Samsung 840 EVO read speeds found to degrade on old data

[Arthur]

Has anyone here experienced slow read speeds with old data on their 840 EVOs?

 

Sources: Overclock.net Forums / Reddit

 

"Both EVO drives have the same exact problem, read speed on old files drops with time as low as ~60 MB/sec. New written files instead are as fast as advertised read/write speed ~450 MB/sec Please note that the data must be "old" in terms of weeks or months to show the problem. Newly-written data read speed is fast as advertised, this is why benckmarks shows always a fast drive. After a reboot the read speed improves a bit, but not much."

 

"I have now seen several reports of this problem on different sites. I own a 1TB EVO since a few days and will try to verify this problem although it supposedly affects 'old' files. But 'old' in all likelihood doesn't mean that you have to leave your SSD lying around for 1 year and then the read data rate even on big 'sequential' files will have dropped significantly. It might rather be related to writing a lot of files or some infrequent event like a sudden system power loss (reset, blue screen, ...)."

 

"Performance drops to as low as 60 MB/s or so for old files. Sadly first hand experience over here."

 

"I have the 840 EVO 1TB, its performance deteriorates heavily the more you use it and it goes down with every benchmark. I would never use that drive as a main drive I just have it in my system to store videos and mp3s / pics. I treat it as a supercharged HDD nothing more nothing less".

 

[zag]

I've been using ours on our backup server writing 10's of GB a week for 6 months now. I didn't notice anything dramatic yet.

[sonofsanta]

Bit worrying, we've got 150+ of these in clients and all their data will be "old" data as it's just the OS install. Hopefully the lack of new data being written will prevent the issue occuring, but it's less than ideal.

 

Buying the MX100s at the moment anyway. Can't argue the price of that 256GB drive when recommending to staff, either

[Arthur]

I've been using ours on our backup server writing 10's of GB a week for 6 months now. I didn't notice anything dramatic yet.

Does your backup software only keep six months worth of data? If so, you probably won't have this issue.

[Arthur]

Hopefully the lack of new data being written will prevent the issue occurring, but it's less than ideal.

I wonder if over-provisioning the drives would help since it usually increases performance?

 

That said, a firmware update which permanently fixes the issue would be preferable.

[zag]

Does your backup software only keep six months worth of data? If so, you probably won't have this issue.

 

Good point, yes its only 30 days on our VEEAM server.

 

I think these drives use TLC memory so it was always expected that they would slow down.

 

I wouldn't use them as a primary drive, but they seem fine for storage.

[sonofsanta]

$technician says one fix is, counter-intuitively, to defrag the drive. Normally a no-no on SSDs because of the wear, but running a defrag cycle marks all the data as new and restores the performance, reputedly.

[Arthur]

Good news! Samsung are working on a fix.

 

Samsung Acknowledges the SSD 840 EVO Read Performance Bug - Fix Is on the Way « AnandTech

 

I just got off the phone with Samsung and the good news is that they are aware of the problem and have presumably found the source of it. The engineers are now working on an updated firmware to fix the bug and as soon as the fix has been validated, the new firmware will be distributed to end-users. Unfortunately there is no ETA for the fix, but obviously it is in Samsung's best interest to provide it as soon as possible.

 

I do not have any further details about the nature of the bug at this point, but we will be getting more details early next week, so stay tuned. It is a good sign that Samsung acknowledges the bug and that a fix is in the works, but for now I would advise against buying the 840 EVO until there is a resolution for the issue.

[Arthur]

Samsung will be releasing an update in 18 days time.

 

Firmware update to fix the Samsung 840 EVO SSD read performance bug coming on 15th October « AnandTech

 

A week ago Samsung acknowledged the existence of the read performance bug in the SSD 840 EVO and I just received a note that the fixed firmware is in validation process and is expected to be released to the public on October 15th. Unfortunately I don't have any further details about the bug or the fix at this point, or whether the update is coming to the 'vanilla' SSD 840 and OEM models, but I hope to get more details as the public release gets closer, so stay tuned.

[Arthur]

An early look at the EVO's Performance Restoration tool « The Tech Report

 

Fixing the problem is a two-part process managed by Samsung's appropriately named Performance Restoration tool. First, the utility flashes the SSD's firmware with the new EXT0CB6Q revision. Then, it reboots the system and completes the restoration process, which we're told involves "conditioning" the NAND. The entire process is non-destructive, so it shouldn't compromise any data on the drive. It also only needs to be run once. That said, the target drive must have at least 10% free capacity for the tool to work its magic.

 

Given how the bug manifests, we won't know if the EVO is truly fixed until at least a month from now, when freshly written data will be old enough to exhibit slower read performance. Samsung is confident that the problem has been addressed, though, and the company has updated its validation procedures to look for similar issues in future products. For what it's worth, Samsung says the problem is unique to the 840 EVO and not an artifact of that drive's TLC NAND. The older, TLC-based 840 Series is unaffected, and so is the upcoming 850 EVO, which will use the three-bit version of 3D V-NAND.

 

If all goes well with final testing, the 840 EVO Performance Restoration tool will be released this week. There are still some associated caveats, like the updater not working with SSDs in RAID arrays or with older AMD SATA drivers, but those restrictions aren't too different than those that typically come along with new SSD firmware. We'll post a link to the update tool once it's available.

[Arthur]

The new firmware update is available!

 

www.samsung.com/global/business/semiconductor/samsungssd/downloads.html

 

[Arthur]

I updated a 120GB 840 EVO and it only took 15 minutes. Now to do all of the rest!

[sonofsanta]

So we've got 160 of 120GB 840 EVOs on site... is there a way of pushing this out centrally (install with SCCM is fine, it's a startup script command I'd need), and if not, is it really going to be worth running? I'm leaning towards no, until we see the SSDs in the normal course of events (i.e. move to a new machine).

[zag]

So we've got 160 of 120GB 840 EVOs on site... is there a way of pushing this out centrally (install with SCCM is fine, it's a startup script command I'd need), and if not, is it really going to be worth running? I'm leaning towards no, until we see the SSDs in the normal course of events (i.e. move to a new machine).

 

Realistically it is not worth running.

 

You would have to write a hell of a lot of data to get to the stage that some people have go to.

 

If they are simple client machines, you won't get any where near the amount of writes needed to degrade performance in the real world.

[sonofsanta]

Realistically it is not worth running.

 

You would have to write a hell of a lot of data to get to the stage that some people have go to.

 

If they are simple client machines, you won't get any where near the amount of writes needed to degrade performance in the real world.

 

That's what I'm figuring - the only real writes they get are image + installation, then the odd user profile from there on out. There was some implication early on that the issue was with old data, but I was never 100% clear on whether that meant old as in temporal age, or old as in how many writes there have been since.

[Arthur]

You would have to write a hell of a lot of data to get to the stage that some people have go to.

This is how Samsung describes the issue:

 

SSDs usually calibrate changes in the statuses of cells over time via the flash management software algorithm. Due to the error in the software algorithm, the 840 EVO performed read-retry processes aggressively, resulting in a drop in overall read performance. This only occurs if the data was kept in its initial cell without changing, and there are no symptoms of reduced read performance if the data was subsequently migrated from those cells or overwritten.

 

is it really going to be worth running?

Yes (but I'm in no rush to do it).

 

... there's a definite improvement in read performance for older files. Some of the individual files load more than 2X faster after the fix, and one of them experienced a massive 6X speedup. Your mileage may vary, of course, but Samsung's fix seems to work as advertised.

 

After applying the EVO update, Bruno also noticed snappier application load times. He says Windows now boots faster, too, particularly after it reaches the desktop and starts filling out the taskbar. We don't have hard data on how the fix impacts load times, but OS and application files are likely to be older, so they're good candidates for a speed boost. (Source)

 

^ I have noticed a difference in logging on/off times so it is worth doing IMO.

[Arthur]

Realistically it is not worth running.

Wouldn't you class a 1340% increase in average read speed as worth doing?

 

Before (35MB/s)

 

After (504MB/s)

[sonofsanta]

I suppose the answer is "run a benchmark on a random system and find out"

[Arthur]

Some 840 EVOs still vulnerable to read speed slowdowns

 

Last year, users discovered a problem with Samsung's 840 EVO SSD that caused dramatic slowdowns when reading older data. Samsung attributed the issue to an algorithmic error in the management routine that tracks the status of cells over time. A firmware fix and accompanying Performance Restoration utility were issued in October, and they seemed to do the trick. However, new evidence suggests that the problem persists.

[Arthur]

This firmware update is for Surface Pro 3 owners with Samsung PM851 SSDs.

 

SSD Performance Restoration Software for Microsoft Surface Pro 3

 

Samsung is releasing a new firmware and performance restoration software package for Samsung SSD for Surface Pro 3 users.

 

The firmware update is recommended to users who may have experienced any drop in read performance.

 

Running the software packages provided will first update the firmware (EXT42M0Q), then scan and calibrate all existing data on the drive.

[Arthur]

Samsung investigating 840 EVO slowdowns, prepping another fix

 

Last month, we reported that some Samsung 840 EVO SSDs are still experiencing slow read speeds with old data. The issue was supposedly fixed with a firmware update issued in October, but some patched drives continue to exhibit the problem, including one of our own. Late Friday afternoon, Samsung issued the following statement on the matter:

 

In October, Samsung released a tool to address a slowdown in 840 EVO Sequential Read speeds reported by a small number of users after not using their drive for an extended period of time. This tool effectively and immediately returned the drive’s performance to normal levels. We understand that some users are experiencing the slowdown again. While we continue to look into the issue, Samsung will release an updated version of the Samsung SSD Magician software in March that will include a performance restoration tool.

 

The drive we tested last month exhibits particularly extreme symptoms, and it's been sent to Samsung for further testing. We'll keep following this story as it develops.

[mac_shinobi]

Duplicate of Arthurs post. Deleted Edited February 23, 2015 by mac_shinobi

[Arthur]

not sure if this is a new article

It's basically saying the exact same thing as the post directly above yours.

[Arthur]

New fix for 840 EVO slowdowns due 14 April

 

When we last checked in on the ongoing saga related to read-speed slowdowns on some 840 EVO SSDs, Samsung pledged to deliver a new version of its Magician software and performance restoration tool in March. The 31st passed with no update in sight, so we asked the company for a status update. We're now told the due date is 14 April 2015.

 

Samsung says it's prepping new EVO firmware in addition to the Magician refresh. Hopefully, the fresh firmware tweaks do a better job of addressing the underlying problem than the initial fix from October. The read-speed slowdowns take time to manifest, so we won't know for sure until weeks or months after the update is issued.

 

One of our affected drives was sent to Korea to validate the new fix. We have another one in our labs, and we'll be loading it with a fresh batch of data as soon as the update is available. Stay tuned.

[Arthur]

New Samsung SSD 840 EVO Read Performance Fix Coming Later This Month

 

The Samsung SSD 840 EVO read performance bug has been on the table for over six months now. Initially Samsung acknowledged the issue fairly quickly and provided a fix only a month after the news hit the mainstream tech media, but reports of read performance degradation surfaced again a few weeks after the fix had been released, making it clear that the first fix didn't solve the issue for all users. Two months ago Samsung announced that a new fix is in the works and last week Samsung sent us the new firmware along with Magician 4.6 for testing, which will be available to the public later this month.

 

I covered the reason behind the issue in one of our earlier articles, but in short the read performance degradation is a result of cell charge decay over time that caused extensive read-retry cycles to retrieve the correct data. The new firmware fixes this by periodically refreshing (i.e. rewriting) old data, which recovers the cell charge back to its original state and ensures that no read-retry or ECC that would degrade the performance is needed. Samsung says that the refresh operation does not impact user performance, suggesting that it's a relatively low priority process that is run when the drive is idling.

 

The new Magician 4.6 also includes an Advanced Performance Optimization feature, which is similar to the performance restoration tool that Samsung released earlier. Basically, it's a command that tells the SSD to rewrite all of its internal data, which resets all cell charges and hence recovers performance. It's merely a supplementary tool as the firmware upgrade itself should be enough to restore performance, but in case the performance isn't fully restored after the firmware upgrade (and some idle time to let the drive refresh the cells), the tool can be used to force a cell charge refresh.