Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

ericinboston

macrumors 68020
Original poster
Jan 13, 2008
2,060
569
Hi all. I have several Mac Mini M1 and M2 baseline models and they run 24x7 a distributed computer project called World Community Grid.

These machines only use the CPU cores to crunch data...no GPU use at all. The machines run whisper quiet and very cool.

Does anyone here know if the new Mini M4s (base model) will run just as quiet and cool with these 24x7 CPU-only tasks? I've seen some general topics and videos stating the new Mini M4s in fact get very hot under load as well as the fan starts screaming. However, these articles I've read don't specify if "under load" in their tests is utilizing CPU, GPU, or both.

Any help would be great. Maybe you know of a better article or video I can watch.
 
Last edited:
Cinebench R23 loads only the processor. After three minutes, 5000 rpm and noise that can be heard in the next room. I do not recommend the mini M4 to anyone, it is better to buy an M2 with 16 GB in a classic case.
1.png
 
Thanks for your help, Dimvol! It's a shame that it's so loud (and likely hot) as other tests have shown.
 
I can only speak about M4 Pro...

If the task puts a high load on more than half of the CPU cores, you will probably have audible fan noise.

The M1, M2, and M3 are still very capable processors. Nonetheless, for example, M4 (10-core, non-Pro) has about the same multicore performance as a 12-core M2 Pro. And the M4 Pro surrounds the M3 Max multicore CPU performance.

As for the heat, it’s only likely to cause any kind of performance hinderance in an environment of >26ºC.
 
  • Like
Reactions: G5isAlive
...The M1, M2, and M3 are still very capable processors. Nonetheless, for example, M4 (10-core, non-Pro) has about the same multicore performance as a 12-core M2 Pro. And the M4 Pro surrounds the M3 Max multicore CPU performance...

Yes, my M2 machines are FANTASTIC for this distributed computing project. I don't think it's worth it for the Mini M4 given the fan noise, extra heat, and likely no more than 10% more CPU power for this project. I can grab a bunch of Mini M2s refurbished from Apple for $419 or $350 on ebay.
 
I have the M4 Pro 12/16 48GB Mini and the temperature/noise issue is blown out of proportion. It's silent probably about 90% of the time, even when I'm editing videos in FCP or DaVinci Resolve. For tasks where the CPU/GPU gets maxed out, the temps hit 100C but it's rare for that to happen. For me it only happens when I'm generating image in Stable Diffusion or exporting heavy After Effects compositions. By default, the fan doesn't go above 1000RPM until the temps hit 100C so it's pretty much always silent. I have a custom fan curve set in TG Pro to always keep the temps below 90C with the tradeoff of hearing low fan noise at around 70C. Overall I'm really happy with the machine.
 
  • Like
Reactions: G5isAlive and EugW
Wow, I guess I'm the only on keeping temps down, 50C is alarming to me.
I use Mac Fan Control app manually, I don't like thermal saturation, I like to stop it before happening.
I started this after seeing temps in the 90C ranges when playing with local LLMs.
 
Wow, I guess I'm the only on keeping temps down, 50C is alarming to me.
I use Mac Fan Control app manually, I don't like thermal saturation, I like to stop it before happening.
I started this after seeing temps in the 90C ranges when playing with local LLMs.

Apple Silicon was designed to handle those type of temperatures. Given that roughly 105-107C seems to be the max temp for M4, even a 90c reading is nothing to be overly concerned about. 50C is low enough that it's really not even worth mentioning unless you're in a competition for the lowest operating temperatures.


Cinebench R23 loads only the processor. After three minutes, 5000 rpm and noise that can be heard in the next room. I do not recommend the mini M4 to anyone, it is better to buy an M2 with 16 GB in a classic case.
View attachment 2479904

Cinebench is a poor test for thermal performance because it is designed to push the CPU cores to their limit, and is unrepresentative of the vast majority of real-world usage. Even heavy gaming does not push the SoC to the same limits as Cinebench.
 
This is a good test in this context and it loads only the perfomance cores. Games load the GPU, and it heats up even more. So, there will be the same 105 degrees. Throttling. Here is the dependence of the results on cooling.
4.png
 
Would anyone here with a Mini M4 be willing to load World Community Grid on their machine? Just set it to run the Mapping Cancer Markers project, set it it to use 100% of CPUs and set it to run 100% of the time. You should know in about 5 minutes if your machine is loud and/or getting hot.
 
Would anyone here with a Mini M4 be willing to load World Community Grid on their machine? Just set it to run the Mapping Cancer Markers project, set it it to use 100% of CPUs and set it to run 100% of the time. You should know in about 5 minutes if your machine is loud and/or getting hot.
anything using 100% of the CPUs 100% of the time is going to get hot and have fan noise. No getting around that.
 
I did a re-test with CB2024, both the 10 minutes per test and 30 minutes per test benchmark suites consecutively. Room temp was ~71ºF/21.7ºC.

Mac-mini_M4-Pro-14-20_Cinebench-2024_10min-and-30min-tests.jpg

The first test is the 10-minute GPU. At the 15-minute mark, the benchmark transitions to the multi-core CPU test. The 10-minute single-core test occurs between 26 and 38 minutes into the run. The first of the 30-minute benchmarks, the GPU test, lasts until the 73-minute mark. The final test, 30-minute single-core CPU, begins at around 106 minutes in.

The case and exhaust became warm and very warm at times. The fan, as in previous tests, was audible only during the multicore benching. Nonetheless, as you can see, the more than two hours of continuous heat generation didn’t (notably) cripple performance.
 
anything using 100% of the CPUs 100% of the time is going to get hot and have fan noise. No getting around that.
Of course AND it varies by machine design. The point of this thread is to compare the Mini M4 vs the Mini M2 in regards to hotness and fan loudness. As I stated, the Mini M2 base running 100% CPU utilization 100% of the time gets warm and I do not hear any fan noise.

How does the Mini M4 base compare? That's my point.
 
Of course AND it varies by machine design. The point of this thread is to compare the Mini M4 vs the Mini M2 in regards to hotness and fan loudness. As I stated, the Mini M2 base running 100% CPU utilization 100% of the time gets warm and I do not hear any fan noise.

How does the Mini M4 base compare? That's my point.
The M4 is definitely not silent at 100% CPU use... but it's not loud either. A TV or room fan in the background easily drowns it out. Coming from a Windows desktop PC, I'll take a little fan noise over the jet engine PC I used before. The Mac is silent 90% of the time when I'm not pushing it.
 
My M4 basic gets hotter than satan's nuts when I'm playing in X-Plane. It's fine.

Uses about 500 watts less than the old PC did doing the same thing.
 
I have my second M4 Pro Mac mini--the first one the fans died, so Apple let me swap it out for a new one (with some pressure from me). Doing Zoom sessions 9 hours a day. Outside gets quite warm, but not hot. Activity Monitor Energy Usage shows that even during the Zoom sessions, with Ableton Live for audio processing, plus all the normal stuff (Safari, Firefox, messages, and 5 or 6 other things), energy usage is probably 10 to 20% (not sure what that means), CPU Usage shows about 85% idle. Fan noise is very low, barely noticeable, but present when running Zoom.
 
I have my second M4 Pro Mac mini--the first one the fans died, so Apple let me swap it out for a new one (with some pressure from me). Doing Zoom sessions 9 hours a day. Outside gets quite warm, but not hot. Activity Monitor Energy Usage shows that even during the Zoom sessions, with Ableton Live for audio processing, plus all the normal stuff (Safari, Firefox, messages, and 5 or 6 other things), energy usage is probably 10 to 20% (not sure what that means), CPU Usage shows about 85% idle. Fan noise is very low, barely noticeable, but present when running Zoom.
That's concerning that the fan died on your first mini so soon. Were you using a fan control app like TGPro? That's what I use to keep the temps down. I only run the fans at temps of 70C and up though not full blast all the time.
 
Last edited:
Yes, that is concerning, the fan dying on a new machine.
I use Mac Fan Control on my M4 Mac Mini Pro, I can only wonder if using the fan out of AUTO is concern for decreased life span?
 
Would anyone here with a Mini M4 be willing to load World Community Grid on their machine? Just set it to run the Mapping Cancer Markers project, set it it to use 100% of CPUs and set it to run 100% of the time. You should know in about 5 minutes if your machine is loud and/or getting hot.

So I dont have a base macmini 4, but a min 4 pro. I downloaded and tested WCG on it AND a MBP M3 Max 16 and on both set CPU to 100%. On both machines I was impressed at how quickly it ramped up to 100% CPU. On the MBP, the fan started almost instantly and held at 42 decibels. Sitting in front of it I could hear it. from across the room I couldn't. Interestingly I was testing during a teams meeting and that ran fine, no one complained. Similar with the MM 4 pro, but it took longer for the fans to ramp up, and maybe a tad louder at 44 decibels. But again, not an issue 10 feet away. My guess given the similarity of the two different Macs, that the mini 4 will be similar and that Apple controls things to an internal standard of theirs.

no temp data, neither machine got hot to the touch. maybe warm? Not enough for a cold day. I will play with it more later.

edit: after 5 hours of work, both machines have not gotten hot...both had the fans running but not overly loud, both crunched through 400 units of work. hope that helps
 
Last edited:
OP, I'm running both folding@home and World Community Grid simultaneously for a few weeks now, non-stop. It's just a quiet sound, really - comparable to my Windows laptop that is idling nearby.
 
Thanks @G5isAlive and @MandiMac! Would you be willing to share the results of your Mac Mini M4 stats from WCG? Simply go to the stats on the WCG webpage and click Devices to see your machine. Here's a shot of mine.

Screenshot (2845).png
 
Thanks @G5isAlive and @MandiMac! Would you be willing to share the results of your Mac Mini M4 stats from WCG? Simply go to the stats on the WCG webpage and click Devices to see your machine. Here's a shot of mine.

View attachment 2481455


sure but keep in mind I dont know what I am doing lol...I had two computers I was playing with, at least one of them was turned off last night... not even sure about the mini to be honest. and on what I found, not sure which is which. but if you want to send me a message walking me through this will be glad to get you better results
 

Attachments

  • stats?.jpg
    stats?.jpg
    59.1 KB · Views: 32
Apple Silicon was designed to handle those type of temperatures. Given that roughly 105-107C seems to be the max temp for M4, even a 90c reading is nothing to be overly concerned about. 50C is low enough that it's really not even worth mentioning unless you're in a competition for the lowest operating temperatures.




Cinebench is a poor test for thermal performance because it is designed to push the CPU cores to their limit, and is unrepresentative of the vast majority of real-world usage. Even heavy gaming does not push the SoC to the same limits as Cinebench.
105 °C = 221 °F

It means that you should be able to fry an egg directly on the M4 Mini's case?
~Just kidding with you :)
 
  • Haha
Reactions: MikePx96
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.