OP won't explain, OP won't elaborate, OP won't describe.
The process itself isn't user error, but the fact that he can't fix it is.
He has a pattern of doing this, please take that into context next time you try to lecture someone.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ARM MAC = new Windows Vista machines in 2025
- Thread starter Zest28
- Start date
- Sort by reaction score
The process itself isn't user error, but the fact that he can't fix it is.
He has a pattern of doing this, please take that into context next time you try to lecture someone.
Oh absolutely, they're fantastic machines! Meant it fully as a compliment to how capable they indeed are.
It takes so much processing power it needs the processing might of the A15 to run well!
I would be fuming if I would have been so dumb to spend 4000$ on a machine I hate afterwards. Respect!
Why worry? It is unimportant. A modern OS will always try to use all the RAM you have. Unused RAM is a waste. It is better to keep a process that runs periodically in RAM than to periodically restart it. It costs nothing. The OS has a priority-based scheduler and as long as the taks you are interacting with are not having to wait, you are fine, not just fine, but there is no room to improve.
Did you ask Apple to do that? Yes. You would be upset if you plugged in an external device and macOS failed to notice. But just do not worry, it does not take away from the task you are doing.
He uses it because it works well. He complains about stuff he does not understand.
Have fun spending more for that video card than the rest of the computer combined (possibly even more than the MacBook) and only getting roughly a 10% uplift in real performance over the 4090. But it's your money to burn...
But the 5090 has the side benefit of being able to cook dinner using all that extra heat. And further warm the planet as well.
My Mac has 8 GB of said memory, and 512 GB of flash store. I am yet to encounter a major annoyance.
P.S. I tend to dislike big Macs (pun intended), but since the OP's struggling so much with theirs, I can help them out and accept it as a gift. Postage is on me!
P.S. I tend to dislike big Macs (pun intended), but since the OP's struggling so much with theirs, I can help them out and accept it as a gift. Postage is on me!
I’ve had similar on my Studio, too - pushing my 128GB to its limits. My Mac didn’t “run out” of memory as such - I knew what I was doing to push it to its limits.
I’m just amazed a little box on my desk can do as much as it can do.
You're not using it right if you are not gobbling up all the RAM 
; or you have chosen to engage in workflows that exceed the limits of the machine, in which case you should have purchased more RAM.
The only thing better than more RAM, is even more RAM.
; or you have chosen to engage in workflows that exceed the limits of the machine, in which case you should have purchased more RAM.The only thing better than more RAM, is even more RAM.
On top of this you also need to look at bugs in the software too. At one point on my M2 Ultra 192GB RAM setup I was working on just compressing a 720p video from DaVinci Resolve. My entire computer became unresponsive where clicks literally took minutes to register and my memory pressure was red. I also saw swapping at the 10+ TB range during the export. Yes terabytes. In no way for 720 or 1080p or heck even 4K and 8k work do I need MORE than 192GB of RAM.
Turns out it was a massive bug and I filed a bug report then uninstalled Davinci. Went back to Final Cut Pro and Compressor. It barely took 2GB of RAM to process the same video.
So red doesn’t always mean YOU NEED MORE RAM!!
Last edited:
Not in all cases. There are some Windows applications that write to swap even if you have plenty of RAM. The biggest offender I saw on this was a few years ago Chrome was writing to cache when I only had 17GB out of 64GB RAM used.
Windows and macOS manage memory slightly differently, and it’s really complicated. Both operating systems have memory managers that were written by incredibly smart people to meet the needs of very broad and diverse user bases, and reflecting the design decisions of each OS kernel (WNT for Windows, XNU for macOS). Windows prefers to page to disk; macOS prefers to page to compressed memory.
I’m not an expert, but if I were to speculate, I’d say the broadest interpretation is that Windows’ approach optimises for developer flexibility and configurability (you can, via the registry, do all sorts of weird and wonderful memory tuning), whereas macOS’s approach trades all that off for a simpler user-facing regime (you can’t really adjust a thing - for example, I don’t think you can put the swap file on a different volume or device) which prioritises system responsivity.
A post with some of the gory details illustrating the complexity
Last edited:
The "cache" is managed by Windows, not Chrome.
Chrome wasn't "writing to cache", it was reading/writing to/from disk and windows was caching the storage.
Modern windows also uses memory compression, you can see this in the task manager in windows 10/11 listed as "compressed".
Out of curiosity, can you comment on how the Windows 11 performance is in UTM (and what general tasks you are doing). I use UTM for throwaway macOS VMs, but currently have Windows running in VMware Fusion because I remember reading that performance UTM almost unusable. I’d really prefer not to have VMware installed.
As far as I'm aware, UTM uses the Apple Hypervisor for virtualisation (on Apple Silicon), so it shouldn't be that different in performance to VMware Fusion or Parallels Desktop. I mostly use Parallels Desktop myself, but the disposable MacOS guests I've used on M4 Pro machines with UTM have worked quickly enough without issues.
Edit: The heat's getting to me .. just remembered I migrated service hosted on an M1 Mac Mini into a virtual hosted under UTM on an M4 Mac Mini 6 months ago ... the M4 virtual is quicker than the native M1 machine was (though some of that is down to 10GBe on the M4 and only USB-C 2.5GBe on the M1)
Edit: The heat's getting to me .. just remembered I migrated service hosted on an M1 Mac Mini into a virtual hosted under UTM on an M4 Mac Mini 6 months ago ... the M4 virtual is quicker than the native M1 machine was (though some of that is down to 10GBe on the M4 and only USB-C 2.5GBe on the M1)
Last edited:
It was a chrome issue. They fixed it with a later update. And it no longer does it.
I obviously know windows manages it but there was some code in chrome that was causing it to run too aggressively. It has since been fixed.
Last edited:
As usual. Can't wait for when A19/M5 suck as they've totally purchased the highest-end versions of them and concluded they're awful based on usage no one can ever be shown!So, I can see that OP disappeared after setting everyone on fire 🤣
As @eddie_ducking mentioned, UTM uses Apple Hypervisor for virtualization.
My initial primary use for Win11 on my M2 Mac Mini was to run ancient software for the Casio Pocket Viewer PV-400... I was shocked that it runs not only on Win11 but the ARM version of Windows... AND the USB-to-serial port dongle I need to connect this ancient hardware to a modern device is fully supported through UTM!
But having such good success with UTM w/Win11, I'm using it for more things and am pleased with the performance. Mostly office productivity stuff (no multimedia content creation). My ThinkPad runs Win10 and I refuse to put Win11 on it, so it is nice to have an option to tinker a bit on Win11.