SP3 SSD, possible firmware bug, same as just fixed in another Samsung drive.

Anonymous
2014-10-20T22:49:51+00:00

I've just had to update the firmware in another device that had a Samsung 840EVO SSD because they had a serious firmware bug with it.  Out of curiosity I tested the Samsung drive in my Surface Pro 3 and it looks like it has exactly the same issue (the SP3 drive is closely related to the 840 series I believe, but an OEM version.

The test is simple, just look at the read rate across the drive.  Older data (OS, installed apps) suffered from terrible read performance after a month or so.  Newly written data is fine.

As you can see, much of the drive is down to 50MB/s now.  It should be a fairly level line right the way across at about 350+MB/s.  This test was run with AC power connected and nothing else running.

If this is the same issue as the 840EVO, the good news is it can be corrected with a firmware update and refresh of the data.  Bad news would of course be Microsoft need to look into it first, then get the Samsung engineers looking into it.

Perhaps some other people could run the test too, and post results.  Remember the drive needs to be a month or more old, so that the data can have been sitting there only being read during that time.

Further details of the bug and the fix that Samsung released can be found here:

http://www.anandtech.com/show/8617/samsung-releases-firmware-update-to-fix-the-ssd-840-evo-read-performance-bug

Surface | Surface Pro | Install and update

Locked Question. This question was migrated from the Microsoft Support Community. You can vote on whether it's helpful, but you can't add comments or replies or follow the question.

0 comments
Answer accepted by question author
Anonymous
2015-09-30T22:01:46+00:00

Hello all

We wanted to let you know that an update to resolve this is still being worked on.  Please continue to post your experiences and technical details.

Mod Edit: Samsung SSD Optimization Utility for Surface Pro 3 now available.

Best regards

Was this answer helpful?

0 comments

151 additional answers

Sort by: Oldest
Most helpful Newest Oldest
  1. Anonymous
    2014-10-28T11:24:01+00:00

    Hi Frank,

    I'd have to say, your results look awful as well!

    Yes I ran it read only (the free 2.55 HDTune only does read tests afaik).  If you check out the AnandTech article I linked to in the first post I think it does a good job of explaining the issue and why it is all about read performance of OLD data.  Benchmarks that write data and then read it straight back (such as CrystalDiskMark) would only see good results.

    I did a quick search for examples of what the graph should be looking like (as in flatter and higher) and there are some here:

    http://www.cdrlabs.com/Reviews/samsung-ssd-840-pro-series-256gb-solid-state-drive/Performance-AS-SSD-and-HD-Tune.html

    Something seems very wrong here. The performance of these drives should be blazing fast across all of the data and it is nowhere near that.  Any ideas how we can get someone from the Surface engineering team to dig into this and check?

    Was this answer helpful?

    0 comments
  2. Anonymous
    2014-10-28T16:43:12+00:00

    Hi Frank,

    I'd have to say, your results look awful as well!

    Yes I ran it read only (the free 2.55 HDTune only does read tests afaik).  If you check out the AnandTech article I linked to in the first post I think it does a good job of explaining the issue and why it is all about read performance of OLD data.  Benchmarks that write data and then read it straight back (such as CrystalDiskMark) would only see good results.

    I did a quick search for examples of what the graph should be looking like (as in flatter and higher) and there are some here:

    http://www.cdrlabs.com/Reviews/samsung-ssd-840-pro-series-256gb-solid-state-drive/Performance-AS-SSD-and-HD-Tune.html

    Something seems very wrong here. The performance of these drives should be blazing fast across all of the data and it is nowhere near that.  Any ideas how we can get someone from the Surface engineering team to dig into this and check?

    But the diskfresh did you run it using the checkbox read only?

    ms is monitoring this forum 😉

    Was this answer helpful?

    0 comments
  3. Anonymous
    2014-10-28T17:02:03+00:00

    But the diskfresh did you run it using the checkbox read only?

    Ah, DiskFresh I specifically ran with writing enabled, as I wanted it to read and then write back the data to 'refresh' the flash memory cells (yes I know that's not good for SSD's in general, but just as a one off test, I wouldn't do it regularly, hopefully never again).  The performance tests had already been reading data across the whole drive.

    Was this answer helpful?

    0 comments
  4. Anonymous
    2014-10-29T00:15:51+00:00

    But the diskfresh did you run it using the checkbox read only?

    Ah, DiskFresh I specifically ran with writing enabled, as I wanted it to read and then write back the data to 'refresh' the flash memory cells (yes I know that's not good for SSD's in general, but just as a one off test, I wouldn't do it regularly, hopefully never again).  The performance tests had already been reading data across the whole drive.

    I went and ran the refresh and Wow! What a difference!

    Was this answer helpful?

    0 comments