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.
ssd swap - high usage of Terabytes Written
- Thread starter Forti
- WikiPost WikiPost
- Start date
- Sort by reaction score
- Status
It’s still a niche market to manually replace failed components. Businesses almost never do this unless it’s a small business. There are typically 5+ identical systems that are spares, IT swaps your broken one out and it gets sent for repair. IT in business typically has a 3 year cycle so warranty is always covered. That way if your system fails again, and it really needs to go back to Dell/HP/Apple, there will be no question about custom hardware you installed.
Personally I upgraded from a 2020 Intel 2.0 GHz 13 inch to an M1 MBP, so i guess i’m pretty qualified to answer this (also having intel macbooks since 2012).
Despite the SSD writes which, on my M1, i’ve fixed, this laptop is still miles and miles better than my intel macbook was. like night and day difference.
Even if the SSD was a permanent issue i’d still think it’s better. The battery life, thermals, no fan noise and the performance make this macbook SO much better than the intel one it replaced. plus it will have much longer support. And using ios apps is nice too.
Well after over 6 hours of uptime, running Edge with around 8-16 tabs consistently, Netbeans for Java Programming, GoodNotes, Discord as a permanent Edge tab, Spotify, Calendar, Reminders, and various other apps throughout today, I've only written around 9GB to the SSD.
That works out to around 1.33GB written per hour, which at this rate would give me an SSD lifespan of around 50-60 years! So yes, personally I consider this fixed and not a problem anymore.
I've done various things really since beginning trying to fix this issue, pretty much all of those I've reported here throughout this thread, I could probably write out every single step that I personally did at a later date. But essentially I'm using M1 Edge with Auto Tab Discarder, I've done a safe boot and first aided all my drive partitions, I've excluded my whole disk from time machine while not using it to backup, I've turned off spotlight indexing for things I never search for and I've excluded some weird startup items that I didn't recognise. Also I limit myself from using Rosetta 2 apps (for example I've side loaded the iOS spotify for the equalizer, but it turns out its actually better for my SSD too!) Oh and I've set the cache of Edge to Read Only so that it's forced to use the RAM instead of writing to disk cache.
That being said, I can't guarantee anything I did would help any other M1 system the way it did for me, because I'm not Apple. What I managed to fix personally might not be what others have to fix in their case. I also do use an external SSD in my work for the programming side of things, which may or may not explain the results. And since owning this Macbook from launch day, I do have 21.2 TBW in sysmontools due to experiencing the SSD write issue for around 3 months before starting to fix it and/or before Big Sur 11.2.3 which in itself seems to have improved the situation a lot.
Last edited:
Thanks for reply !
So as i understand;
- M1 Edge browser with auto dab discarder
- Partition aid
- Stopping time machine
- Stopping unnecssary indexing
- Using native apps as much as can
- Edge cache type = read only.
- Updating OS
Oh yes indeed I forgot to mention, I did the whole Onyx thing too, and that seemed to help also. And with edge I had to manually go to the Caches folder in the system folder and set it to read only, as there is no option for doing so in the browser itself
A guide with all those steps will be good for m1 mac users. Hope you or someone will write step by step or record a video whole process.
So, basically its a Software issue that can be fixed in next Big Sur updates? Or its a pure hardware issue?
There really aren't any serious problems for the *vast majority of people*, and the M1 Air is so much better than the Intel version, that I wouldn't even consider the latter. There will be OS updates that will address the issue if it proved to be serious for a significant proportion of users.
If you need a computer now (and can't wait up to 6 months for the next gen), then buy an M1 Air now. If you can wait, you may well get a big spec jump on a 2nd generation Apple Silicon Mac.
If you buy an Intel Mac now, you will make a big loss trying to sell it after the 2nd gen Apple Silicon macs arrive, because it would be completely outdated.
I don't know what happened, but I switched to Safari Technology Preview v122 and I noticed less RAM usage in general and less SWAP.
I did grow the habit of emptying cache. I did a keyboard shortcut to empty cache from System Preferences and when I am closing plenty of tabs (i.e. when I am working on a paper and have several tabs for papers and such), I empty cache easily with the shortcut.
The keyboard shortcut I used was Cmd + ` (the key left to the "1" key).
Personally, I really can't live without without my iCloud Keychain (I'm far too lazy to get 176 passwords from the keychain to BitWarden or other password managers) and Safari's Reader Mode.
I did grow the habit of emptying cache. I did a keyboard shortcut to empty cache from System Preferences and when I am closing plenty of tabs (i.e. when I am working on a paper and have several tabs for papers and such), I empty cache easily with the shortcut.
The keyboard shortcut I used was Cmd + ` (the key left to the "1" key).
Personally, I really can't live without without my iCloud Keychain (I'm far too lazy to get 176 passwords from the keychain to BitWarden or other password managers) and Safari's Reader Mode.
I’m confident Apple is both aware and busy working on fixes to these issues, so i’m not surprised to see technology previews show improvements!
I’ve grown used to Edge now but Keychain with Touch ID and the tab overview are definitely features i miss.
Nope! I'm on macOS 11.2.3.
Safari Technology Preview is v122, equivalent to Safari 14.2.
11.3 Beta 4 has Safari 14.1.
Agreed. Also, I think Edge has tab overview as well. It should be in the experimental flags somewhere I think.
I'll probably regret asking this, but three questions:
1) How involved would it be to actually replace the soldered SSD if needed? What would Apple charge? How hard would it be for an independent repair shop to perform the operation, and would it be worth the labor costs?
2) Why the heck does Apple insist on soldering SSD's and memory? If it's not to hinder users being able to repair or supplement their own computer hardware, what in the world else justifies such decisions?
3) What is the technical reason for the 16 GB memory limit to the M1 minis?
In any case, I have a 3 year old iMac that needs replacing soon, and given that iMacs are in the short term tied to Intel architecture, I'm considering getting a Mini. Over the last 10 years Apple has continued to disappoint me in various ways, such that I no longer buy Apple laptops (my last was a 2006 MBP), iPad reliability problems have led me to refrain from buying more of those, and now iMacs are in transition (thank god I didn't invest in an iMac Pro,which is being discontinued). I could go on and on, but my wife really needs the continuity provided by remaining in the Mac Apple garden with photos and so forth, so I'd really like to be confident in getting a Mini soon. Any answers appreciated to the above. I myself have just about gone full Linux with laptops - and soon - desktops as well.
1) How involved would it be to actually replace the soldered SSD if needed? What would Apple charge? How hard would it be for an independent repair shop to perform the operation, and would it be worth the labor costs?
2) Why the heck does Apple insist on soldering SSD's and memory? If it's not to hinder users being able to repair or supplement their own computer hardware, what in the world else justifies such decisions?
3) What is the technical reason for the 16 GB memory limit to the M1 minis?
In any case, I have a 3 year old iMac that needs replacing soon, and given that iMacs are in the short term tied to Intel architecture, I'm considering getting a Mini. Over the last 10 years Apple has continued to disappoint me in various ways, such that I no longer buy Apple laptops (my last was a 2006 MBP), iPad reliability problems have led me to refrain from buying more of those, and now iMacs are in transition (thank god I didn't invest in an iMac Pro,which is being discontinued). I could go on and on, but my wife really needs the continuity provided by remaining in the Mac Apple garden with photos and so forth, so I'd really like to be confident in getting a Mini soon. Any answers appreciated to the above. I myself have just about gone full Linux with laptops - and soon - desktops as well.
As I pointed out before some people have reported similar problems on Intel Macs. So this isn't unique to the M1 or Big Sur.
You snipped the other possible candidates: trying to use a 8 GB Mac as if it was an 16 GB one, running Chrome (or another other browser that eats RAM like candy)
If everyone was having this problem then I would agree but not everyone is having this issue. Logically if is was Apple screwing up again (as with that Safari OS update fiasco late year) then everybody using Big Sur on an M1 and checking the specks would be seeing this issue.
The real biggest issue is most people are freaking out on the raw TB and not the percentages.
Lonehorn shows 1% in about 2 months - that is 200 months (16.6 years) for 100%.
Dan Moren's numbers are even better as 1% in 3 months is 300 months or 25 years.
His 2017 iMac's 14% in 42 months (3.5 years) produces the same life span - 300 months (42*100/14).
These are examples of people freaking out over the raw TB without following though on the math. If the tools they are using can be trusted than the percentages produced must also be trustworthy.
I'm old school and use 1Password 6 (the non online one) to store my passwords with open source KeePassXC as a backup.
I have a theory as I have both an 8GB and 16GB, and see a difference between the 2 in terms of overall written, did some testing, and seems to correlate with, in my case optimized and Rosette apps ... Ran both for a day on only "Native" apps and next day to use "Rosette". Day 1 nearly the same low figure, day two the 8GB had higher (nearly double) of the 16GB .... YMMV
In answer to your questions:
1) It would be very difficult, and probably not cost-effective, to replace a soldered-on SSD. Apple would almost certainly just replace the whole motherboard, which is essentially the whole computer minus the fan and external connectors & switches
2) Soldered components are more reliable because there is no possibility of physical shifting after manufacture. I've had SSDs and RAM chips move and develop intermittent failures. They are also probably faster because (a) you are removing physical interconnects that may have capacitive effects that affect maximum transmission speed (b) you can make the physical conductor distances shorter without having to have a socket into which the SSD/RAM chips need to sit. At these frequencies, even a few millimeters makes a difference.
3) The RAM chips are located on the SoC package, so the physical size of the SoC limits the amount of RAM. At the time of manufacture, there weren't readily available (& cheap) 16GB memory chips, so the M1 could only fit 2x8GB. There might also be a limitation on the number of memory controllers that could fit on the M1 die, which limited the maximum addressable memory. I doubt it is an inherent shortcoming of Apple Silicon, just a decision about the limits to apply to the first generation, entry-level chip. Apple (fairly reasonably) decided that for an entry-level Mac, 16GB was enough, and that people with a need for more would be happy to wait until the more powerful 2nd generation.
Do we know exactly what the smartctl percentages actually mean?
Does 1% mean that *all* the SSD cells have, on average, been written to for 1% of their guaranteed number cycles (e.g. 30 times for an SSD that has 3000 write cycles per cell)
Or does it mean that 1% of the cells have now exceeded their guaranteed number of writes?
The first of these would be less alarming because you could (as you have done) estimate the point at which the entire disk (assuming some kind of wear-leveling is used) would be about to reach its guaranteed write limit, above which we would expect to see cells marked as permanently failed. The disk would be "on borrowed time" but it doesn't mean that anything has failed yet, and we could still have the full capacity and have time to make backups.
The second would imply that some cells have already exceeded their limit, and have either failed, or could fail imminently. Presumably, if this were the case, we would eventually start to see the available disk capacity decrease, as failed cells were removed from the total available.