Thanks for so many answers. I guess there was a bug or something with my Activity Monitor. Since I rebooted my Macs Actvitiy Monitor show the WindowsServer constantly around 300-800MB. Swap does not really happen, even though I worked yesterday again in Linux VM.
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 Air - Is my swap usage normal?
- Thread starter phl92
- Start date
- Sort by reaction score
More likely that WindowsServer has a memory leak and it was consuming resources but not releasing it
Yes, swap and some other cache files aren’t cleared until a restart.
Agreed.
From what I’ve read, it’s typically:
• 256GB = 150 TBW
• 512GB = 300 TBW
• 1TB = 600 TBW
...
The following article states ~50% endurance improvement rather than 100% increase. However, it is a few years out of date.
How Long Does an SSDs Last? | Calculate Your SSD’s Lifespan
Learn how long your SSD will last and how to maximize their lifespan with our comprehensive guide.
www.enterprisestorageforum.com
Also, FYI, you cannot calculate the wear linearly. For example, when half of the cells are unreliable/unusable, the remaining half will presumable be written twice as often (assuming the workload remains constant).
I have MacBook Pro M1 16GB (500 SSD), swap for me is standard practice, when working in FCP - my record 🤭 is 43 GB. Only one rules, what I have - when project is really large, I work from external Thunderbolt SSD Samsung 980 Pro 2TB. To make max free space in MacBook Pro SSD for swap memory.
Attachments
sorry, bit of a newb with some of this stuff, have searched but can't find the answer..
My 8/256 macbook air has been running for about a week without a reboot, in Activity monitor it shows 0 bytes of swap used, is that a measurement of the amount used since the last reboot, or the amount currently being used?
Thanks,
My 8/256 macbook air has been running for about a week without a reboot, in Activity monitor it shows 0 bytes of swap used, is that a measurement of the amount used since the last reboot, or the amount currently being used?
Thanks,
Yes, it is excellent practice to reboot any digital device from time to time (Mac, PC, tablet, phone, camera, etc.). It clears everything out and reinitializes the software and hardware. I reboot my M1 Mac Mini about once a week. I have the Reopen windows when logging back in option checked so in a few seconds I am good to go right where I left off.
Honestly it doesn't look too bad. Here's what mine looks like after about two weeks of usage - mine's a base model 8/256, and it's still plenty snappy for my needs 
Exactly this. It’s about 1.6PB of writes per 256GB.
People should just use the computer and stop checking Activity Monitor to worry themselves about things that they shouldn’t be worrying about.
Would someone please provide a link/source for the 1.6PB thing? Thanks!
I’d like to hear that too. Typically for older higher end 512 GB drives it was around 300-400 TB. For low end third party 256 GB consumer drives it may be as low as 60 TB.
Here is the table for the Samsung 960 Pro:
FWIW, DriveDx reports the OEM Apple Samsung 256 GB drive in my 2015 MacBook Pro has 67% wear health after 100 TB writes.
You cannot compare a regular consumer SSD with a specialized hardware designed for Apple, exclusively for Apple Silicon, there's no similar model in the market, the SSD plays a key role in the Apple silicon design so it makes sense it is overprovisioned. The 1.6PB info comes from the hardware controller itself using smartmontools if you do a simple calculation based on the Percentage Used threshold informed. Ex: My SSD is currently at 118TBW but only used 8% of its rated capacity after more than a year of heavy usage.
What is your SSD size? And the 960 Pro isn't a regular consumer SSD. It was their high end prosumer model.
Anyhow, colour me skeptical that Apple's 256 GB drives are designed for 1.6 PB writes. That's not suggesting it will die at 300 TB though of course.
Also, I don't see what you mean by "the SSD plays a key role in the Apple Silicon design". It's an SSD, and functions like an SSD (albeit with a different location for the controller).
base model 256gb. Don't need to be skeptical, there are users on this forum that crossed the 1PB mark because of the early memory leak issues on Big Sur.
Lol, the SSD is literally glued in the SOC and it doesn't even have its own controller, it's fully managed by the SoC. The memory management is also dramatically different on Macs with Apple Silicon as Apple let it swap a lot more since the IO is so fast it almost almost behave like more RAM, specially because the RAM is shared with the GPU.
Just because an SSD can reach 1 PB writes doesn’t mean it’s designed for it. Conventional consumer SSDs can also reach 1 PB writes too but they aren’t specifically designed for it.
And no, the SSD is not “glued in the SoC”. That’s just flat out wrong. Hell, on the Mac Studio, the SSD is actually even removable.
As for the controller, yes it’s the SoC. That was my point by saying the controller was in a different location.
Dude I'm just showing facts, the information is accessible with simple tools, if you wanna live worried about the SSD go for it.
Also just because you can remove the SSD on the Studio doesn't mean it's "designed for it". The SSD is actually dumb storage controlled by the Apple Silicon as noted by the Asahi Linux developers. You keep comparing it to consumer SSDs when it's just not. You can't buy one of those, you can't replace one of those, so they are not for consumers.
Apple doesn't sell SSDs, they sell complete products, so they don't have to rate the SSD life like if it was separated product, they likely extensively tested these SSDs to reach those numbers reported by the controller.
Last edited:
Nice pivot. First you falsely claimed the SSD is glued into the SoC, but now you have shifted to talking about how Apple is not in the SSD sales business.
Anyhow, basically you’ve posted conjecture, along with some flat out completely wrong claims. That’s not helping anyone.
You're the one spreading useless skepticism. What is false about the SSD being glued into the SoC? Have you seen the pictures? You'd need special tools to remove the SSD chips and for what? You can't just buy another. So for the consumers it is glued as in NOT USER REMOVABLE.
I'm done with this nonsense. Have a nice day.
SmartMonTools and the like are only reporting what Apple has recorded via their SoC's SSD controller. If you are skeptical then you'll have to take it up with Apple. If the tools are reporting 8% used after 118 TB then you can expect to get somewhere around 1.4 PB writes. And before you speculate that the tools are wrong, they aren't. I verified that Apple's own NVMe SMART APIs are reporting the same as SmartMonTools. So if the tools are wrong, it is purely Apple's fault.
My issue here is that the data is not necessarily being used appropriately by some. Let me give you the example of a Samsung 860 EVO's SmartMonTools report. This is an inexpensive consumer drive, meant for stuff like business desktops and mainstream home computers. Well, some dude stuck a 500 GB Samsung 860 EVO in a server, and wrote 944 TB of data to it. According to SmartMonTools, it still had WLC of 1% left. So, if you extrapolate that, it means the drive should have a life expectancy of around 950 TBW right?
Well, no, because Samsung's own spec rating is 300 TBW for this specific model at this specific size.
Does that 300 TBW spec rating mean it's going to suddenly die once it hits 300 TBW? No, obviously not, considering it already had hit 944 TBW with zero problems. I wouldn't be surprised if it hit well over 1 PB with no issues, but that doesn't mean it was designed to hit 1 PB.
Last edited:
Except Samsung specs out their TBW as part of their warranty. Apple doesn’t so they have no reason not to report the remaining life accurately.