling, it's running that way because the only thing it's doing is running the browser right now (Chrome).

Mainly because your making assumptions about what I mean by throttling and I'm not running benchmarks, just normal work. I do not overclock, I like stability, and the only thing I've done to this desktop is add more RAM, otherwise it's off the shelf. (It's a Lenovo)

Apologies if I misunderstood. But you wrote yourself "I do have an i9 desktop and it doesn't throttle. (it's got cooling that can handle it.)" and you previously wrote that you define throttling as running below the maximal clock. Running at or close to maximal clock is definitely not normal operation for any of these CPUs (although it can be done with appropriate power delivery and cooling. Mac Pro can run Xeons at full turbo — takes over 300W thought).

Maybe it would be helpful to the discourse to define what throttling actually means. Here is how I understand throttling (and as far as I know, this is the industry-standard definition): "running below the advertised spec (base clock) under heavy sustained load"

 

[iPadified]

Funny that the same persons are always criticising the M chips. Every metric indicate that Apple have made a very strong, even surprisingly strong, entry into the "PC" CPU/GPU chip market. Can you find exceptions? Sure and the reason is likely unoptimised code such as games rather than poor performance of the Si or that apples are compared with oranges.

Is the Pro/Max chip worse because Intel (finally) spits out something competitive? Hardly.

What I lack in these metrics is fan noise under maximum load and performance under sustained load where the heat generation is balanced by the heat dissipation. It is good that most reviewers run on battery and look for energy usage. At least laptops and supercomputers engineers are interested in energy usage and so should everybody else!

PS. Lying is not the same as making biased statements. A person that does not know the full science behind a technology often makes statements that is not perfectly true. It is natural to be biased, we all are. I want the M1 Pro to be good because I plan to buy a 16 inch so I have a tendency to look for positive reviews. Not good, I know. DS.

 

[magbarn]

Well, yes, that's why it throttles. The sustained power dissipation of M1 Air is around 7-10W. Once it reaches high temperatures it will reduce it's clock to bring down the power consumption to those levels Third-party reviewers observed throttling of around 20% under heavy load (https://www.notebookcheck.net/Apple...le-M1-CPU-humbles-Intel-and-AMD.508057.0.html)

You mention that your M1 gets laggy and sluggish. That is definitely not normal behavior. I have no reason to doubt you, but lagging under load is simply not common experience with these laptops. There must be some third factor that causes this.



Apologies if I misunderstood. But you wrote yourself "I do have an i9 desktop and it doesn't throttle. (it's got cooling that can handle it.)" and you previously wrote that you define throttling as running below the maximal clock. Running at or close to maximal clock is definitely not normal operation for any of these CPUs (although it can be done with appropriate power delivery and cooling. Mac Pro can run Xeons at full turbo — takes over 300W thought).

Maybe it would be helpful to the discourse to define what throttling actually means. Here is how I understand throttling (and as far as I know, this is the industry-standard definition): "running below the advertised spec (base clock) under heavy sustained load"

He needs to check his memory usage. On the 8gb M1 MBA I returned, it would start to slow down and stutter when it was swapping heavily due to the memory being full.

 

[bobcomer]

Well, yes, that's why it throttles. The sustained power dissipation of M1 Air is around 7-10W. Once it reaches high temperatures it will reduce it's clock to bring down the power consumption to those levels Third-party reviewers observed throttling of around 20% under heavy load (https://www.notebookcheck.net/Apple...le-M1-CPU-humbles-Intel-and-AMD.508057.0.html)

You mention that your M1 gets laggy and sluggish. That is definitely not normal behavior. I have no reason to doubt you, but lagging under load is simply not common experience with these laptops. There must be some third factor that causes this.

Curious. You don't think 20% slower is not laggy and sluggish? Maybe I'm just more sensitive to it. (if it's only 20%, it feels even slower to me)

VM's is the third factor I think. they're awfully hard on a machine, especially one that is doing some emulation inside the VM, not to mention its taking cycles away from the host. (and IO, and RAM). My Air is 16G of RAM, and that's pretty much my bare minimum. I don't know if Java apps cause more strain, but I usually have my Midrange Access/development software running. Plus I usually have Chrome and Safari open, with multiple tabs and maybe also firefox. Thunderbird email. If I'm home, the TV app is open to airplay to my TV.

Then in the VM, Domino Notes email client, Excel and Access, A VPN (IKEv2) up and running. Maybe Visual Studio, maybe not. Maybe Word.

On my Windows PC's i can have more than one VM open at a time, but I never even tried on the Air.

And then there's my want of running an emulation package like UTM to run x86 Windows when the software gets a bit better. I reached my highest temp running Windows in a UTM VM. (118F) And no, I didn't leave it running to see how fast it would lower and stabilize the temp.

Apologies if I misunderstood. But you wrote yourself "I do have an i9 desktop and it doesn't throttle. (it's got cooling that can handle it.)" and you previously wrote that you define throttling as running below the maximal clock.

I don't remember I ever said that throttling is defined by running below the maximum clock, but if I did, that was wrong. It's possible I was talking about something like intentional underclocking.. Sorry for the misunderstanding!

Running at or close to maximal clock is definitely not normal operation for any of these CPUs (although it can be done with appropriate power delivery and cooling. Mac Pro can run Xeons at full turbo — takes over 300W thought).

Agreed! Not something I would usually do these days. In my youth, I might have done some overclocking before the age of turbo, but it always caused some instability, and I hate that more than a machine feeling slower than i think it should.

Maybe it would be helpful to the discourse to define what throttling actually means. Here is how I understand throttling (and as far as I know, this is the industry-standard definition): "running below the advertised spec (base clock) under heavy sustained load"

That's fair, but I would add "because of heat" to the end of your definition. The heat based throttling is really all I'm concerned about. And precisely why I never even tried running an i9 laptop -- WAY too many people complaining about sluggish i9's in laptops.

My desktop has a couple of serious fans on it to keep it from throttling (my definition). Yes it makes a lot noise, but noise is acceptable.

In any case, I just ordered a 14" MBP M1 Pro and I'm using the Air as a trade in. The MiniLED, active cooling, more cores, was too much of a draw for me, even with the notch, which I will use that new app to black the menu bar out and probably run more apps full screen. <g>

 

[bobcomer]

He needs to check his memory usage. On the 8gb M1 MBA I returned, it would start to slow down and stutter when it was swapping heavily due to the memory being full.

I have 16G, the max possible, and yes, I pushed it hard.

 

[leman]

Curious. You don't think 20% slower is not laggy and sluggish? Maybe I'm just more sensitive to it. (if it's only 20%, it feels even slower to me)

Well, with M1 20% slower is still faster than most premium laptops, so I doubt that it’s the reason of your experience. There can be many reasons for perceived lag, but usually it has to do with uneven distribution of response times. But these CPUs are so fast that they should be able to produce a response below your perception threshold most of the time, so I believe your experience has other causes.

VM's is the third factor I think. they're awfully hard on a machine, especially one that is doing some emulation inside the VM, not to mention its taking cycles away from the host. (and IO, and RAM). My Air is 16G of RAM, and that's pretty much my bare minimum. I don't know if Java apps cause more strain, but I usually have my Midrange Access/development software running. Plus I usually have Chrome and Safari open, with multiple tabs and maybe also firefox. Thunderbird email. If I'm home, the TV app is open to airplay to my TV.

Then in the VM, Domino Notes email client, Excel and Access, A VPN (IKEv2) up and running. Maybe Visual Studio, maybe not. Maybe Word.

On my Windows PC's i can have more than one VM open at a time, but I never even tried on the Air.

And then there's my want of running an emulation package like UTM to run x86 Windows when the software gets a bit better. I reached my highest temp running Windows in a UTM VM. (118F) And no, I didn't leave it running to see how fast it would lower and stabilize the temp.

It does sound like your workload is rather intense and relies on low quality software (such as Domino Notes). There might really be a lot of reasons why you experience lag, from RAM or cache trashing to rogue threads to hardware interrupt abuse. But I am fairly certain they have to do with software.

That's fair, but I would add "because of heat" to the end of your definition. The heat based throttling is really all I'm concerned about. And precisely why I never even tried running an i9 laptop -- WAY too many people complaining about sluggish i9's in laptops.

There are generally two ways to manage the dynamic performance - either using predefined power consumption threshold or using temperature thresholds. Apple traditionally relied more on the latter - the performance you get is restricted by the thermal capacity of the chassis. For Apple CPUs to reach high temperatures and then clock down is part of normal operation, it’s more efficient than setting arbitrary limits as these limits do not take the environmental factors into account.

The problem with i9 is a bit different - Intel has been stagnating for so long that they were forced to adopt a very aggressive dynamic boost turbo with latest models to get sales. The result are CPUs that consume way much power they have any right to.

 

[bobcomer]

It does sound like your workload is rather intense and relies on low quality software (such as Domino Notes). There might really be a lot of reasons why you experience lag, from RAM or cache trashing to rogue threads to hardware interrupt abuse. But I am fairly certain they have to do with software.

I don't have that type of problems on Windows hardware. (lagging) using the same workload, and Yes, Domino Notes is a pig, but it's a corporate pig. (required) It could be interrupt abuse, but the laptop also gets very hot to the touch when it's lagging. Not sure what you mean by rogue threads, but like I said, the lagging only happens when it's hot.

 

[Pressure]

SKU	Cores P+E/T	E-Core Base	E-Core Turbo	P-Core Base	P-Core Turbo	IGP	Base W	Turbo W	Price
i9-12900K	8+8/24	2400	3900	3200	5200	770	125	241	$589
i9-12900KF	8+8/24	2400	3900	3200	5200	-	125	241	$564
i7-12700K	8+4/20	2700	3800	3600	5000	770	125	190	$409
i7-12700KF	8+4/20	2700	3800	3600	5000	-	125	190	$384
i5-12600K	6+4/16	2800	3600	3700	4900	770	125	150	$289
i5-12600KF	6+4/16	2800	3600	3700	4900	-	125	150	$264

Up to 241 W on the spec sheet for the desktop parts.

 

[Andropov]

Regarding the TDP and power levels (PLs) of the Alder Lake-P (mobile) chips, if this leaked information is to be believed... https://www.tomshardware.com/uk/news/intel-alder-lake-p-and-m-processor-power-limits-listed

	Alder Lake-P (6+8+2, 12900HK?)​	i9 11980HK​
TDP​	45W​	45W​
PL2​	115W​	107-135W​
PL4​	215W​	-​

Assuming the leaked 6+8+2 is the 12900HK (6 efficiency cores + 8 performance cores + level 2 graphics with 96 execution units). There seems to be a new power level 4, using up to 215 watts, for maximum short-burst performance (up to 10ms). The i9 11980HK has a configurable PL2 of 107 watts in the normal profile and 135W on the 'performance' profile. The leaked PL2 of the Alder Lake-P is 115W, likely for the normal profile.

 

[Appletoni]

The new Intel cpu is much faster and that’s only because Apple decided to use only 8 performance cores instead of 32.
In other words: It’s Apples fault.

 

[amartinez1660]

Really Intel's star will be 12 gen Intel i5-12600K - "budget" CPU (260$) faster than
i9-11900K so near M1 Max .

Intel Core i5-12600K mid-range CPU shines again in new benchmarks

Intel's new mid-range Core i5-12600K 'Alder Lake' CPU is up to 46% faster than the AMD Ryzen 5 5600X, more impressive in-person.

www.tweaktown.com

I’ll take this with a grain of skepticism until it’s out, the anandtech’s article SPEC benchmarks posted here by several users show a different picture, the interesting tidbit (extracted also from a previous poster):

“In the SPECfp suite, the M1 Max is in its own category of silicon with no comparison in the market. It completely demolishes any laptop contender, showcasing 2.2x performance of the second-best laptop chip.

The M1 Max even manages to outperform the 16-core 5950X – a chip whose package power is at 142W, with rest of system even quite above that. It’s an absolutely absurd comparison and a situation we haven’t seen the likes of.”

The Intel scores on there are almost half of that AMD’s to boot.

It might happen, who knows, but it means that suddenly the Intel CPU’s after many years of like 5% performance increases is gonna do 200+% in one shot.

 

[bobcomer]

In any case, I just ordered a 14" MBP M1 Pro and I'm using the Air as a trade in. The MiniLED, active cooling, more cores, was too much of a draw for me, even with the notch, which I will use that new app to black the menu bar out and probably run more apps full screen. <g>

I actually ordered a Max, I forgot that I switched it at the last minute. It wont be here for a month and a half, switching it to 32G was the kicker shipping-wise.

 

[dgdosen]

We'll know for sure in a few days, but it looks like the desktop Alder Lake needs 240+ watts to beat the M1X... That increase alone starts to add up to hundreds of dollars per year (~say 200W (300-400 for a system?) x 24 hours / 1000 x 365 days x $0.10 kWh). If I were building a rig, I'd absolutely take electricity costs over the life of the machine into consideration.

 

[magbarn]

We'll know for sure in a few days, but it looks like the desktop Alder Lake needs 240+ watts to beat the M1X... That increase alone starts to add up to hundreds of dollars per year (~say 200W (300-400 for a system?) x 24 hours / 1000 x 365 days x $0.10 kWh). If I were building a rig, I'd absolutely take electricity costs over the life of the machine into consideration.

That’s a bit overblown. Unless you were foolishly trying to mine crypto on a cpu, hardly many of us, even gamers, are going to run a cpu full tilt 24/7. I’m pretty sure an alder lake system with a gold/platinum power supply will be running well under 100 watts the majority of the time.

 

[raknor]

I don't have that type of problems on Windows hardware. (lagging) using the same workload, and Yes, Domino Notes is a pig, but it's a corporate pig. (required) It could be interrupt abuse, but the laptop also gets very hot to the touch when it's lagging. Not sure what you mean by rogue threads, but like I said, the lagging only happens when it's hot.

Is your windows hardware passively cooled? List the model number and specs of said windows hardware. Post a video of both running the same work load with system monitoring tools to demonstrate your claims.

I would guess you have the base model Macbook Air with the 7 core GPU and are comparing it to a massive Core i9 gaming rig.. where you are running a load that the MacBook Air was never designed to run.

 

[bobcomer]

Is your windows hardware passively cooled?

Nope! Nothing in the last 20+ years or so, I wouldn't even think of trying it. The only reason I got the MBA is because it was cheaper than the MBP and I wasn't sure I'd like it. I shouldn't have cheaped out in the first place.

I would guess you have the base model Macbook Air with the 7 core GPU and are comparing it to a massive Core i9 gaming rig.. where you are running a load that the MacBook Air was never designed to run.

Nope. 16G RAM, 1 TB SSD, 8 core GPU. And no, not comparing it to a gaming rig, I don't game on PCs, so don't have any gaming rigs. (though my desktop is an i9, it's only a middle of the road one, but it does have lots of RAM, it makes running multiple VM's so much nicer.)

But yes, my normal workload didn't fit the MBA, hence why I said it was a mistake in buying it. It got way too hot way to much of the time. I expect the M1 Max 32G that I ordered will be as fast as my i9 desktop a lot of the time, it has the cooling and enough RAM to use a couple VM's at once, plus all the other stuff I run.

I already erased the MBA in preparation for sending it back as my trade in, and I don't expect my new MBP until mid December.