Surface Laptop (Intel and AMD) vs M1 MacBook Air. Oh my...

[TheReal1980]

Summary

In a new video by Max Tech, three different laptops are tested against each other.

Apple Macbook Air (M1)

Microsoft Surface Laptop 4 (Intel)

Microsoft Surface Laptop 4 (AMD)

 

Here are some results.

 

 

 

 

 

Quotes

Quote

In this video, we compare the M1 MacBook Pro to Two Microsofts Surface laptop 4's the 16GB AMD surface laptop 4 vs 8GB Intel Surface laptop 4. We test everything from Geekbench and Cinebench R23 Benchmarks for CPU and GPU Metal and Open CL performance, Gaming performance with GFXbench, Lightroom Classic photo editing, Davinci Resolve 17 4K video editing, with Performance when unplugged and on battery power, web browsing performance with Speedometer 2.0, single-core, Multi-core, Fan noise, CPU temps and overheating/thermal throttling, smoothness & More!

 

My thoughts

I knew the M1 would be fast but not so much faster compared to the Surface. The BIG problem I saw is that running the AMD (and somewhat Intel) on just battery will kill the performance. Intel also lost a little in performance but not nearly as much as AMD and it ran faster than the AMD in multiple tests.

M1 is no joke, we all knew Intel was dripping in their pants but I think AMD just got a little moist as well.

 

Sources

 

 

[jaslion]

I mean nothing unexpected here. The air isn't thermally limited in such short runs and it is current gen tech. Amd and intel stuff is last gen stuff already.

 

Still impressive for a arm chip and excited to see where it goes.

[jaslion]

1 minute ago, Benji said:

I have no idea how a comparison between X CPU and the Apple M1 qualifies as "news" anymore. There are comparisons with other devices that run these chips as well already (1185G7) and the 4980U is just a slightly modified 4800U. We also know that Apple's WebKit is far superior in comparison to anything else performance-wise, and that has been a thing for years already.

Well, given the fact that apps with certain x86 codes doesn't run and them being rumoured to just not be included in their design, I wonder how much "Arm" these things actually still are. They might be drastically different at this point in time for all we know. After all, Apple calls them "Apple Silicon" and not explicitly "Arm".

Fair enough. I'm well aware of the extremely limited support the m1 macs currently have especially with drivers as many peripherals simply do not work at all.

[Helpful Tech Witch]

this was barely news the week after the MacBook air came out.......

now its just another day as a computer enthusiast

[Zodiark1593]

8 minutes ago, Benji said:

I have no idea how a comparison between X CPU and the Apple M1 qualifies as "news" anymore. There are comparisons with other devices that run these chips as well already (1185G7) and the 4980U is just a slightly modified 4800U. We also know that Apple's WebKit is far superior in comparison to anything else performance-wise, and that has been a thing for years already.

Well, given the fact that apps with certain x86 codes doesn't run and them being rumoured to just not be included in their hardware design, I wonder how much "Arm" these things actually still are given the fact that they can run x86 programs at almost original speed, but with way slower clockspeeds. They might be drastically different at this point in time for all we know. After all, Apple calls them "Apple Silicon" and not explicitly "Arm".

Calling the M1, Apple Silcon is an entirely fair assessment. Apple doesn’t license ARM cores, but rather, the ISA. The CPU cores are 100% Apple’s design that is compatible with the ARMv8 ISA, and in fact, is known that Apple has added at least one extension set (Apple AMX) that doesn’t exist in ARM’s ISA. 

[TheReal1980]

Don't know if anyone who posted in the thread watched the video but the AMD model had, in one test a reduction in performance of about 50% when running it on battery compared to charger, that is totally unacceptable.

I thought AMD would be this fantastic comeback of a company with great hardware but something feels shady.

[Dabombinable]

2 minutes ago, TheReal1980 said:

Don't know if anyone who posted in the thread watched the video but the AMD model had, in one test a reduction in performance of about 50% when running it on battery compared to charger, that is totally unacceptable.

I thought AMD would be this fantastic comeback of a company with great hardware but something feels shady.

Something is not quite right with the power limits.

[Kisai]

3 hours ago, TheReal1980 said:

Don't know if anyone who posted in the thread watched the video but the AMD model had, in one test a reduction in performance of about 50% when running it on battery compared to charger, that is totally unacceptable.

I thought AMD would be this fantastic comeback of a company with great hardware but something feels shady.

 

That's because the x86-64 chips cheat on TDP. On battery, you can't max-out the cpu because the battery will last 5 minutes and be hot enough to cook an egg on. So what happens is they alternate between full power and restricted power in cycles to avoid this. Sure an advertised speed/turbo is achievable... but only on AC. On Battery, even if Windows power management is told otherwise, it's going to avoid pulling so much energy from the battery that the battery immediately fails to supply it.

 

That's why in "thin and light" laptops, usually the second you pull the usb-c PD charger, the laptop switches to the lowest brightness and cpu performance mode so it can lower the amount of fan noise since the fans also consume energy.

[xtroria]

As someone who owns intel mbp 13, i tried opening safari on m1 mac and it made me cry

 

The performance jump made old macbooks feels like ancient machine

[leadeater]

6 hours ago, TheReal1980 said:

Don't know if anyone who posted in the thread watched the video but the AMD model had, in one test a reduction in performance of about 50% when running it on battery compared to charger, that is totally unacceptable.

I thought AMD would be this fantastic comeback of a company with great hardware but something feels shady.

That's just simply down to how the OEM configured it, other laptops can be entirely different when running on battery. There are Intel examples in the past as well with drastic performance drops on battery. You tend to see this on laptops with relatively small batteries and/or poor cooling for the battery so they power limit to avoid issues.

 

I believe the Asus ZenBook isn't as bad in this respect, also comes with Ryzen Mobile 5000 U series too.

[Spindel]

6 hours ago, TheReal1980 said:

Don't know if anyone who posted in the thread watched the video but the AMD model had, in one test a reduction in performance of about 50% when running it on battery compared to charger, that is totally unacceptable.

I thought AMD would be this fantastic comeback of a company with great hardware but something feels shady.

That is because the TDP number on AMD and Intel CPUs is a complete fantasy (if they say 25 W it really isn't 25 W if you want full performance). 

 

Meanwhile the M1 Mini draws around 35 W from the wall socket, when the CPU and GPU runs full tilt. And remember it is 35 W from the wall the full system draws 35 W under full load not just the CPU. 

[leadeater]

1 hour ago, Spindel said:

That is because the TDP number on AMD and Intel CPUs is a complete fantasy (if they say 25 W it really isn't 25 W if you want full performance). 

No this is wrong, you can install Intel XTU or use Ryzen Master and a 15W or 25W TDP chip under a sustained load will use no more than this, or really the PL1 value which is the TDP (almost always).

 

The reason for the drop in performance is because the battery and applicable circuitry simply isn't capable of supplying the current safely so when on battery power settings get adjusted automatically like peak current, turbo time length, turbo power (PL2/PPT) etc.

 

It has literally nothing to do with TDP being wrong, which is not the case, otherwise the plugged in performance should be equally limited due to cooling. These slim devices can only cool so much and being on battery or off battery does not change this and the cooling must be adequate for either. For slim devices cooling is the first and primary factor for performance.

 

So there is only two reasons to lower performance on battery, increase run time or the battery simply cannot provide the current. 

 

See below review and comments for the Asus ZenBook

Quote

Like our previous U-series processor testing, today we are benchmarking the Ryzen 7 5800U in both its 15W and 25W power configurations (long term power limits). However, as with all laptops tested, there's also a boost period, in the case of this laptop up to about 30W.

Multicore workloads will be power limited to 15W or 25W depending on cTDP configured. You will not see more than this simply by plugging in to power.

 

Quote

For this review, we should also note that we re-tested the Ryzen 7 4800U with boost behavior that closely matches the 5800U in this ZenBook 13, just so we could have the absolute best apples-to-apples comparison between AMD processor generations.

 

 

R23 Multi Only:

Spoiler

 

https://www.techspot.com/review/2235-amd-ryzen-5800u/

 

And as you can see there is next to zero performance difference between 15W or 25W cTDP for a single core workload as the power required for a single core is low.

 

The Apple M1 TDP is 15W, Ryzen Mobile U is also 15W with the ability on less thin/better cooled devices to configure it to 25W but the cooling has to actually be able to handle it.

 

P.S. Plugged in the AMD Surface laptop is running 25W cTDP, damn sure based on the benchmark figures.

[PeachGr]

(I don't take the part of anyone here)

If I would be on the market for a laptops that costs about 1500€ on my country, I d like a comparison of M1 vs a similar priced laptop that has a GPU in it, even a 1650. CPU vs arm CPU is not really useful for me

[n0stalghia]

4 hours ago, xtroria said:

As someone who owns intel mbp 13, i tried opening safari on m1 mac and it made me cry

 

The performance jump made old macbooks feels like ancient machine

The only result that matters, to be honest. Who cares about benchmarks. A user here tried the tech and was able to feel the difference, so he's a potential buyer now. Job successful.

[Spindel]

4 hours ago, xtroria said:

As someone who owns intel mbp 13, i tried opening safari on m1 mac and it made me cry

 

The performance jump made old macbooks feels like ancient machine

I've naged about this in a lot of threads about the M1. 

 

As an owner of an M1 Mini one thing no benchmark can't communicate and most (if not all reviews) do not communicate is the over all responsiveness of the system when using it. I understand its omitted because it is a thing that is hard to quantify, but no computer I have ever used before has the same feeling of responsiveness when actually using the computer. 

[hishnash]

15 hours ago, Benji said:

They might be drastically different at this point in time for all we know. After all, Apple calls them "Apple Silicon" and not explicitly "Arm".

Apple only have an ARM instruction set license it is by definition not at all related to the ARM Cortex IP. The Apple Silicon CPU has as much ARM Cortex IP as AMDs CPUs have Intel IP they only share the instruction set. Apple also extends the the instruct with a load of their own things. 
 

It is also worth noting that the CPU is not running x86 applications, the x86 apps are translated in advance into an extended version of the ARMv8.4 instruction set, this is possible since apples cpus are able to switch the memory access mode so that memory operations on their cpus act the same as they do on x86 systems. This means the translation from x86 to ARM is quite simple since arm is a RISC instruction set every x86 instruction as 1 or more (sometimes many) corresponding ARM instructions, Rosseta2 does this before the app starts for the first time and writes to disk the output that is an ARM64 application that runs with this x86 memory mode. The reason that MS are not able to do this is all down to this x86 memory mode, without it to run x86 applications that have more than one thread you need to intercept all memory operations and add lots of sleeps etc. 

 

2 hours ago, Spindel said:

is the over all responsiveness of the system when using it.

This is all down to the 0 copy when it come to passing things off to the GPU for rendering you can just pass the pointer in memory that has the image and it renders it, this is a massive benefit for things like system UI were the cpu sometimes renders some of the output and the GPU renders other parts and then the GPU composite it all tougher with both CPU and GPU being able to rw to the memory just the same as 2 cpu cores operating on the memory the latency is massive reduced. (note intel cpus with integrated GPUs use the system memory but are not able to use shared pointers in the same way so almost all memory passed to intel integrated GPUs needs to be copied, taking up twice the space it needs to in memory). 



 

[HenrySalayne]

8 hours ago, Kisai said:

On battery, you can't max-out the cpu because the battery will last 5 minutes and be hot enough to cook an egg on. So what happens is they alternate between full power and restricted power in cycles to avoid this. Sure an advertised speed/turbo is achievable... but only on AC. On Battery, even if Windows power management is told otherwise, it's going to avoid pulling so much energy from the battery that the battery immediately fails to supply it.

Do you have any sources for these claims? Modern batteries can supply insane currents. Most power supplies, on the other hand, are underpowered and the battery is used for current peaks. I had several notebook draining the battery under high load while plugged in.

Throttling the performance while running on battery is a design choice to improve battery life. The battery should be rarely a bottleneck.

[leadeater]

15 minutes ago, HenrySalayne said:

Do you have any sources for these claims? Modern batteries can supply insane currents

That depends on the battery, see all the problems with Android phones resetting even with large mAh batteries and the iPhone performance capping scandal. In general batteries can supply a large amount of current but the issue is voltage drop, there are different ways to design a battery and the ones for the mobile and laptop market are not the designs that can handle high current loads all that well. It's a necessity to get those larger mAh ratings.

 

The two main category of designs are Energy Cell and Power Cell.

 

Quote

Li-ion Energy Cell
The Li-ion Energy Cell is made for maximum capacity to provide long runtimes. The Panasonic NCR18650B Energy Cell (Figure 1) has high capacity but is less enduring when discharged at 2C. At the discharge cutoff of 3.0V/cell, the 2C discharge produces only about 2.3Ah rather than the specified 3.2Ah. This cell is ideal for portable computing and similar light duties.

 

Quote

Li-ion Power Cell
The Panasonic UR18650RX Power Cell (Figure 2) has a moderate capacity but excellent load capabilities. A 10A (5C) discharge has minimal capacity loss at the 3.0V cutoff voltage. This cell works well for applications requiring heavy load current, such as power tools.

 

If laptops came with Power Cell batteries then current/wattage limiting wouldn't be a necessity but you would also have a lot lower run time on battery for lighter loads. Since the most typical is lighter loads the more suitable type is Energy Cell.

[TheReal1980]

34 minutes ago, HenrySalayne said:

Do you have any sources for these claims? Modern batteries can supply insane currents. Most power supplies, on the other hand, are underpowered and the battery is used for current peaks. I had several notebook draining the battery under high load while plugged in.

Throttling the performance while running on battery is a design choice to improve battery life. The battery should be rarely a bottleneck.

Yeah I mean there are LOADS of more powerful laptops on the market that CAN handle the higher wattage needed so if this is a development choice based on having "more life" in the computer then they need to show this when sold.

If I buy a laptop and it has 50% less performance when it runs on just battery.... Let's just say I'll be sending it back. The Intel version of the Surface had no where near that loss of performance.

[Franck]

I'm just curious as of why is his web browsing benchmark is even there. He's running probably Opera vs chrome. That's not even fair benchmark

[TheReal1980]

5 minutes ago, Franck said:

I'm just curious as of why is his web browsing benchmark is even there. He's running probably Opera vs chrome. That's not even fair benchmark

I am more curious why he doesn't show for example how many times you can render something on one full battery charge? Linus should do this in his laptop reviews so you understand better what it is capable of on one charge.

[WolframaticAlpha]

18 hours ago, TheReal1980 said:

Summary

In a new video by Max Tech, three different laptops are tested against each other.

Apple Macbook Air (M1)

Microsoft Surface Laptop 4 (Intel)

Microsoft Surface Laptop 4 (AMD)

 

Here are some results.

 

 

 

 

 

Quotes

 

My thoughts

I knew the M1 would be fast but not so much faster compared to the Surface. The BIG problem I saw is that running the AMD (and somewhat Intel) on just battery will kill the performance. Intel also lost a little in performance but not nearly as much as AMD and it ran faster than the AMD in multiple tests.

M1 is no joke, we all knew Intel was dripping in their pants but I think AMD just got a little moist as well.

 

Sources

 

 

While the M1 is good, Max Tech is a is a complete apple shill. He literally said that a PC which was a third the price of the Mac Pro, and was 60% faster was overall worse because it made 13 dB more noise. What a fucking joke.

[maartendc]

4 hours ago, Spindel said:

I've naged about this in a lot of threads about the M1. 

 

As an owner of an M1 Mini one thing no benchmark can't communicate and most (if not all reviews) do not communicate is the over all responsiveness of the system when using it. I understand its omitted because it is a thing that is hard to quantify, but no computer I have ever used before has the same feeling of responsiveness when actually using the computer. 

Interesting...

 

But really the speed at which apps open is more dependent on storage speed than CPU speed, isn't it? If you are comparing an NVME SSD equipped Intel Macbook vs an NVME SSD equipped M1 macbook, is the difference noticeable?

 

I own a 2019 Macbook Pro with Core i9 (9880H) and it has a blisteringly fast NVME SSD. So I wonder if I would notice much of an improvement, I wouldn't think so? Storage speed should be the same, and the i9 CPU can at least turbo up to very high clock speed momentarily while opening an app.

[Uji Ninja]

4 hours ago, Spindel said:

I've naged about this in a lot of threads about the M1. 

 

As an owner of an M1 Mini one thing no benchmark can't communicate and most (if not all reviews) do not communicate is the over all responsiveness of the system when using it. I understand its omitted because it is a thing that is hard to quantify, but no computer I have ever used before has the same feeling of responsiveness when actually using the computer. 

They are nice machines, not denying that.

But there are quite a few things (such as using Linux, using old OS/software through Virtual Machines...) that I would not be able to do. And don't get me started on the headache and cold sweat inducing task of checking if my music hardware and software would work.

[HenrySalayne]

44 minutes ago, leadeater said:

If laptops came with Power Cell batteries then current/wattage limiting wouldn't be a necessity but you would also have a lot lower run time on battery for lighter loads. Since the most typical is lighter loads the more suitable type is Energy Cell.

The behaviour while reaching the lower limit is not an excuse to throttle something across the entire capacity range. And larger batteries of the same type are actually more capable than lower capacity ones. I don't mind some throttling and power saving routines when the battery drops below 20%, but there is no technical reason to do these things all the time. It's a design choice, not an intrinsic battery characteristic.