Do You Think M1 iMac will Outperform M1 Mac mini/MacBooks?

[imdog]

I haven’t seen any discussion about this on forums or from YouTubers, but I’m wondering if the M1 iMac will have better performance than the M1 Mac mini, MacBook Air & MacBook Pros? As this is the first Mac that was built from the ground up for Apple Silicon, perhaps the thermals & such are much better? We do see different performances from each of the previous M1 Macs (mini > MBP > MBA). What are you guys expecting?

 

[Abazigal]

I haven’t seen any discussion about this on forums or from YouTubers, but I’m wondering if the M1 iMac will have better performance than the M1 Mac mini, MacBook Air & MacBook Pros? As this is the first Mac that was built from the ground up for Apple Silicon, perhaps the thermals & such are much better? We do see different performances from each of the previous M1 Macs (mini > MBP > MBA). What are you guys expecting?

We may see slightly better sustained performance for more intensive tasks like exporting video, but I expect them to perform largely the same for most part.

I think the whole idea behind putting the same M1 chip is all these Macs is so that the user doesn't feel like he is making a tradeoff by opting for one form factor over the other. Since they will all perform largely the same, he is free to choose the form factor which best suits his needs, knowing that he is going to get the same performance regardless.

 

[Juicy Box]

I think it will be on par with the M1 Mac Mini.

I agree with @Abazigal that there probably won't be any difference, with maybe the exception for long tasks.

My M1 Mac Mini does a really good job staying cool while performing really well during video encoding, unless it is something taking more than an hour, then it starts to throttle a bit.

My Intel Macs on the other hand, the fans start blasting off within 10 seconds of encodes starting and about one minute into encoding, the performance drops significantly.

What are you guys expecting?

It is too soon to tell, but imo, if you want the best performing M1 Mac and don't want to wait for iMac reviews and tests, go with the M1 Mac Mini.

You can use it with a monitor with a lower native resolution than the M1 iMac, which I would guess would lead to better performance due to the iMac having to push that 4.5k resolution.

 

[Lemon Olive]

No. It should be exactly the same as the Mac mini.

 

[Sydde]

Curiously, when you go to Apple and look at tech specs, you see that Intel iMacs and Minis run at 3~3.8Ghz with "turbo boost" to 4.5. AS versions, by contrast, do not mention clock speed at all. It seems unlikely that the iMac would have significantly different performance characteristics compared to the Mini.

 

[eneisch]

It will most likely perform better then the M1 Air as that only has passive cooling (no fans). I would expect similar performance to the M1 mini, but we won't really know until it is tested.

 

[iPadified]

They did say "faster than any Mac" or similar in the keynote, did they not? That means the M1 in current Macs as well. Looking forward to see some comparison. What are the first benchmarks released? At first shipment or before?

 

[imdog]

They did say "faster than any Mac" or similar in the keynote, did they not? That means the M1 in current Macs as well. Looking forward to see some comparison. What are the first benchmarks released? At first shipment or before?

They usually leak early, but I believe they can be faked. I remember when the original M1 Macs benchmarks started leaking a lot of people thought they were fake because of how powerful they were

 

[pmiles]

They didn't change anything in the M1 chip. They all share the same SoC configurations. Technically, they should all perform exactly alike because the only difference between them is the case they are housed in.

 

[cltd]

It is too soon to tell, but imo, if you want the best performing M1 Mac and don't want to wait for iMac reviews and tests, go with the M1 Mac Mini.

You can use it with a monitor with a lower native resolution than the M1 iMac, which I would guess would lead to better performance due to the iMac having to push that 4.5k resolution.

This. Remember iPad 2 vs iPad 3 (Retina) performance ?

 

[imdog]

They didn't change anything in the M1 chip. They all share the same SoC configurations. Technically, they should all perform exactly alike because the only difference between them is the case they are housed in.

The Intel MacBook Pro 13":

The M1 MacBook Pro 13":

The fan is literally the exact same:

Cases can have a fairly significant effect on performance & thermals. Completely redesigned fans, placed in a completely different part of the casing that has access to much better airflow (right by the bottom air vents of the computer, compared to right behind the screen in the center of the computer) can as well. You'll find some tests where a different casing with better airflow can drop temperatures by over 20 degrees.

Intel 21.5" 4K iMac:

M1 24" 4.5K iMac:

I

 

[Maconplasma]

I thought at least one of the YouTubers would've gotten a pre-release copy from Apple for review.

 

[iPadified]

A mild overclocking of the M1 seems possible. It seems logical to me that the a bigger machine can have a slightly higher clock rate because the cooling is better. Easy way to improve performance. Likely, the iPad Pro can get a slightly slower M1.

 

[uller6]

I have an M1 air 16/1T and an M1 iMac 16/1T. Here are some Cinebench results I got from playing around this past weekend.

Note that according to powermetrics (go to terminal and type "sudo powermetrics") both M1 chips operate at the same speeds. The M1 air actually hits higher frequencies for longer (until the heat saturates), so I think the M1 in the Air is a higher binned chip (lower power operation). From watching powermetrics on both machines the iMac M1 typically uses ~500 mW more power for the high performance cores at the same operational frequency as the MacBook Air M1. Also according to powermetrics the GPU in the M1 iMac runs at 712 MHz, while the GPU in the M1 Air runs at 396 MHz. I ran the tests while also running Slack, messages, safari, terminal, activity monitor, excel, calendar, music, and mail in the background on both machines.

The tests were run in a room with 25C ambient temp. M1 air was tested while on battery power. I ran three sequential Cinebench R23 multicore tests and recorded the CPU temperature using System Monitor, the CPU speed using powermetrics, and the resultant Cinebench scores. I pressed go on the next test as soon as the previous test was finished. All three plots are attached below.

It's clear that the M1 air starts to throttle CPU speed as the chassis heats up (I did the thermal pad mod). The iMac fans go crazy during the test (7150 and 6650 RPM) but even while blasting at maximum speed they're amazingly quiet.

 

[iPadified]

The GPU has higher frequency than the air and yet the same metal score? Is that not the inconsistent? So this design cannot cool more than the M1 at full blast. I think this is strange as the new MBP will likely have the same thickness and yet will have a more powerful chip. The larger iMac will then be thicker.

 

[uller6]

The GPU has higher frequency than the air and yet the same metal score? Is that not the inconsistent? So this design cannot cool more than the M1 at full blast. I think this is strange as the new MBP will likely have the same thickness and yet will have a more powerful chip. The larger iMac will then be thicker.

I was running Cinebench CPU tests; the GPU wasn't involved so I'm assuming this is the baseline GPU core frequency. These scores are not metal scores.

The M1 iMac was the first Mac designed around the M1 chip, so yes, by definition it cannot cool more than the M1 at full blast. It is designed to cool the M1 at full blast, nothing more. Running flat out, my M1 iMac pulls 45 watts from the wall. And that's including the screen at maximum brightness. The new MBP will also be designed with a specific CPU power dissipation in mind.

 

[Sydde]

The GPU has higher frequency than the air and yet the same metal score? Is that not the inconsistent? So this design cannot cool more than the M1 at full blast.

The GPU is part of the M1. Integrated in one chip (i.e., SoC). The M1 is basically all the fan needs to cool.

 

[iPadified]

I was running Cinebench CPU tests; the GPU wasn't involved so I'm assuming this is the baseline GPU core frequency. These scores are not metal scores.

The M1 iMac was the first Mac designed around the M1 chip, so yes, by definition it cannot cool more than the M1 at full blast. It is designed to cool the M1 at full blast, nothing more. Running flat out, my M1 iMac pulls 45 watts from the wall. And that's including the screen at maximum brightness. The new MBP will also be designed with a specific CPU power dissipation in mind.

You reported GPU frequencies for iMac M1 to be 700+ and the M1 Air as just below 400. The metal benchmarks are the same and usually higher GPU frequencies means higher performance but we do not see this. There is the inconsistency.

 

[uller6]

You reported GPU frequencies for iMac M1 to be 700+ and the M1 Air as just below 400. The metal benchmarks are the same and usually higher GPU frequencies means higher performance but we do not see this. There is the inconsistency.

Right. I read the GPU frequencies directly from powermetrics, so the speeds should be correct. The discrepancy could be between the idle GPU frequencies for each binned M1 chip, since I was running CPU tests only? Maybe the M1 air idles far lower to help save heat & battery, while the iMac can idle at a faster frequency because of the improved power dissipation. When they're running a metal benchmark they may boost to the same "full load" frequency, which would give the same benchmark performance.

Edit: can you recommend a good metal benchmark for me to try?

 

[iPadified]

The GPU is part of the M1. Integrated in one chip (i.e., SoC). The M1 is basically all the fan needs to cool.

See above.

 

[iPadified]

Right. I read the GPU frequencies directly from powermetrics, so the speeds should be correct. The discrepancy could be between the idle GPU frequencies for each binned M1 chip, since I was running CPU tests only? Maybe the M1 air idles far lower to help save heat & battery, while the iMac can idle at a faster frequency because of the improved power dissipation. When they're running a metal benchmark they may boost to the same "full load" frequency, which would give the same benchmark performance.

Edit: can you recommend a good metal benchmark for me to try?

I am sorry but no. I only looked at Geekbench metal score. Idle frequency might be an explanation as frequency under load is >1200 MHz. Curious difference though.

 

[4sallypat]

Check out real world tests - note that M1 Air is not able to keep up with the M1 iMac.

The M1 Mini is still pretty much the reference M1:

 

[Juicy Box]

Here is Max Tech comparing the two M1 iMacs:

Basically, the base model runs hotter and throttles often due to the weak cooling. Pretty sad.

Check out real world tests - note that M1 Air is not able to keep up with the M1 iMac.

The M1 Mini is still pretty much the reference M1:

The video hasn't aged well considering it was made pre-teardowns and with the assumption that the cooling was the same on the two M1 iMac models.

Apple has really misled their customers when it came to the base model M1 iMac, and I can't believe this isn't getting more coverage.

The M1 Mini is still pretty much the reference M1

It looks like the Mac Mini is the best performing M1 Mac. The M1 Mac Mini is a beast, and a really great value.

I would like to see a more in-depth face off between the M1 Mac Mini and higher end 24" M1 iMac, maybe Max Tech will do one.