Performance Consistency

Performance consistency tells us a lot about the architecture of these SSDs and how they handle internal fragmentation. The reason we do not have consistent IO latency with SSDs is because inevitably all controllers have to do some amount of defragmentation or garbage collection in order to continue operating at high speeds. When and how an SSD decides to run its defrag or cleanup routines directly impacts the user experience as inconsistent performance results in application slowdowns.

To test IO consistency, we fill a secure erased SSD with sequential data to ensure that all user accessible LBAs (Logical Block Addresses) have data associated with them. Next we kick off a 4KB random write workload across all LBAs at a queue depth of 32 using incompressible data. The test is run for just over half an hour and we record instantaneous IOPS every second.

We are also testing drives with added over-provisioning by limiting the LBA range. This gives us a look into the drive’s behavior with varying levels of empty space, which is frankly a more realistic approach for client workloads.

Each of the three graphs has its own purpose. The first one is of the whole duration of the test in log scale. The second and third one zoom into the beginning of steady-state operation (t=1400s) but on different scales: the second one uses log scale for easy comparison whereas the third one uses linear scale for better visualization of differences between drives. Click the dropdown selections below each graph to switch the source data.

For more detailed description of the test and why performance consistency matters, read our original Intel SSD DC S3700 article.

Samsung SSD 850 EVO 250GB
Default
25% Over-Provisioning

The 850 EVO presents a healthy increase in IO consistency. The 840 EVO wasn't exactly inconsistent in the first place, but the 850 EVO takes the steady-state IOPS from ~3,000-5,000 IOPS to 5,000-8,000 IOPS, which is actually nearly on par with the 850 Pro. The 850 EVO has without a doubt one of the highest performance consistencies out of the value/mainstream drives we have tested.

Samsung SSD 850 EVO 250GB
Default
25% Over-Provisioning

 

Samsung SSD 850 EVO 250GB
Default
25% Over-Provisioning


TRIM Validation

To test TRIM, I filled a 120GB 850 EVO with sequential 128KB data and proceeded with a 30-minute random 4KB write (QD32) workload to put the drive into steady-state. After that I TRIM'ed the drive by issuing a quick format in Windows and ran HD Tach to produce the graph below.

And TRIM works as expected.

Endurance: Close to Planar MLC NAND AnandTech Storage Bench 2013
POST A COMMENT

99 Comments

View All Comments

  • TheWrongChristian - Monday, December 08, 2014 - link

    Performance consistency wise, the MEX controller on the 1TB drive looks better than the MGX controller on the smaller drives. I guess this is the loss of a cpu core at work, but I figure there'll be no discernible difference to the user experience.

    All in all, looks like a nice drive, a reasonable upgrade to the existing 840 evo drives. Just hope the V-NAND cost brings the whole price down to compelling levels as the process matures.
    Reply
  • Solid State Brain - Monday, December 08, 2014 - link

    As for the Wear Leveling Count, it might actually be taking into account that a fixed portion of the installed NAND is used in SLC mode for the TurboWrite buffer. The 120GB model has 128GiB of NAND, of which 9 GiB are used for 3 GiB TurboWrite Buffer, so that makes 119 GiB of TLC capacity for both overprovisioning and user addressable space.

    By the way, this also implies that because of TurboWrite these Samsung EVO SSDs (including the previous 840 EVO) have less overprovisioning space than most other SSDs.
    Reply
  • Solid State Brain - Monday, December 08, 2014 - link

    Yikes! This was not meant to be in response to TheWrongChristian. Reply
  • hojnikb - Monday, December 08, 2014 - link

    >What I do know is that Samsung started the mass production of TLC V-NAND later, which suggests that the two aren't completely uniform. Moreover, from what I know TLC NAND requires some changes to the peripheral circuitry in order to read three bits from one cell, so while the NAND memory arrays could be alike the die size is still likely at least slightly different.

    Is it possible, that samsung designed 2nd gen 3D with TLC in mind (eg requred peripheral circuitry) and simply set the controller in 850PRO to use 4 states instead of 8 (so MLC). I mean, it kinda makes sense to go this way, but not the other way around....
    Reply
  • Solid State Brain - Monday, December 08, 2014 - link

    @Kristian Vättö
    Do you have any information from Samsung about whether the TurboWrite SLC buffer also helps decreasing the write amplification like on the SanDisk Ultra II with nCache 2.0?
    Reply
  • Kristian Vättö - Tuesday, December 09, 2014 - link

    Unfortunately I don't. It was among the questions I sent but unfortunately Samsung couldn't get any of my questions answered on time for the review. Reply
  • Solid State Brain - Wednesday, December 10, 2014 - link

    Too bad that you couldn't get an answer in time. According to my observations from other people's drives, it really looks like the Turbowrite does help on that regard. Try checking out my thread on the Memory and Storage forum.

    This phenomenon (write amplification just over or below 1.0x) is likely not going to show up during reviews or heavy usage since generally the drives get secure erased often and/or get hammered with writes which end up filling the Turbowrite SLC buffer.
    Reply
  • Daniel Egger - Monday, December 08, 2014 - link

    It would be nice to have some older drives in the charts to get some perspective on whether a drive would make a good upgrade choice. For an upgrade the available space and the space/price ratio are probably the most important aspects but some features or a major speed increase might sweeten the deal even more. ;) Reply
  • sheh - Monday, December 08, 2014 - link

    Not perfect of always complete, but usable:
    http://www.anandtech.com/bench/SSD/65
    Reply
  • Kvaern - Monday, December 08, 2014 - link

    It's my anecdotal experience that all these awesome benchmark figures means absolutely nothing for the average user.
    Case in point I upgraded from an old 60GB corsair drive to an EVO840 which on paper is like twice as fast as the Corsair but in reality my user experience is exactly the same.
    Reply

Log in

Don't have an account? Sign up now