p.1 #1 · A tale of two batteries: Mac vs. PC laptops and photo editing
As a follow up to the Lightroom benchmarking post of a few days ago, I decided to do something I've never seen done/posted anywhere else--test how an Intel i7-12800H based laptop with Windows 11 compares to a new M2 Max MacBook Pro when run on battery. Details on each device:
Mac: 16" M2 Max MacBook Pro, 12-core processor, 96GB of RAM, 38 GPU cores
Based on synthetic benchmarks, the CPUs are pretty comparable, using GeekBench 6 the i7 gets a single core score of 2200 and a multicore score of 9500; it has 14 cores compared to the M2's 12 cores. The i7 runs at a base frequency of 2.4GHz and turbo boosts to 4.8GHz; the M2 runs at 3.68GHz but clocks down when not being pushed. The M2 gets an average single core score of 2600 and a multicore score of 14500, so a little bit higher but it's also a little bit newer.
The GPUs are a bit different; based on just OpenCL scores you'd think the Nvidia is a lot better, at 105,000 vs. about 85,000 but most modern apps are using Metal on the Mac platform or coding to the capabilities of the Nvidia cards better. The Metal score for the M2 Max GPU is in the 140,000 range.
In terms of battery they are similar in capacity--the ThinkPad as a 90W battery and the MacBook Pro has a slightly larger 100W battery.
So now for the test results. I decided to just compare the two intensive tasks that take time, as they would show battery drain best. Both laptops were started with a 100% charge for each task. Both were left to the OS to set the power mode when on battery; the PC went to "battery saver" mode when unplugged, and the Mac was allowed to auto-throttle if needed, but apparently didn't do so, as the results show.
First, preview rendering on 1465 Sony A1 files. As the graph shows, when plugged in the Mac is about 13% faster than the PC; not a big difference and probably the faster i9 CPU could close some of that gap, although theoretically the PC GPU should have closed some of the gap and didn't. The real differences show up on battery power: The Mac chugs along and performs EXACTLY the same on mains power or on battery. Meanwhile the PC has to throttle back so it's 9% slower on battery than mains power, and now it's 23% slower than the Mac on battery power. More amazing is how much of the battery was consumed during the process. The PC used up a whopping 51% of its battery capacity while the Mac used only 10%.
The difference was even more stark on the JPEG exports which hit the GPU even harder than preview rendering. In this case the PC already starts out slower than the Mac, even with what's supposed to be more a powerful GPU--it takes 85% longer to render all of the previews when plugged in. On battery, it takes almost 3x as long to render the previews as the Mac, whereas once again the Mac does the job in the same amount of times as when it's plugged in. The battery situation is remarkable--the Lenovo used up a 87% of its battery capacity to complete the task and gave me warnings about being needed to be plugged in. The Mac managed to use only 20% of its battery capacity.
p.1 #3 · A tale of two batteries: Mac vs. PC laptops and photo editing
armd wrote:
Super, real world comparison and you're fortunate to be able to afford such nice (envy) systems. Thanks for sharing!
They're work expenses, so that helps; I can't play with them all of the time so before I give them to the employee(s) using them I put them through some paces as I figure it will help some people out! Glad you found it useful.
p.1 #4 · A tale of two batteries: Mac vs. PC laptops and photo editing
And this is the crux of it, right here. I hear people use the argument “Apple can’t break the laws of physics”, which is true, but this test shows that they have managed to find some pretty eye-popping efficiency gains with these processors, without sacrificing performance.
C
p.1 #5 · A tale of two batteries: Mac vs. PC laptops and photo editing
Chris Court wrote:
And this is the crux of it, right here. I hear people use the argument “Apple can’t break the laws of physics”, which is true, but this test shows that they have managed to find some pretty eye-popping efficiency gains with these processors, without sacrificing performance.
C
Kinda reminds me of how we got more "energy efficient" when we went from T40 to T32 fluorescent bulbs. Really, only less of same technology = less power = less output (lumens), so it was "smoke and mirrors" efficiency gains.
Then, enter LED lighting, and it was an different way of doing things, and had a real efficiency gain.
I'm a big fan of physics and laws of energy conservation (i.e. neither created, nor destroyed ... only changes form) ... so, I think the fact that Apple is also showing a significant heat signature reduction (i.e. less wasted energy), is indicative of their harnessing the energy they are working with well, beyond traditional (waste) heat dissipation efforts.
Suffice to say, it's been enough of a difference to catch my attention way more than T32 bulbs did.
p.1 #6 · A tale of two batteries: Mac vs. PC laptops and photo editing
RustyBug wrote:
Kinda reminds me of how we got more "energy efficient" when we went from T40 to T32 fluorescent bulbs. Really, only less of same technology = less power = less output (lumens), so it was "smoke and mirrors" efficiency gains.
Then, enter LED lighting, and it was an different way of doing things, and had a real efficiency gain.
I'm a big fan of physics and laws of energy conservation (i.e. neither created, nor destroyed ... only changes form) ... so, I think the fact that Apple is also showing a significant heat signature reduction (i.e. less wasted energy), is indicative of their harnessing the energy they are working with well, beyond traditional (waste) heat dissipation efforts.
Suffice to say, it's been enough of a difference to catch my attention way more than T32 bulbs did. ...Show more →
While people love to cite "physics" it's not really the barrier here. Like you said, it's a new way of doing things. The whole x86 architecture is extremely long in the tooth. Its gone through a few reiterations along the way that have really changed it--first the integration of some RISC-like superscalar architectural changes in the Pentium era (which I should point out Intel swiped from Digital), then the rise of the multi-core world in the Pentium Core Duo and following, but the reality is that it was never designed as a power-efficient architecture. Intel actually had other efforts in that arena, but bet incorrectly and sold it off right before the whole smartphone revolution took place.
The massive explosion of the smartphone market has resulted in massive investments in R&D in processors that by necessity are designed from the ground up to be extremely power-efficient. In some ways, in hindsight, it shouldn't be surprising to realize these processors could be slightly amped up to run full-blown desktop systems, and the complexity and power of the handhelds has pretty much converged with desktops in the 15 years since the launch of the iPhone.
Apple has the added advantage of developing the entire stack from hardware to software, so performance and efficiency can be highly optimized. Even Microsoft realized the advantage of that approach and has been building their own hardware using ARM chips and highly optimized versions of Windows and Office for that platform.
IMO we are entering a different era in computing. In the past it was all about very rapid changes in generic hardware and software that leveraged that. In recent years that has stagnated and computers have taken a back seat to handheld devices. I don't think we'll continue to see as much of a separation between desktop and handheld hardware and software as time moves forward. Today's "killer apps" aren't desktop applications that require tons of power like in the past, they're web-based and have to be able to run well on handhelds and not just a desktop computer. In some ways that explains why we've seen some stagnation on the desktop computer hardware side, outside of the increasing power of GPU's, but that's driven more by their use for a high-spending demographic--gamers--and their use in things like machine learning and other specialized analytics.
Apple has shown a different future where the investments in power-efficient hardware for mobile handhelds can be re-imagined for desktop/laptop use without sacrificing performance, and while gaining huge power efficiency. I have no doubt that's also why Microsoft has invested so much in building ARM versions of Windows and their apps--they seem a similar future. What this means for Intel/AMD is another question--AMD seems to have an edge in their Ryzen and chiplet architecture right now, but it's still desktop/laptop limited. Intel and AMD will react, and have already started to do so. What the future holds is open at the moment.
p.1 #8 · A tale of two batteries: Mac vs. PC laptops and photo editing
Cnyphotoguy wrote:
Windows by default manages peak CPU capabilities while on battery, did you look to verify that both machines where similarly capable in your testing?
Yes, if you read what I wrote in detail, I did, and Windows was--believe or not--set to be on "battery saver" mode when the test was being run. Windows users seem to struggle to believe this is all possible, but remember, the M1/M2 chips are basically just the same chips that run ALL DAY in an iPhone. Ramp them up to run a laptop, and use a large laptop battery, and the results are not surprising at all. The x86 architecture has never been designed to be power-efficient, vs. the new Apple silicon, which has its roots in highly power-efficient mobile chip design.
p.1 #9 · A tale of two batteries: Mac vs. PC laptops and photo editing
jhapeman wrote:
Yes, if you read what I wrote in detail, I did, and Windows was--believe or not--set to be on "battery saver" mode when the test was being run. Windows users seem to struggle to believe this is all possible, but remember, the M1/M2 chips are basically just the same chips that run ALL DAY in an iPhone. Ramp them up to run a laptop, and use a large laptop battery, and the results are not surprising at all. The x86 architecture has never been designed to be power-efficient, vs. the new Apple silicon, which has its roots in highly power-efficient mobile chip design. ...Show more →
I don't struggle with anything other than curiosity as to if this is really a valid comparison.
jhapeman wrote:
Both were left to the OS to set the power mode when on battery; the PC went to "battery saver" mode when unplugged, and the Mac was allowed to auto-throttle if needed, but apparently didn't do so, as the results show.
I did miss this, how did you verify that the Apple didn't use the auto throttle config to use full hardware capabilities? knowing that one OS decided to limit hardware capability hardly makes for much of a comparison.
p.1 #10 · A tale of two batteries: Mac vs. PC laptops and photo editing
Cnyphotoguy wrote:
I don't struggle with anything other than curiosity as to if this is really a valid comparison.
I did miss this, how did you verify that the Apple didn't use the auto throttle config to use full hardware capabilities? knowing that one OS decided to limit hardware capability hardly makes for much of a comparison.
The Mac is set to the exact same thing as the PC: Throttle the power if needed to save the battery. As the results clearly show, that was not necessary--the battery was just fine and the CPU was able to run at full power. That's the whole point, isn't it? Not only is the PC forced to throttle and get slower results, even when it is throttled the battery life is absolutely abysmal. I'm sure I could *force* the Mac into a power savings mode that would have conserved the battery even more than it already did, but that's pointless.
The basic point remains: The PC has to dramatically slow performance and yet still kills the battery. The Mac doesn't have to slow performance at all and basically just sips on the battery, using very little power. That's an amazing difference, and honestly, if you are buying a laptop to use as a laptop--that is to say, not plugged in 100% of the time--the Apple Silicon MacBooks perform like nothing else out there.
p.1 #11 · A tale of two batteries: Mac vs. PC laptops and photo editing
My lap test ... Better performance, cooler temps, battery all evening.
The only way that all three can improve together is if they have a common denominator ... efficiency.
Others utilize tradeoffs power for heat, power for battery, etc.
It is a paradigm shift in approach, and it shows, no matter how much you slice and dice it ... the results keep revealing the same thing, no matter if it is benchmark or real world, fair or unfair ... the answer keeps being the same.
p.1 #12 · A tale of two batteries: Mac vs. PC laptops and photo editing
jhapeman wrote:
The Mac is set to the exact same thing as the PC: Throttle the power if needed to save the battery.
Thats not how Windows works, in battery saver mode as a user, your software is completely unable to achieve full hardware performance. I appreciate your testing, but don't find it that telling of anything specific not already known to varying degrees.
p.1 #13 · A tale of two batteries: Mac vs. PC laptops and photo editing
Cnyphotoguy wrote:
Thats not how Windows works, in battery saver mode as a user, your software is completely unable to achieve full hardware performance. I appreciate your testing, but don't find it that telling of anything specific not already known to varying degrees.
That's *exactly* how it works: Both Windows and the Mac will throttle back the CPU if needed when on battery, to conserve battery power. For some reason you don't seem to get this. The difference is that now, with the Apple Silicon, the Mac *doesn't need to* throttle back the performance, because it uses so little battery power even at full performance. As the battery gets low, it does then throttle back to extend the life, but that was never the case here, as it used so little of the battery to complete these tasks.
If I had forced the PC to stay at full performance, we already know the stats on how it would do (go look at my other thread, it still loses out to the Apple M1/M2), but it wouldn't have been able complete either task without completely exhausting it's battery.
It most certainly shows something I've never seen anyone else publish--all you ever see are inane battery tests using a web browser or streaming a video. Those are not wholly representative of any kind of day to day use scenarios, particularly for what a photography user here might experience. I use Lightroom on the road, and often a plug isn't available. In the past with my Intel-based MacBook Pros, it was a problem, they just devoured the battery. The point is that's completely not the case anymore. The level of performance one can now have on battery power is unprecedented. It's not web browsing, its very intensive tasks like preview rendering, exports, etc.
when we went from T40 to T32 fluorescent bulbs. Really, only less of same technology = less power = less output (lumens), so it was "smoke and mirrors" efficiency gains.