• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

MX500 rapid erase cycles, seems to be poor wear levelling

Okay hit me with it chief. Is this good or bad?
Capture.PNG
 
Hi,
Temperature is terrible at 45c
 
Hi,
Think crucial has gone down hill personally
Samsung has some crappy series too
I remember their 840 ssd's needing that emergency firmware, to stop bitrot.

In my opinion this issue isnt as bad as that as the ssd still functions normally and probably will do for the duration of warranty, just with accelerated erase cycles.

Okay hit me with it chief. Is this good or bad?
View attachment 218097
How old is the ssd?

You have 245 erase cycles with 4131 hours, 28956GB write.

When I first looked I thought thats quite bad as only about a 1/3 more hours than mine but with 2.45x the cycles, however you have written 29TB of data. About 15x the amount I have written.

The guys on toms hardware reckon if you have many power cycles the problem doesnt show up, you have nearly 3k power cycles, more than half of the power on hours, so averaging less than 2 hours per power up.

So to me given the host writes I think it looks much better than mine, your F7 is also much closer to your F8.

You are I think roughly 2.4x write amplification at a quick guess. Everything looks good, just you are writing to the drive heavily.
 
Last edited:
I remember their 840 ssd's needing that emergency firmware, to stop bitrot.

In my opinion this issue isnt as bad as that as the ssd still functions normally and probably will do for the duration of warranty, just with accelerated erase cycles.


How old is the ssd?

You have 245 erase cycles with 4131 hours, 28956GB write.

When I first looked I thought thats quite bad as only about a 1/3 more hours than mine but with 2.45x the cycles, however you have written 29TB of data. About 15x the amount I have written.

The guys on toms hardware reckon if you have many power cycles the problem doesnt show up, you have nearly 3k power cycles, more than half of the power on hours, so averaging less than 2 hours per power up.

So to me given the host writes I think it looks much better than mine, your F7 is also much closer to your F8.

You are I think roughly 2.4x write amplification at a quick guess. Everything looks good, just you are writing to the drive heavily.
Holy fuck! 29TB of write? I even moved my Users folder to the Western Digital hard drive to mitigate wear and tear on the SSD. It should be just AppData, ProgramData, both Program Files, Microsoft software (edge and whatnot) and Windows itself. Think it's the pagefile?

And I bought it in early 2019 or at the end of 2018.
 
View attachment 218099
I'm in tropic of cancer btw.
Hi,
We aren't that far apart
Difference might be central air conditioning and not.

I remember their 840 ssd's needing that emergency firmware, to stop bitrot.

In my opinion this issue isnt as bad as that as the ssd still functions normally and probably will do for the duration of warranty, just with accelerated erase cycles.
Yep plus if you do rma you'll get a refurbished back not a new one this is crucial's rma policy.
 
Holy fuck! 29TB of write? I even moved my Users folder to the Western Digital hard drive to mitigate wear and tear on the SSD. It should be just AppData, ProgramData, both Program Files, Microsoft software (edge and whatnot) and Windows itself. Think it's the pagefile?

And I bought it in early 2019 or at the end of 2018.
It could be a few things.

First windows logs a ton of data, Microsoft take the approach to run their logs in very verbose mode. Every time sync is logged, every svchost startup is logged, and much more on top of that, they also use cyclic logs so once the logs hit the size cap the full size of the log is written for every line that gets updated. So e.g. if a log is 20MB, when one line is added to the log, the 20MB is rewritten. Some of these logs are adding lines every few seconds. (they are also the source of many scams as the scammers know these logs fill up full of crap).
Next of course as you said is the pagefile, how much this writes I dont know, Microsoft themselves have said on a blog pagefile is much more reads vs writes, this will likely depend on memory pressure's on your system.
Hibernate another possible cause if you use hibernation.
In regards to AppData and ProgramData, AppData is used much more than you think, some examples.
Web browsers by default store their profile in AppData, and web browsers write a lot of data, not just temporary cache, but also now have their own virtual filesystem, app cache, code cache, logs, the journaling for the logs, their databases, the journaling for the databases, then there is browser extensions which can also be write heavy depending on the extension. Also almost forgot, some will cache media as you stream it. Netflix, youtube etc.
Google drive will cache every file thats synced temporarily in its AppData folder.
Steam will download any game first inside its Program Files folder, before moving it to the destination.
Macrium logs inside ProgramData.
The default system temp folder is C:\Windows\temp. The default user temp folder is inside AppData.
Finally modern windows update has very large patches now due to the cumulative patch system

To give you an idea, I use both volume shadow copy and also do Macrium incremental backups, I moved my documents, pictures folders to a spindle, I also moved the user temp folder to a spindle (most temp files are in user temp), as an experiment I also currently on my main PC have my browser profile folder on my second ssd (when I put on spindle the browser was noticeably laggy), I moved my browser temp files to ram disk. Even with all this done I can see from size of volume snapshots and size of incremental backups there is still 100s of megs writes per day to my OS drive.

As long as you have confidence your ssd will be able to function for its rated erase cycles, you will probably be fine, it will outlive its warranty easily.

Hi,
We aren't that far apart
Difference might be central air conditioning and not.


Yep plus if you do rma you'll get a refurbished back not a new one this is crucial's rma policy.

Thanks for this info, I will think twice on the RMA then if that is offered.
 
A little update, nothing from Crucial yet, but I ran the smart self test script for about 3-4 days and the erase cycle hasnt moved, however that did raise the temp of the drive, so I stopped it, I power cycled and have since left it running normally, and will monitor erase cycles for a bit to see how it behaves, the laptop usually is rebooted once a month for windows updates, but occasionally more often for other reasons, I plan to power cycle on every reboot now so at the minimum there is a power cycle every month.
 
I plan to power cycle on every reboot now so at the minimum there is a power cycle every month.
You don't need to force a power down. Just choose "Shut down" on the Start menu.
 
I hooked up a brand new fresh out of the package MX500 and confirmed that the Average Block Erase Count does in fact start at 100. Which confirms that it is a percentage that counts down and OP's drive is still at 100%. There is no point in bugging Crucial support about this, trying a workaround, or returning the drive, the number starts at 100 and goes down from there. Your drive is functioning completely normally.

That then begs the question where does CrystalDiskInfo get it's health status from? One would think it is likely reading Percent Lifetime Used, which from what I can figure is just the Total Host Writes compared to the rated TBW. But those numbers don't add up to the health status CrystalDiskInfo is providing(OP's drive should still be at 98 Percent Life Used but they claim CDI is saying 94% health). So I'm going to go with my theory that I've always had about CDI's health status %. It's just pulled straight out of thin air, it doesn't relate to any actual SMART data and doesn't relate at all to the actual drive's health.

View attachment 218012


By "Data Field" I assume you mean Raw Values field. And to most people that field is totally garbage information because they can't read hexidecimal. The current field is just the raw value translated into a number most people can understand.
Ive never liked that tool
 
Well it stayed on 100 cycles for a while, but yesterday there was 2 cycles in one day, now on 102 cycles. It seems only a few days uptime without a power cycle is enough to set it off again.

Some data from the snapshots I took from smart. Note they not done the same time each day, I just check my laptop when I finish work, and hit save in crystal diskinfo.

24 sept host writes 2044GB cycles 100 idle state prevented by self tests
25 sept host writes 2053GB cycles 100 idle state prevented by self tests
26 sept host writes 2059GB cycles 100 idle state prevented by self tests
26 sept (later) host writes 2062GB cycles 100 power cycled
27 sept host writes 2064GB cycles 100 allowed to idle
28 sept host writes 2071GB cycles 100 allowed to idle
29 sept host writes 2076GB cycles 102 allowed to idle

Ive never liked that tool
Well to answer his questions, he misunderstood the cycle data I provided.

The lifetime remaining on the MX500 matches the erase cycle rating.

The 'current' value for erase cycle count matches the 'current' value for percent life time used. Both are 94 on my MX500.
The raw value for erase cycles is now 102, which he misunderstood as to be the 'current' value.

On my 970 evo which doesnt provide any erase cycle figures, I noticed the predicted life goes down in crystal diskinfo when it starts using spare sectors.
 
This issue only affects the SMI2258 drives (square box) with the M3CR02x firmware, and possibly the SMI2259 with the M3CR03x firmware.

No, it wasn't fixed, even drives obtained with the firmware installed from the factory demonstrate the problem.

Recent drives (in the rectangular box) use the SMI2559 and M3CR043 firmware.

Affected units will see the C5 field occasionally switch to 1 followed by a burst increase of the F8 field, this happens so often on the affected parts that it ends up significantly higher than the F7 field, which for an OS disk should actually always be ahead of F8 on these drives.

Might be a bad batch, RMA it and move on.

It wasn't a bad batch, i have multiple M3CR023 parts bought years apart with differing manufacturing dates with the same issue from different etailers, the recent M3CR043 i have purchased do not demonstrate the issue.
 
Last edited:
I pulled a Toshiba 1TB spinner from my moms dell laptop a couple of years ago. She doesn’t do anything other than look at Facebook and stuff like that. The amount of writes the drive had was mighty impressive. Sleep/hibernate kicks the crap out of drives.
 
I have a replacement in Raid 10 array for one of mine, life left went below 10% on the old one (This is in a data centre where they must be using hundreds of them).

Life left is over 100%... on the new one, Which is strange, I also have Crucial software, it's showing 10 nand blocks reallocated but 36 still available. Only very low amount of actual data written to this drive though.

Bit concerned as don't want it to fail unexpectedly, even if redundancy in the array.


CrystalDiskInfo_20220223074642.png
 
I have a replacement in Raid 10 array for one of mine, life left went below 10% on the old one (This is in a data centre where they must be using hundreds of them).

Life left is over 100%... on the new one, Which is strange, I also have Crucial software, it's showing 10 nand blocks reallocated but 36 still available. Only very low amount of actual data written to this drive though.

Bit concerned as don't want it to fail unexpectedly, even if redundancy in the array.


unknown.png

Theres a fault there, the current and worst is supposed to begin at 100,

unknown.png


Your smart data has exceeded 100% and the value has rolled over from 200, which it shouldn't,
you're actually at 165% life time used
 
Theres a fault there, the current and worst is supposed to begin at 100,

unknown.png


Your smart data has exceeded 100% and the value has rolled over from 200, which it shouldn't,
you're actually at 165% life time used

Yes it's strange , I thought it could have been modified somehow or a bug. It's barely had any data written to it and not high up time hours which makes me question how it can be near endurance.

I'll get it changed out for another one I think anyway.
 
this afternoon i find my SSD suddenly not respond at all, first i think it's from the cable after recheck everything looks normal but i feel something that not right, i remove the ssd and place it on external case then test it on other pc, and this is what i have...... damn
Untitled.png

now i try to back it up to somewhere before it get worse
 
That's the newer firmware version that doesn't have the problem.

It doesn't seem clear to me if the erase cycle issue actually impacts drive life much.

Or if the endurance algo simply can't handle it leading life left to plunge far too quickly when actually the nand will take much more usage than it's suggesting.
 
I updated the firmware myself only recently, the drive had been on the older firmware until now ¯\_(ツ)_/¯
 
That whole lineage is fixed it's M3CR020 in particular that wears ridiculously quickly. M3CR023 has fixed the problem, mainly.

I think M3CR032 had another bug with raid configurations but not the endurance problems.

You can't flash M3CR023 to anything higher though.

M3CR033 should be all good I think.

In conclusion, Crucial need better firmware engineers!
 
Ironic this thread recently got updated, my laptop is down at the moment pending investigation, a bunch of critical window files can no longer be read, so will need to recover from backup and figure out whats going on, dont want to say anything else until I have done this.

Funny Cutechri drives are older than mine but he/she got the newer model. o_O
 
Last edited:
A little off topic; do modern thumb drives wear level?
 
Back
Top