ssd swap - high usage of Terabytes Written

[LowAffinity]

Format SMART data sizes in units based on powers of 2 (1k= 1024) gets you GiB
√ Format SMART data sizes in units based on powers of 2 (1k= 1024) gets you GB

So without a √ it counts in GiB.
With a √ it counts in GB.

Easy to check. Try both.
The smaller number you get is in GB. (Division is by 1024^3)
The larger number you get is in GiB. (Division is by 1000^3)

Okay?

Thank you very much for clearing that up!

 

[Mike Boreham]

According to Drivedx, my M1 is still in the GB range.. Does this mean I can now trust Drivedx is reporting the same numbers as smartmontools would?

DriveDX always has given same as smartmontools if you converted the DriveDx units to TB.

 

[MrGunny94]

I got my Air a few weeks ago, and using every day between 8-10h this is how it looks like. It's a base model

I use this device for everything besides gaming.

 

[smirking]

I picked up a refurb 8GB MBP that's being used very heavily for development. I went with the 8GB just as a trial to see if it could work with the intent of upgrading to a 16GB when the 16" models come out, but this disk use is a little scary.

After 10 days I'm at close to 4TB and 276 power cycles. Some of these must have been already logged by the previous life of the machine, but I saw the data units written go up by .01TB in the 20 minutes I was reading this thread.

 

[Paulo Henrique Baldassi]

I picked up a refurb 8GB MBP that's being used very heavily for development. I went with the 8GB just as a trial to see if it could work with the intent of upgrading to a 16GB when the 16" models come out, but this disk use is a little scary.

After 10 days I'm at close to 4TB and 276 power cycles. Some of these must have been already logged by the previous life of the machine, but I saw the data units written go up by .01TB in the 20 minutes I was reading this thread.

View attachment 1798697

First thing is for you to inform which version of Macos is running on your machine! Big Sur 11.4?

 

[Maximara]

I picked up a refurb 8GB MBP that's being used very heavily for development. I went with the 8GB just as a trial to see if it could work with the intent of upgrading to a 16GB when the 16" models come out, but this disk use is a little scary.

After 10 days I'm at close to 4TB and 276 power cycles. Some of these must have been already logged by the previous life of the machine, but I saw the data units written go up by .01TB in the 20 minutes I was reading this thread.

View attachment 1798697

First rule: if you are doing development do not skimp on RAM. Doing so is penny wise pound foolish.

Did you check this before you did anything? How much of that 4 TB was already there?

The worst case we have is a 518 GB drive going bye bye at the 600 TB mark and there are serious questions about those numbers (they don't jive with others reporting their usage) and the M1 Mac was used in something that is write happy.

The key number is not how much is written on its own but also Percentage used. A 0% means at worst only 0.49999.% of the SSD's rated lifetime capacity has been used. This gives us at worst 800 TBW before it hits 100% (and most SSDs even in 2014 went way past that number)

If 10 days gets 4 TB than it should take 2,000 days (800/4*10) to hit the 100% mark. That is ~5.48 years - more than enough time to get an external or replace the machine. And this is the worst case. May people are coming up with 1000+ TB for their drives.

 

[osplo]

Did you check this before you did anything? How much of that 4 TB was already there?

That's the million dollar question, Smirking. If you did not check before, try to see the delta for a week or so.

 

[smirking]

First thing is for you to inform which version of Macos is running on your machine! Big Sur 11.4?

Yes, it’s 11.4. It’s the latest one.

First rule: if you are doing development do not skimp on RAM. Doing so is penny wise pound foolish.

Entirely accurate. I wanted to try it as an experiment just to see if I could get by and my answer so far after 11 days is that I indeed can... but just because I can, doesn't mean I should. 😂

I'm sure that most of that mileage appeared while I was in possession of the unit. The websites I develop have databases in the GB range and they're constantly making large cache writes to disk even with a very generous allocation of RAM so in my case, skimping on storage would probably also not be advisable.

This experiment worked out better than I expected, but I'm still returning this M1. I wasn't quite ready to upgrade yet but if Apple refuses to allow us options to service our own batteries and drives, I won't feel bad about taking a refurb out for a hard test drive while my daily driver is in the shop. I went into this thinking it'd be a 50/50 shot. If it blew me away on all fronts, I'd keep it.

I'm eager to get my 2018 back from Apple tomorrow so I can run some diagnostics on that and see how hard I've been hitting the SSD there. If RAM will greatly lighten the wear on the SSD for my use case, the tale of the tape should be quite different there as my 2018 is a 32GB/2TB unit.

 

[smirking]

That's the million dollar question, Smirking. If you did not check before, try to see the delta for a week or so.

Well, the idea of checking disk writes wasn't even a vague notion to me before I stumbled into this thread. I've never had any problems with my SSDs failing so it's just not something I ever worried about despite my very disk heavy use cases.

I do believe that I was responsible for most of that 4TB in 10 days. Every time I clear my database cache while developing it goes up a small, but notable amount. On the one day when I didn't do much work and just browsed websites, the disk writes only nudged upward a little.

I'm getting my 2018 MBP back from service tomorrow. The first thing I'm doing is checking to see how much mileage I put on it in its 2+ years of very heavy daily use as a development and photo editing battlestation. I have a lot more RAM in that 2018 unit, but I'm going to guess that the SSD is still going to show some hard mileage.

 

[Maximara]

This experiment worked out better than I expected, but I'm still returning this M1. I wasn't quite ready to upgrade yet but if Apple refuses to allow us options to service our own batteries and drives, I won't feel bad about taking a refurb out for a hard test drive while my daily driver is in the shop. I went into this thinking it'd be a 50/50 shot. If it blew me away on all fronts, I'd keep it.

I don't understand this "options to service" mentality. As I mentioned before based on the numbers you gave us and using the worse case situation you were looking at 5 years before the drive hit 100% and the 2014 test shows odds are the SSD would go way past that point. External SSDs are cheap and small enough that in most cases the "dongle" argument is moot.

As for the battery that is a little more on the usage side - what the expected usable lifespan of the machine going to be? If it is substantially north of 2 years yes that is an issue provided having an outlet nearby is not an option.

 

[smirking]

As I mentioned before based on the numbers you gave us and using the worse case situation you were looking at 5 years before the drive hit 100% and the 2014 test shows odds are the SSD would go way past that point. External SSDs are cheap and small enough that in most cases the "dongle" argument is moot.

I was floating around 20G/hr in writes. On a 500GB drive, I guess that should amount to 5-7 years, but that wasn't anywhere near my worst case scenario. That was me working pretty leisurely for a week without pressure.

So yeah, I know that before the SSD is likely to fail, I'm likely to have moved on to another computer, but the excessive writes were a clear sign to me that my 8GB experiment wasn't going as well as I had hoped.

I'm not obsessed with having a pristine machine. I'm obsessed with avoiding the possibility that my computer is going to act up on me when I can least afford it.

 

[smirking]

Ok, here's my disk write data on my 2018 MBP with 32GB RAM for comparison. To recap, I've been test driving an 8GB M1 MBP while my 2018 MBP was getting battery service. I intentionally got the 8GB model just because I wanted to see if I could drive my workload on an "under resourced" machine. Even on only 8GB, it did quite admirably, but I was racking up very heavy disk writes that weren't attributed to the bug that was fixed in OS X 11.4.

So here's what my disk stats look like for my 2018 that I've been using for almost 30 months. It's at 60TB of writes.

In comparison, here's what my 8GB M1 MBP looked like after 9 days. It's at 4TB of writes.

Either the kind of things I do is just going to be like this on M1's or skimping on the RAM really puts a surprisingly heavy burden on the SSD.

 

[crucifixx]

I am on Mac OS Monterrey Beta 2 currently. Previously when I saw people here saying that this issue has been fixed with 11.4, I tried to return back to Safari since it's my fav browser. But the SSD writes that comes from using Safari just bananas, I'm talking about 100GB+ per hour during my working hours. I switched back to Microsoft Edge and it's been much much better. I can for certain say that it is a Safari problem. I didn't mess with any memory saving extensions or anything like that for Edge. Also didn't change any sleep/wake settings in the terminal. Everything is default and I managed to curtail the problem just by not using Safari. After switching back to Edge the write is a more reasonable 100GB per day.

 

[vs40]

I am on Mac OS Monterrey Beta 2 currently.

I can for certain say that it is a Safari problem

Or Monterrey Beta problem...

 

[crucifixx]

Or Monterrey Beta problem...

I tried the same thing on 11.4 with the same results. Upgrading to Monterrey was mostly to check if they've fixed Safari or not.

 

[vs40]

My results after few days of regular usage(Home Office, browsing, YouTube, streaming movies, light photo editing), when all initial setups/updates/installations are done.

29.06 / 09:00 AM
Read: 393GB
Written: 290GB

30.06 / 01:00 AM
Read: 429GB
Written: 320GB

01.07 / 01:00 AM
Read: 443GB
Written: 334GB

I'm totally fine with it 👍

 

[osplo]

Ok, here's my disk write data on my 2018 MBP with 32GB RAM for comparison. To recap, I've been test driving an 8GB M1 MBP while my 2018 MBP was getting battery service. I intentionally got the 8GB model just because I wanted to see if I could drive my workload on an "under resourced" machine. Even on only 8GB, it did quite admirably, but I was racking up very heavy disk writes that weren't attributed to the bug that was fixed in OS X 11.4.

So here's what my disk stats look like for my 2018 that I've been using for almost 30 months. It's at 60TB of writes.
View attachment 1799268

In comparison, here's what my 8GB M1 MBP looked like after 9 days. It's at 4TB of writes.

View attachment 1799269

Either the kind of things I do is just going to be like this on M1's or skimping on the RAM really puts a surprisingly heavy burden on the SSD.

But yet you still don't know if these 3.99 TB are a result of your working week for sure. You mentioned that this Mac was a refurbished one, right? So most of these 3.99 TB could have been written by the previous owner.

I beg your pardon but, again, I think you need to work for another week and then compare...

 

[smirking]

So most of these 3.99 TB could have been written by the previous owner

It’s possible, but not likely because what led me to my conclusion was that I racked up .6TB over the course of 12 hours. At that point the question of how much of the writes came preloaded was no longer relevant to me. I was on pace to hit 1TB a day and I hadn’t been doing my most resource intensive work.

 

[smirking]

My results after few days of regular usage(Home Office, browsing, YouTube, streaming movies, light photo editing)

8GB or 16GB?

 

[vs40]

8GB or 16GB?

16GB

And small updated after another 3 days

04.07 / 0:25 AM
Read: 494GB
Written: 400GB



It is 101GB read and 110GB written in 5 day with my regular usage.

PS
I just find out, what different video players have drastically different approach on caching while streaming.
For example Movist player caching everything on SSD and it can be pretty intense if you for example often stream heavy BDRemux files 50-70GB each.
Meanwhile VLC and IINA players consume more RAM, but without huge cache on SSD.

 

[pshufd]

16GB

And small updated after another 3 days

04.07 / 0:25 AM
Read: 494GB
Written: 400GB

View attachment 1801817

It is 101GB read and 110GB written in 5 day with my regular usage.

PS
I just find out, what different video players have drastically different approach on caching while streaming.
For example Movist player caching everything on SSD and it can be pretty intense if you for example often stream heavy BDRemux files 50-70GB each.
Meanwhile VLC and IINA players consume more RAM, but without huge cache on SSD.

Thanks.

I installed VLC last night and I’ll just stick with it for now.

 

[MrGunny94]

I'm hoping Apple ends up releasing 16GB model next redesign of the Air and the Pro (not the first base model but the upgraded one they usually sell at amazon and etc)

 

[pshufd]

I'm hoping Apple ends up releasing 16GB model next redesign of the Air and the Pro (not the first base model but the upgraded one they usually sell at amazon and etc)

I don't think that they will. Industry standard practice is still 8 GB and Macs do run just fine on 8 GB for your typical user. The typical user in these forums is more of a power user.

 

[smirking]

I don't think that they will. Industry standard practice is still 8 GB and Macs do run just fine on 8 GB for your typical user. The typical user in these forums is more of a power user.

The typical user in these forums is also unaware that 8GB is actually quite capable and as "future proof" as an average user would need.

 

[MrGunny94]

I don't think that they will. Industry standard practice is still 8 GB and Macs do run just fine on 8 GB for your typical user. The typical user in these forums is more of a power user.

Definitely but might have to jump to the 14" then