Well I'm not currently seeing any issue on my M1 MBA. I have reasonable numbers in the TBW slot; about 7 GB after 100 days or so. But I don't want to have to monitor it so I would love an official response from Apple on what's going on with people having 10x that amount of TBW or more. Obviously there is some trigger to this that we still don't have an explanation for. There are a few reports that Apple may have fixed something in 11.2.2. We'll see.
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.
M1 Mac Users Report Excessive SSD Wear
- Thread starter MacRumors
- Start date
- Sort by reaction score
Instead of everyone fighting, why not respond to this guy? He said 11.2.2 may have fixed this issue.
I hope so but I would still like some sort of official statement from Apple on what's changed and what they fixed. Like I said above, I don't want to have to monitor this. If I knew what caused some others to have very high writes in a short period of time, I could adjust my use to make sure it doesn't happen. Or if Apple states that they fixed a bug, that would be fine too. The absence of any statement by Apple is a problem in my opinion. I would report a bug to Apple except I'm not personally seeing the issue.
There is an unofficial statement from Apple the numbers are wrong.
Do you have any proof that Apple is wrong in this assertion? Do you know what the true and accurate numbers should be?
We have a statement (albiet unofficially) from Apple explaining the numbers used to calculate the extent of use were wrong. It would seem to me that is what "caused" this "issue". What evidence do you have that would counter Apple?
If you are not seeing the issue, this would seem to be evidence this is not a universal problem. This would seem to strengthen the unofficial position of Apple this issue was caused by faulty programming generating numbers which are not accurate.
If you can prove something contrary to this, now is the time.
Read this thread. I've done just that. It is pretty simple arithmetic to verify that smartmontools is working correctly. See this post:
I've also compared the smartmontools code (which is open source) with the official NVMe specification as posted by @killerovsky here: https://forums.macrumors.com/threads/m1-mac-users-report-excessive-ssd-wear.2285892/post-29640500.
The supposed statement that came from AppleInsider is interesting but we have no way to know if it came from someone with knowledge of the issue or was just a PR flack trying to cover the problem. Anonymous statements from a rumor site isn't very convincing while my experiments have convinced me. I was originally pretty skeptical that the tools were working correctly but I put in some effort to read the code and verify against the specification.
I think that the anonymous statement from someone at Apple doesn't hold much weight. I do think it isn't a universal problem since I don't seem to have any issues. But I want to know why some people are seeing an outrageous amount of SSD writes since if I don't know what causes it, I can't avoid it myself. I don't want to have to continuously monitor my MacBook Air if I don't have to.
Edit: I have the link to my post wrong. Sorry for any confusion. The correct link is: https://forums.macrumors.com/threads/ssd-swap-high-usage-of-terabytes-written.2284893/post-29639724
Last edited:
Really it isn't that hard to verify that the smartctl numbers are accurate. You have multiple ways of measuring SSD writes that you can use to compare. It isn't about "I said so" it is about doing a reasonable experiment to verify. See: https://forums.macrumors.com/thread...n.2284893/page-13?post=29639724#post-29639724
Since I installed Big Sur 11.2.2 this morning my swap hasn't been above 64MB (M1 Mac Mini 16GB/512). They must have added something in the update:
Interesting!
Let's hope Apple actually fixed what was causing the problem... not simply changing the reporting.
That would be the equivalent of putting a piece of tape over the Check Engine light...
I received an open box M1 MacBook Air from Best Buy yesterday. It was registered on February 12th, so the first owner barely used it considering when I bought it, and the shipping time. After using it minimally (checking it for defects and such), I downloaded DriveDx and recorded about 825gb written to the hard drive. I moved basically nothing to the computer in terms of files. I downloaded the new update late last night, used the computer VERY minimally today, and just recorded about 852gb written. That's over 25gb written for downloading and installing the update, and looking at a few webpages today. I plan to use it a bit tonight to see how much more is written to the drive before I transfer anything over to it.
8gb/256gb model
8gb/256gb model
Last edited:
It's positively impossible if you don't know what the numbers should be. This seems to be the thrust of the Apple "unofficial" statement. Apple knows, you don't... end of story. Full stop.
The point would seem to be that you don't have any earthly idea what the numbers should be, and you justify your statements by saying what you think they should be is what they are, i.e. because I said so.
Again, selective delusion mode... engage!
Nice!
I'm using sata-ssd's, so NMVe spec does not help me and I'm sorry that I don't have time to review NVMe spec, but maybe you know and can answer:
How does the NVMe SMART calculate power-on-hours?
Let's say that computer writes 1kB to the NVMe ssd every second.
Are all those seconds summed or are they calculated as a fractions of full throughput of that drive?
If the drive full write speed is eg. 100 MB/s, is writing a 1kB/s calculated as a 0.00001 second to power-on-hours for each second?
Hahaha touché!The only way for customers to "push" for anything is with their wallets. (take a scan of the receipt of a competitor's product and mail it with a note to Apple indicating why you went with a competitor) Apple has zero incentive to do anything differently if people complain but buy their products anyways... or giving them more money by paying for upgraded options (if they think that 16GB RAM will extended the life of the SSD).
They can't hear the grumbling and complaining over the sound of the bulldozer engine raking in the mountain of cash.
Can't agree that 256 is too small. especially that a lot of things are in cloud and movies/shows/music comes from streaming services. I have MBP M1 with 256 for 2 months and I barely used the space, sitting on 195GB free.
as of the 60GB per day, doesn't look like barely used for sure but also not that excessive use. I mean, with my daily coding routine i'm eating around 20GB write (10 comes from kernel task and launchd). BUT, a simple thing like updating the OS to 11.2.2 written around 15GB to my disk. so simple tasks can write quite a lot and bump up the average numbers.
not negating the overall problem here as I hit 1TB in under two months on this M1 while my friend on Intel has 3.5TB written in 2 years.
I guess only you can answer that but it has given me pause.
Normally I would never consider a 1st gen release of any kind, Apple gets the benefit of the doubt (from me) most of the time but in this case... the old axiom about avoiding 1st gen products seems yet again apropos.
I see. The thing with this SSD issue is, no one is 100% sure. There are those who say it exists, and those who say it doesnt. Apple is silent.............sigh
I'm somewhat old school and prefer my files on my personal hard drives (largely because grew up when internet speeds sucked and if my internet goes off-line I can still do stuff). Also if you want things at the system level (where a lot of installers put stuff) your Application folder can fill up fast. Xcode for example is about 30 GB, Diable III takes out another 18 GB, Starcraft 29 GB, Warcraft III 30 GB, and so on.
59 GB per day still works to be 21,535 GB in a year or 21 TB, well above what even you are seeing (which is 6 TB).
I am concerned about this potential issue, but it didn't stop me from buying a refurb as soon as they became available. My current Mini is 6+ years old, so I have been feeling the pressure to upgrade. I didn't consider waiting 2-3 more years for the next mini as much of an option. Replacing my 6 year old Intel with a 3 year old Intel wasn't very exciting either.
It is up to you whether it is worth waiting a few days, weeks or months to see how this plays out. Depending on your situation, an M2 iMac or MB might also be an option for you later in the year, maybe very soon.
Last edited:
Thanks! I'm not sure I want to take this on as a real project but that's at least a small hole in my theory.
Apple can’t say anything much. At least they should say “we’re looking into the issue.”
I'm not sure why your response is vaguely ad hominem. I'm not trying to shame you, I'm just throwing out a theory based on my anecdotal experience and the collective experience of others. Personally, I think Chrome is broken but I don't think that mean you're a bad person for using it.
I have an M1 Mac with very normal/nominal wear (two actually, but I didn't do a deep dive on my first other than the SMART data report drive health as "normal"). Yes, there's a chance something is wrong with the software or silicon but SSDs are pretty mature tech. Apple's implementation of putting the controller on the SOC is fairly abnormal in the PC-space but not atypical in other types of computers. The problem also isn't universal. So it seems more likely there is some underlying usage pattern causing this—or at least contributing to it.
I'm all for making some noise, and holding Apple accountable, if there is a real issue. Meanwhile, putting our collective experience and intelligence to use to find underlying usage patterns or causes could help ameliorate the issue in the meantime. Isn't that better?
What if there is an issue but it isn't on Apple's end? Then what? For all we know this is some third party program that is running wild because programing was less then stellar.
But they shouldn’t be. That’s one of the things Linux does correctly. Use RAM as much as possible and never touch the swap file until you absolutely must.
Linux Performance: Why You Should Almost Always Add Swap Space indiaces this may not exactly be true:
"It is normal and can be a good thing for Linux systems to use some swap, even if there is still available RAM."