Compilation of real-world tests on M1 Macs

[Loa]

Hello,

I have two friends who came to me for buying advice on M1 Macs. And while the internet is filled with theorycrafters (has more cores, has more thermal limit, will probably throttle, etc...), there is a definite dearth of real-world tests.

So I'd like this thread to become a list of real world tests. No theory, no hypothesis, no "I think that", no "it should" and no "well 7/8 implies 12.5% gain in performance"...

Just good testing.

Areas that seem to me to be the most interesting:
-----The differences in performance of 7 or 8 GPU cores in various games;
-----The differences in performance of M1 Macs Vs other computers with discrete GPUs;
-----The differences in performance of 7 or 8 GPU cores in non gaming apps and situations;
-----Actual performance effect of the thermal throttling on the Air, in which situations, after how much time;
-----Performance difference of ram (8GB Vs 16GB) on memory intenseive tasks;

If you have done, or have found, such real world tests, please post them here. If this works, this thread could become a useful resource for all potential mac buyers.

Thanks.

 

[Wotcher]

Strictly my personal opinion.

If someone came to me asking these kinds of questions to decide whether they should buy an M1 Mac, I would just tell them to stick with what they have for now.

For better or for worse, Youtube has a plethora of benchmark testing between M1 7-core, 8-core, and Intel Macs. But these are synthetic type benchmarks, and "real-world" testing of when it might thermal throttle depends on actual use case, which depends on the user's individual situation.

One can split hairs a trillion different way and still someone will find something wrong, or yet another way to split the trillionth piece. It's just me, but these kinds of tests feels like beating a dead horse to its millionth death.

I have both Intel mini (i7/64GB) and M1 MBP (16GB) and the M1 is waaay faster for my "everyday tasks". No idea whether it's 12.7% or 82.8% faster or slower, I just know tasks get done faster (or feels that way).

I mean, and no disrespect to OP, do people split hairs like this? I guess they do...

 

[Loa]

Look, it's very simple. A lot of people that need computers are not knowledgeable about technical details about the computers. Also, a lot of people care about how much their computer costs and wether the extras are worth the price.

The intersection of these two groups makes up a very sizeable proportion of people who buy macs.

Nobody is splitting hair. Or, if we are, we'd like to know exactly why, not just believe the opinion of people who can't back their claims with real-world tests. You sentence: "No idea whether it's 12.7% or 82.8% faster or slower, I just know tasks get done faster (or feels that way)." is exactly what most of us don't need.

That's why I'd like to have people who did tests, on a variety of apps, to see if the upgrades are worth their price. It's not that complicated.

 

[The_Interloper]

-----The differences in performance of M1 Macs Vs other computers with discrete GPUs

I have an M1 Mac Mini (16Gb), Macbook Pro (8Gb) and recently sold my 2018 i5 6-core Mac Mini with Intel iGPU (32Gb). I also have a PC with Ryzen 7 1700 8-core CPU, 32Gb RAM and Radeon 480 GPU.

Rendering a 43 minute 1080p video on each:

M1 Mac Mini - DaVinci Resolve: 6 min 20 secs
i5 Mac Mini - DaVinci Resolve: 12 min 33 secs

M1 Mac Mini - Final Cut Pro X: 3 min 51 secs
i5 Mac Mini - Final Cut Pro X: 8 min 18 secs

Ryzen 7 - Davinci Resolve (software render): 22 mins 19 secs
Ryzen 7 (OpenGL accelerated) - ShotCut: 9 min 22 secs

The M1 is faster in every discernible way from fairly decent PC setups (though not with the likes of RTX 3090s etc) and annihilates the 2018 i5 Mac Mini in my experience. I also had an eGPU (with the Radeon 480 in) and found little improvement in this area with the i5. For rendering it was actually slower than the iGPU (due to Intel QuickSync).

 

[Wotcher]

So when the tests tell me on the 78th second it thermal throttles vs on the 82nd second it throttles, what does that mean? Is it clear that I should buy the one that throttles on 78? Do I stop using it before I hit 78 seconds or 82 seconds and let it cool off? Do I even use the same app in the same way as that "real world benchmark"?

Or when the 7-core gets me 55.9 fps vs 8-core gets me 62.6 fps, what does that mean? And these types are averages so you know sometimes FPS dips below and sometimes it's pretty high.

My set of apps and the way I use them are different from yours and different from everyone else, even though some (or most) are common.

The combinations to test--of apps/games (plus which versions), which Mac and configurations--would be pretty huge, no? How may tabs on which browser? 55 on Safari? 98 on Chrome? Or both at the same time? Which sites? What am I doing while I have those browsers/tabs/sites up? final Cut? iMovie, Lightroom (CC or Classic)? Photoshop (which version? plugins)? Facetime? Zoom? Teams? MS Office (Word? Excel? PP? what kind of files do I have open and using)? what kind of wifi/ethernet connection do I have?

And that would mean, as soon as someone tests in a particular set of combination, there would come along someone else saying, "but what about...?", no? When I search Youtube for M1 tests, I get page after page of tests and reviews.

Because "real world", people don't think about fps or exactly when my 7-core Air throttles. You can't tell people in your zoom meeting, "I gotta go, in 5 seconds my Air is going to throttle to 55 fps from 70 fps... dang, I should have bought a Pro!"

I mean, if someone feels they need to know whether it's 82 seconds or 105 seconds to complete the task/before throttle/battery runout/screen dims (and I'm curious, don't get me wrong), that's an exercise of curiosity.

But it's not a buying decision--it's much more nuanced.

 

[TheSynchronizer]

I had an M1 Air during the 14 day return period and it got noticeably hotter to the touch when used on a lap compared to my pro I own currently. That being said, it was never uncomfortably hot or an issue at all, but it’s a genuine difference i’ve noticed and my pro always runs at the same temperature as my clothing when on my lap therefore the heat never being noticeable at all.

I’ve found the screen of the pro also noticeably more enjoyable to use outside in my garden, which means I now often work outside and it’s something I wouldn’t have done often having kept the air due to the inconvenience when the air’s screen looks so much better inside.

Both have performed pretty similarly, and the 7 core of the air was giving me an average of 53 FPS compared to the pros average of 61 in Samurai Jack on apple arcade. Going back to the thermal aspect the pro definitely never throttles, e.g. when playing graphically intensive apple arcade games such as Samurai Jack for a long time hooked up to a 4k TV, the air‘s frame rate would get noticeably lower as time went on (20 minutes +). Not unplayably low dropping of frames but it was a genuine noticeable drop, to the areas of 43-46 FPS. The pro on the other hand sticks with the same average frame rate no matter how long I play on it. Likewise I wouldn’t want to do any intensive tasks like rendering or gaming on my lap with the air, whereas with the pro it’s not an issue.

 

[Wotcher]

So when the tests tell me on the 78th second it thermal throttles vs on the 82nd second it throttles, what does that mean? Is it clear that I should buy the one that throttles on 78? Do I stop using it before I hit 78 seconds or 82 seconds and let it cool off? Do I even use the same app in the same way as that "real world benchmark"?

Or when the 7-core gets me 55.9 fps vs 8-core gets me 62.6 fps, what does that mean? And these types are averages so you know sometimes FPS dips below and sometimes it's pretty high.

My set of apps and the way I use them are different from yours and different from everyone else, even though some (or most) are common.

The combinations to test--of apps/games (plus which versions), which Mac and configurations--would be pretty huge, no? How may tabs on which browser? 55 on Safari? 98 on Chrome? Or both at the same time? Which sites? What am I doing while I have those browsers/tabs/sites up? final Cut? iMovie, Lightroom (CC or Classic)? Photoshop (which version? plugins)? Facetime? Zoom? Teams? MS Office (Word? Excel? PP? what kind of files do I have open and using)? what kind of wifi/ethernet connection do I have?

And that would mean, as soon as someone tests in a particular set of combination, there would come along someone else saying, "but what about...?", no? When I search Youtube for M1 tests, I get page after page of tests and reviews.

Because "real world", people don't think about fps or exactly when my 7-core Air throttles. You can't tell people in your zoom meeting, "I gotta go, in 5 seconds my Air is going to throttle to 55 fps from 70 fps... dang, I should have bought a Pro!"

I mean, if someone feels they need to know whether it's 82 seconds or 105 seconds to complete the task/before throttle/battery runout/screen dims (and I'm curious, don't get me wrong), that's an exercise of curiosity.

But it's not a buying decision--it's much more nuanced.

Replying to myself because apparently I didn't accept all the required cookies to edit:

But it's not (edit to add) [solely for] a buying decision--it's much more nuanced.

Edit for better clarity:
Not saying these benchmarks are useless. I'm genuinely curious.

I've also had people ask me for buying advice, and benchmarks are the last thing on their minds--it's dependent on their use case. Can I run XYZ app? I'm going to school to major in ABC, and the school requires DEF apps and I want to do QRS and maybe play a little game or two.

They need to get work done and have a set budget. In most cases, they're moving from 3-5 year old laptop (of varying specs). Most certainly a 7-core air will get the job done for a Political Science major writing 150-page paper and researching sources w/ Zoom or Teams call with their classmates. 59fps? 89 fps? Throttle? Shrug?

 

[TheSynchronizer]

Replying to myself because apparently I didn't accept all the required cookies to edit:

But it's not (edit to add) [solely for] a buying decision--it's much more nuanced.

Edit for better clarity:
Not saying these benchmarks are useless. I'm genuinely curious.

I've also had people ask me for buying advice, and benchmarks are the last thing on their minds--it's dependent on their use case. Can I run XYZ app? I'm going to school to major in ABC, and the school requires DEF apps and I want to do QRS and maybe play a little game or two.

They need to get work done and have a set budget. In most cases, they're moving from 3-5 year old laptop (of varying specs). Most certainly a 7-core air will get the job done for a Political Science major writing 150-page paper and researching sources w/ Zoom or Teams call with their classmates. 59fps? 89 fps? Throttle? Shrug?

Ofcourse the 7-core air will get the job done. We’re in a time presently where a very large amount of our devices are more than able to get the job done, at much lower prices than the air. I could literally find the cheapest barebones chromebook from a few years ago, install zoom/teams on that and be writing away at political science papers while doing all the research i’d ever need in the browser, all of this at the price of 1/6 a macbook air. Hell, even a raspberry pi for $35 with linux and a headset can do all of this, so long as you have any display to hook it up to. Likewise could do the same on my ipad or even my phone.

Going up from these baselines you’re always just going to be paying for extra features if your whole work consists of just writing papers, researching and conference calling. The pro compared to the air simply has extra features for the higher price. And the majority of these features are literally improvements in the performance aspects of the laptop, i.e the battery lasts longer, the performance is better, the screen is brighter. And these are all things to measure with benchmarks.