Intel Alder Lake vs. Apple M1

[Technerd108]

Watching the D2D video with Alder Lake confirms exactly what I thought. They can get the performance but they have to crank everything up including heat. For Desktop as Dave says it is not as much a problem if you can adapt to the extra cooling needed and have wired power source.

What happens when you put Alder Lake in a laptop. It will be impressive no doubt but it will still have thermal and wattage issues. Gaming laptops will be ovens and ultra lights will run hot as well and none of the laptops will get the advertised performance under load for long.

Intel is heading in the right direction but their architecture and process node are a disadvantage right now. Eventually I believe they will be able to turn things around but it is going to take them a while-maybe two years before they get both performance and heat as well as battery life under control.

 

[Kpjoslee]

What happens when you put Alder Lake in a laptop. It will be impressive no doubt but it will still have thermal and wattage issues. Gaming laptops will be ovens and ultra lights will run hot as well and none of the laptops will get the advertised performance under load for long.

Looking at the Alder Lake tested at 6+8 (which is going to be top Alder Lake laptop SKU) at 35w performing favorably against M1 Max at CB23. It feels like Alder Lake is designed for the laptops first and decided to just feed the extra wattage on desktop variants to compete against top AMD desktop SKUs.

 

[Taz Mangus]

Looking at the Alder Lake tested at 6+8 (which is going to be top Alder Lake laptop SKU) at 35w performing favorably against M1 Max at CB23. It feels like Alder Lake is designed for the laptops first and decided to just feed the extra wattage on desktop variants to compete against top AMD desktop SKUs.

CB R32 is not a good measure as it seems to not push all the cores if the M1 Max properly. There is speculation that the program not working correctly for the M1 Max.

 

[Ceed]

I realize a handful here want to believe Apple is some exclusive luxury brand, but people who work at McDonald's are buying iPhone Pro Maxs, AirPods Pro, and MacBook Pros without much stress about it. It is, through and through, a product for the commoner.

It’s the best decision Apple could’ve made for 99% of their users.

That seems about right, this decision is mostly attractive to their current loyal following. They haven't been doing so well with the recent Intel designs and this will improve that situation tremendously. As for "the majority of users", the majority of computer users are on PCs, including the majority of creative professionals (step away from campus or California). Whether this new platform will attract more of them (and it may), time will tell. I personally do not anticipate the curve bending upward all that exponentially, though. I honestly can see a situation where they might shed more than they gain with this dramatic changing of guts, especially in business sectors who depend on stability and certainty.

 

[Technerd108]

Looking at the Alder Lake tested at 6+8 (which is going to be top Alder Lake laptop SKU) at 35w performing favorably against M1 Max at CB23. It feels like Alder Lake is designed for the laptops first and decided to just feed the extra wattage on desktop variants to compete against top AMD desktop SKUs.

Again more cores equal more heat. I have yet to see any intel cpu stick to a hard wall on wattage. All Intel mobile cpu's use way more power when plugged in than the advertised TDP. So plugged in with proper cooling they may be competitive until they start to throttle. On battery forget it-The M1 Pro or Max on battery on heavy load sustained for a long period of time will crush Alder Lake!! Mark my words! Sure a single core Geekbench test run once then Intel may have an advantage when the device is plugged in. Alder improves a lot the battery life when the cpu is doing light tasks but under load it suffers the same problems Intel has had for a while.

 

[crazy dave]

Looking at the Alder Lake tested at 6+8 (which is going to be top Alder Lake laptop SKU) at 35w performing favorably against M1 Max at CB23. It feels like Alder Lake is designed for the laptops first and decided to just feed the extra wattage on desktop variants to compete against top AMD desktop SKUs.

Even for CB23, I’d like to see that confirmed as I’m a little dubious it’s that good at 35 W *turbo* - 35W base …. maybe. If ADL i9 desktop is that efficient at low wattage then the i5 configuration makes no sense, which scores only 18% better for 5x the power draw in a regime where increasing performance shouldn’t be increasing power draw by that much. That would say Intel could’ve made an desktop i5 with both better performance and perf/W without much if any increase in cost and then chose not to.

As you can see in this post:

https://twitter.com/i/web/status/1455961649282752525

Dropping off Turbo below 150W starts to have a near linear effect on CB23 i9 performance and it should start dropping off faster than that the lower you go. Now I grant you that the laptop silicon shouldn’t behave the same way and turning off 2 P-cores should also have a big bump in MT efficiency, but this was a desktop part and dropping those cores also drops the performance score a ton.

Maybe I’m wrong though. After all, he has it to actually test, I don’t and am extrapolating from numbers. Regardless, we’ll find out in a couple of months when laptops drop or until another reviewer does the same.

 

[Technerd108]

I realize a handful here want to believe Apple is some exclusive luxury brand, but people who work at McDonald's are buying iPhone Pro Maxs, AirPods Pro, and MacBook Pros without much stress about it. It is, through and through, a product for the commoner.



That seems about right, this decision is mostly attractive to their current loyal following. They haven't been doing so well with the recent Intel designs and this will improve that situation tremendously. As for "the majority of users", the majority of computer users are on PCs, including the majority of creative professionals (maybe step away from campus or California). Whether this new platform will attract more of them (and it may), time will tell. I personally do not anticipate the curve bending upward all that logarithmically, though. I honestly can see a situation where they might shed more than they gain with this dramatic changing of guts, especially in business sectors who depend on stability and certainty.

I completely agree with most of your points except the last. Business cares solely about cost. I don't think Apple will lose customers due to this redesign. I think it will spur a lot of people to upgrade that already own a Mac of some kind just like you were saying whether in business or personal. A lot of Mac users are going to upgrade and from a business perspective it is a write off for those who need it. Large businesses are not going to deploy these machines. These are for small businesses, individual entrepreneurs, creative pros, and possibly big digital art studios but not for mainstream business. So I don't see it doing anything but increasing sales. These are huge upgrades for those that need them and those that want them!!

 

[aevan]

I realize a handful here want to believe Apple is some exclusive luxury brand, but people who work at McDonald's are buying iPhone Pro Maxs, AirPods Pro, and MacBook Pros without much stress about it. It is, through and through, a product for the commoner.

The idea that someone believes Apple is some luxury brand usually doesn't come from Apple users. It is usually an attempt to discredit Apple users as "elitist" and not as professionals who chose these machines because they are the best option for their work.

Macs are a tool. And yes, a lot of people buy them and use them. It is a product for the "commoner" as much as it is for a professional. Just like PCs.

That seems about right, this decision is mostly attractive to their current loyal following. They haven't been doing so well with the recent Intel designs and this will improve that situation tremendously. As for "the majority of users", the majority of computer users are on PCs, including the majority of creative professionals (maybe step away from campus or California). Whether this new platform will attract more of them (and it may), time will tell. I personally do not anticipate the curve bending upward all that logarithmically, though. I honestly can see a situation where they might shed more than they gain with this dramatic changing of guts, especially in business sectors who depend on stability and certainty.

Why the "loyal" following? You don't have to be "loyal" to a brand to appreciate these machines. You're trying to paint this as if some niche diehard group of fans is excited about this - and not people who are looking for the best tools for the job. There are a lot of people who are not loyal to any company who will see these laptops for what they are: best in class.

Majority of users are on PC, no doubt about that (mostly due to price, let's be honest) - but if anything gets them on the Mac, it will be the performance of these machines. But either way, Apple was never about chasing market share. Yes, of course, PC will always dominate the market share, just like Android dominates the phone market compared to iPhones.

You could also look at other metrics, like customer satisfaction. Most of the people I know from the creative industry are on PC, true, and I don't know a single person who is actually a fan of them (I've heard the term "necessary evil" to describe Windows more times that you'd think).

At the end of the day, Apple doesn't have to make the most popular laptop - that title will always go to the "best value" products. Even in their own lineup of devices - the most popular iPad is the regular iPad - by far. But it certainly isn't the best iPad. Same with computers - Macs won't be the most popular computers and they don't have to be. They just have to be among the best computers - and I'd say that, for many people, Apple just made the best laptops there are. Loyal or not.

 

[Ceed]

Why the "loyal" following? You don't have to be "loyal" to a brand to appreciate these machines.

You're right, I'm not loyal and I can definitely appreciate the work Apple has done here. These are some seriously impressive SoCs. However, I'm already comfortable with my workflow, and this is a rather virgin platform. I'll take a wait-and-see approach, and I think a lot of us are in the same boat. ARM for desktop computing is a nascent technology that is only going to get more competitive (beyond Apple) and mature in the coming years.

 

[C-Dubs]

I'll be getting something lighter that will run Windows apps, that happens to be almost as fast at most things. And most likely it will be cheaper too. Fair tradeoff!

I think the ‘running windows apps’ would be the main reason to not care about the M1 performance revolution. You can argue that your PC might be cheaper, but the amount of power draw means you’d probably end up paying way more for the PC over the course of its shortened lifespan.

 

[C-Dubs]

Lol. The person who posted that people choose windows because of ‘stability and reliability’ I guess doesn’t use windows much. If you need that, you go Linux

 

[Leifi]

I don't understand this obsession with power efficiency in isolation to overall perfromance... M1 Max is less power efficient than M1 just as Alder Lake is less power efficient than M1 Max. But a snapdragon 662 is even more power-efficient than an M1 etc.

Would all people here choose an M1 laptop purely because it is more power-efficient than an M1 Max? Just as they would choose an M1 max because it is more power-efficient than Intel's faster CPUs. It seems hypocritical to only compare power efficiency with faster higher-core CPUs and not between the M1 versions...

 

[Leifi]

Lol. The person who posted that people choose windows because of ‘stability and reliability’ I guess doesn’t use windows much. If you need that, you go Linux

LOL.. I guess he compared the realibility with MacOS not with Linux.. Its becoming hard to argue that MacOS is more stable and reliable than Windows these days...

 

[bobcomer]

I think Apple is more concerned about profits, they are probably the most profitable laptop manufacturer on the market, even with a smaller market share (they are 4th in laptop shipments and growing, though, so I wouldn't call them "small").

I don't doubt you're right.

I'm sure you understand how, from their business perspective and the perspective of the majority of their customers, the decision to switch to their own silicon is a game changer.

Actually no, I don't understand why they did that, even for a majority of their customers, but I don't have to -- it's just a quirk of my own that i pay attention to Apple for personal reasons. Their change hurt my usage, that can't be argued about. I don't think Apple is going to be hurt overall by their decision, but I also think it wont help them either, they'll continue on as they have. Losing some and gaining some...

 

[Technerd108]

I don't understand this obsession with power efficiency in isolation to overall perfromance... M1 Max is less power efficient than M1 just as Alder Lake is less power efficient than M1 Max. But a snapdragon 662 is even more power-efficient than an M1 etc.

Would all people here choose an M1 laptop purely because it is more power-efficient than an M1 Max? Just as they would choose an M1 max because it is more power-efficient than Intel's faster CPUs. It seems hypocritical to only compare power efficiency with faster higher-core CPUs and not between the M1 versions...

In my uneducated opinion These are the first consumer chips that can give both. Sure scaling and cores matter but M1 series is the first where you can have very high performance and at the same time have very high efficiency and thermal profiles. So now you can do serious tasks for a long time on a battery which before these chips was simply not really possible.

I like the design of the M1 Pro and Max because they are suited perfectly for the use of each. M1 Max gives about 50-80% worse battery life than the M1 Pro under heavy load but makes sense because of higher bandwidth, more memory and more GPU. Both give huge battery life considering the performance.

Intel doesn't have any faster CPU's at the moment. Certainly not in the mobile space. And while their cpu's are not faster they are much LESS power efficient. No matter the metric Intel cpu's lose. Alder Lake is not out yet. Once it is and it is implemented in real devices we will see. One thing is for sure the M1 Pro and Max out perform on both performance and power and heat efficiency. Intel Alder Lake which is not out for consumer products yet under very specific conditions using desktop processors probably water cooled beat the M1 Pro and Max. That really doesn't bode well for Intel. At least for right now...

 

[bobcomer]

I think the ‘running windows apps’ would be the main reason to not care about the M1 performance revolution.

Yep. And I wouldn't call it a revolution in any way, it's just the latest bump in performance/power savings -- the idea behind it is still the same old digital computer. We haven't seen true innovation yet and I'm not sure I ever will.

You can argue that your PC might be cheaper, but the amount of power draw means you’d probably end up paying way more for the PC over the course of its shortened lifespan.

Shortened life span, no, most PC's last quite a long time, no different than Macs -- mostly the same components. As for power draw, have you actually calculated the difference in cost for power? Not so much. Cheaper up front is still cheaper overall.

 

[toobravetosave]

In my uneducated opinion These are the first consumer chips that can give both. Sure scaling and cores matter but M1 series is the first where you can have very high performance and at the same time have very high efficiency and thermal profiles. So now you can do serious tasks for a long time on a battery which before these chips was simply not really possible.

I like the design of the M1 Pro and Max because they are suited perfectly for the use of each. M1 Max gives about 50-80% worse battery life than the M1 Pro under heavy load but makes sense because of higher bandwidth, more memory and more GPU. Both give huge battery life considering the performance.

Intel doesn't have any faster CPU's at the moment. Certainly not in the mobile space. And while their cpu's are not faster they are much LESS power efficient. No matter the metric Intel cpu's lose. Alder Lake is not out yet. Once it is and it is implemented in real devices we will see. One thing is for sure the M1 Pro and Max out perform on both performance and power and heat efficiency. Intel Alder Lake which is not out for consumer products yet under very specific conditions using desktop processors probably water cooled beat the M1 Pro and Max. That really doesn't bode well for Intel. At least for right now...

Here are notebookcheck's battery tests on the m1 pro. Definitely more efficient than say a similarly powered gaming laptop unless under load

 

[crazy dave]

I don't understand this obsession with power efficiency in isolation to overall perfromance... M1 Max is less power efficient than M1 just as Alder Lake is less power efficient than M1 Max. But a snapdragon 662 is even more power-efficient than an M1 etc.

Would all people here choose an M1 laptop purely because it is more power-efficient than an M1 Max? Just as they would choose an M1 max because it is more power-efficient than Intel's faster CPUs. It seems hypocritical to only compare power efficiency with faster higher-core CPUs and not between the M1 versions...

Because M1 Max is actually more efficient than the Snapdragon and as if not more efficient than the M1 depending on the task? Efficiency is best measured in joules, or the amount of energy it takes to complete a task, rather than watts. Although watts can be important too for heat dissipation and throttling. Previous generations of Intel chips for instance could offer the same performance as the M1 CPUs but at terrible wattage so the joules they used were off the chart. Literally in the case of some of Anandtech’s charts where the Apple silicon was so much more efficient the Intel/AMD chips couldn’t be displayed on the same graph.

This means for laptops (MacBooks), AIO (iMac), and SFF (minis), Apple can offer better performance for the thermally constrained environments and keep them quiet in operation. Even more importantly for MacBooks it gives them great battery life. The 16” M1 Max model gets the same battery life as the 13” M1 but for substantially higher performance at the cost of portability. An Intel chip would decrease the portability further and still have poorer battery life under load.

Now it should be stated that Alder Lake is a massive improvement in this regard over Intel’s previous generations. Not enough in my opinion to make up the difference which is still substantial. For instance there’s an earlier post in which a video review talks about operating desktop Alder Lake in thermally constrained environments and having to use either exotic cooling or suffer from throttling reducing their actual performance. It also means that when Apple increases its core counts for the eventual Mac Pro, it can stick 4 times the number of P-cores in those chips in the same or smaller thermal envelope as an i9 and on something like CB23 it will get scores of well over 30,000. And do even better in even more sensible benchmarks.

This does however also highlight a deficiency in Apple’s strategy which has been true even before the switch to their own silicon: Apple only makes a relatively small number of product lines that can’t overlap and there can sometimes be big gaps between the lines. The difference in performance between the M1 CPU and M1 Pro/Max is 50% - for the GPU it’s 2-4x! The gap between the Pro/Max and whatever comes next may be similarly large. Meanwhile AMD and Intel make a dozen different chips that can be then paired with a huge range of discrete GPUs for a huge range of configurations and price points. This means that especially for something like a desktop tower, sure it may be bigger and/or louder than an iMac with an M1 Max, but likely will it run faster and be cheaper even if paired with a decent if not phenomenal screen (which you can keep if you change towers). And fundamentally they’re both big desktops even if the iMac is “sleeker”. And the only option to get something faster for the Mac is to then get the Mac Pro which ends up overshooting both in price and performance for most people. This issue is less prevalent for the laptop line where every OEM is constrained by thermals, weight, and battery life. But it’s still an advantage of the PC platform (there are disadvantages to having this jungle of devices but that’s another post).

Sorry for the wall of text but this is a huge subject with a lot of nuances and … you did ask.

 

[NotTooLate]

Nah the laughable bit is that Mac buyers,myself included, have had to do a lot of hand waving to excuse Apples middling performance with extremely high prices for years. This paradigm change with the M1 chips brought out a lot of hilarious people strutting around here. Now that they have had their performance feather plucked from their hats we’re on to perf per watts wanking. Cute story though.

What are you talking about … you compare a 250w part to a SoC that goes into a laptop , when was that ever a thing ? Did ppl ask why their xps laptop doesn’t have a 5950x instead of a 5800h ? The fact you are unable to grasp the reason as to why it’s important in the mobile space doesn’t make it untrue , also fyi , if I would buy a x86 desktop , it would still be AMD today and not Intel.
But I guess your entire comment has no real technical or meaningful contribution aside from name calling.

 

[Ceed]

Doesn't Intel have a lot more headroom to improve, though? They are at 10(7)nm and could essentially cut that in half, if they would find the talent to actually do it. No wonder the M1 is so much more thermally efficient, they hired well.

 

[leman]

Little interesting Cinebench R23 score results.
Alder Lake i9 12900k limited at 35w and disabled 2 P-cores to simulate laptop version of Alder Lake with 6+8.
vs M1 Max at 30w

12900k - 14288
M1 Max - 12326

https://twitter.com/i/web/status/1456966531460501507

By the way, using this as a projection: this R23 score represents a ~45% reduction in performance agains the full i9 12900K. If we apply the same to the SPEC scores, we are in the ballpark of 44 (int) and 45 (fp), which would be 20% and 45% slower than M1 Pro, respectively.

I think the final binned mobile i9 at 45W will be more or less as fast as M1 Pro in SPECint (maybe 5-10% slower) and around 30-40% slower in SPECfp. But it will have a slight edge in short burst workloads, and of course, Cinebench.

 

[NotTooLate]

By the way, using this as a projection: this R23 score represents a ~45% reduction in performance agains the full i9 12900K. If we apply the same to the SPEC scores, we are in the ballpark of 44 (int) and 45 (fp), which would be 20% and 45% slower than M1 Pro, respectively.

I think the final binned mobile i9 at 45W will be more or less as fast as M1 Pro in SPECint (maybe 5-10% slower) and around 30-40% slower in SPECfp. But it will have a slight edge in short burst workloads, and of course, Cinebench.

What ppl are still missing is that 1) You need to compare similar products to similar products , I cannot understand why dont we throw a thread ripper in the mix as well , it blows M1X in those heavy multithreaded workload , I hope we dont do it because it doesn't make any sense , the same way comparing a power hungry desktop vs a laptop SoC that can perform to its max potential on battery does not make sense.

Once alder lake is in a laptop , lets see the comparisons then !

Its also funny that ppl pick and choose their "best performing" benchmark and call it a day , its if ppl with the new M1 Max will run prores workload and say that's the differences between the machines (which is not) , I am not sure when Cinebench became the defacto benchmark across the internet , for sure Spec is a much better benchmark , as it actually checks a lot of aspects of the Uarch , but alas , you cannot run it as easily as you do those GB and CB.

I will wait for both the Intel laptop offering and Apples high end desktop machines to do apples to apples comparisons.

 

[Kpjoslee]

By the way, using this as a projection: this R23 score represents a ~45% reduction in performance agains the full i9 12900K. If we apply the same to the SPEC scores, we are in the ballpark of 44 (int) and 45 (fp), which would be 20% and 45% slower than M1 Pro, respectively.

I think the final binned mobile i9 at 45W will be more or less as fast as M1 Pro in SPECint (maybe 5-10% slower) and around 30-40% slower in SPECfp. But it will have a slight edge in short burst workloads, and of course, Cinebench.

Memory bandwidth plays bigger role on SPEC scores than Cinebench which is largely uninfluenced by memory bandwidth. So there is a good chance we will see a smaller reduction in performance on SPEC score.

 

[leman]

What ppl are still missing is that 1) You need to compare similar products to similar products , I cannot understand why dont we throw a thread ripper in the mix as well , it blows M1X in those heavy multithreaded workload , I hope we dont do it because it doesn't make any sense , the same way comparing a power hungry desktop vs a laptop SoC that can perform to its max potential on battery does not make sense.

Once alder lake is in a laptop , lets see the comparisons then !

Its also funny that ppl pick and choose their "best performing" benchmark and call it a day , its if ppl with the new M1 Max will run prores workload and say that's the differences between the machines (which is not) , I am not sure when Cinebench became the defacto benchmark across the internet , for sure Spec is a much better benchmark , as it actually checks a lot of aspects of the Uarch , but alas , you cannot run it as easily as you do those GB and CB.

I will wait for both the Intel laptop offering and Apples high end desktop machines to do apples to apples comparisons.

I think it’s at least interesting to compare these products at custom configs to get an idea what performance we can expect. Of course, you have to plan in a certain margin of error. And sure, you can throw a Theeadripper into the mix if you manage to run it on 40W, but that would be less interesting, as it’s a completely different product. We already know that mobile i9 ADL will use a 6+8 config at 45W, so whats wrong with emulating this SKU using the same chip? Of course with the caveat that the final mobile SKU is likely more efficient.

Regarding benchmarks, I prefer SPEC for obvious reasons. Cinebench is too one-sided and has already been demonstrated to be problematic (it does not properly represent the CPU performance in wide range of workloads). Geekbench is actually a not too bad proxy for SPEC, but it overestimates sustained performance since it’s too short.

 

[Zwopple]

Doesn't Intel have a lot more headroom to improve, though? They are at 10(7)nm and could essentially cut that in half, if they would find the talent to actually do it. No wonder the M1 is so much more thermally efficient, they hired well.

By the way, is anyone here environmentally conscious enough to self-sacrifice? That is, sacrifice your time and use a base M1 for the good of the Earth.

On a pure NM metric the number doesn't matter that much anymore. Also getting your new fab up and running on a new process takes a lot of money and time. TSMC is not sitting still by any means and you'll see continued gains there too.

Intel literally just sat on their butts for waaaay too long thinking nobody would come around and knock them but TSMC is showing them just how complacent they became.