Interesting. I read through those posts but nobody mentioned whether restoring a backup image fixed the problem. (temporarily)
I would think that a secure erase, followed by restoring a backup/reinstalling the OS would fix it. Actually, I've been doing that for a while now. I thought this was a commonly known issue with all SSDs, although the specifics of this may be a bit different.
It's a fundamental problem with TLC. The industry needs to stop shrinking the process and back up a step.
I think this is specific to the 840 models, it is a firmware bug - http://www.anandtech.com/show/8550/...40-evo-read-performance-bug-fix-is-on-the-way
Not read about TLC slowing with age, do you have more info ? Also, better firmware does account for degradation, than end up in states of error.
Not really more info. But that's what happens with TLC. The cells' charges flatten out over time. And the firmware needs to apply a different algorithm to discern the levels and effect more frequent refreshes. There is simply too much overhead in error checking and correcting required in TLC designs. Just get the data stored right and cleanly and you have less to worry about later. TLC planar is simply a result of the industry pushing their processes too fast on road to higher densities. Perhaps when this V-NAND gets underway we'll once again see reliable SSD.
The German site ComputerBase is telling that Samsung will release firmware-update on 15 October. In German: http://www.computerbase.de/2014-09/840-evo-ssd-schwaechelt-bei-alten-daten-loesung-in-arbeit/
http://www.samsung.com/global/business/semiconductor/minisite/SSD/us/html/support/downloads.html http://techreport.com/review/27212/samsung-840-evo-update-fixes-slow-reads-with-old-data
Have just ran this on my Samsung 840 Evo 250gb drive, the firmware update goes quickly & then says the PC will either reboot & the optimisation will continue when PC has rebooted but mine's shutdown, powered back on then the optimisation runs & took about a hour.
I also just ran the Performance Restoration Software with similar results, although the optimization took considerably less time. Regardless, I was a bit surprised that this software required installation as I was under the impression that it would be a "run once" sort of deal. I decided to uninstall the s/w, but am left wondering whether it needs to be run on some sort of regular basis. Of course, the firmware update is a one-and-done, but what about the rest of the program? Dumb question perhaps, but does the performance component of the package ever need to be run after it has been run initially?
As far as I know the performance bit only has to be run once, once it had completed I rebooted & then uninstalled it & no issues as yet.
does the optimization program need to be run after a clean install of windows ? eg. run after new software is installed or is it automatically enabled as default onboard the drive as part of the FW update and all newly written data is optimized automatically ?. also does the FW update wipe any existing data as part of the update ? Gazzer
If the SSD is a system drive with Windows on it then I would make a backup first to be on the safe side, my evo is just used for games so didn't do a backup & all my data is still there & the games are working fine. You also need at least 10% free space as a minimum for it to do the process. I'm not 100% sure but I think after it updates the firmware & the PC is restarted it then continues with the optimisation it moves the data so it can do what it needs to do the NAND cells.
Has anyone tried the fix with an encrypted SSD (LUKS)? I am hesitating because I read that performance restoration won't work „if Encrypted Drive standards are in place.“ http://www.extremetech.com/computin...but-be-careful-of-the-caveats-when-you-update
Samsung Releases Statement on 840 EVO Performance - Another Fix Is In the Works February 20, 2015 http://www.anandtech.com/show/8997/...0-evo-performance-another-fix-is-in-the-works
My 840 EVO read terrible, but was performing good again after the first fix, now a month later it is performing compared to a '98 harddrive. I suspect this next "fix" is nothing more than another bandage, in order to not have to acknowledge that it's just a (cheap) hardware decision.
Any drive with TLC (and even some MLC) are cost cutting unreliable designs. The density is simply too high.