Performance Consistency

Our performance consistency test explores the extent to which a drive can reliably sustain performance during a long-duration random write test. Specifications for consumer drives typically list peak performance numbers only attainable in ideal conditions. The performance in a worst-case scenario can be drastically different as over the course of a long test drives can run out of spare area, have to start performing garbage collection, and sometimes even reach power or thermal limits.

In addition to an overall decline in performance, a long test can show patterns in how performance varies on shorter timescales. Some drives will exhibit very little variance in performance from second to second, while others will show massive drops in performance during each garbage collection cycle but otherwise maintain good performance, and others show constantly wide variance. If a drive periodically slows to hard drive levels of performance, it may feel slow to use even if its overall average performance is very high.

To maximally stress the drive's controller and force it to perform garbage collection and wear leveling, this test conducts 4kB random writes with a queue depth of 32. The drive is filled before the start of the test, and the test duration is one hour. Any spare area will be exhausted early in the test and by the end of the hour even the largest drives with the most overprovisioning will have reached a steady state. We use the last 400 seconds of the test to score the drive both on steady-state average writes per second and on its performance divided by the standard deviation.

Steady-State 4KB Random Write Performance

The 512GB ADATA SU800 delivers lower steady-state random write performance than the 525GB Crucial MX300 despite the latter having less built-in overprovisioning. The 256GB SU800 isn't the slowest in its class, but the 128GB is in last place with barely half the performance of the 120GB Samsung 750 EVO. The Silicon Motion engineering sample with more overprovisioning outperformed the SU800 with twice the capacity.

Steady-State 4KB Random Write Consistency

The performance consistency scores from the SU800 are typical for a low-end SSD, but there's clearly room for improvement. Samsung's 750 EVO scores far higher than any of the other budget SSDs, and the Crucial MX300 has the next highest score.

IOPS over time
Default
25% Over-Provisioning

For the first few seconds of the test, the SU800 delivers close to the advertised IOPS, then it drops into a fairly typical pattern of widely variable performance above 5k IOPS as it burns through the spare area. Once the SLC cache and spare area are full, the SU800 spends most of its time slowed down by garbage collection, with regular short bursts of higher performance.

Steady-State IOPS over time
Default
25% Over-Provisioning

The garbage collection cycles on the SU800 appear to last about a minute each, separated by several seconds of faster performance. Overall performance and variability are higher for the larger capacities.

Introduction AnandTech Storage Bench - The Destroyer
Comments Locked

35 Comments

View All Comments

  • RamIt - Wednesday, February 1, 2017 - link

    Why is there no comparison to the Intel 600p recently reviewed?
  • extide - Thursday, February 2, 2017 - link

    Probably because it is a PCIe drive. A shopper who is shopping for SATA drives is probably not interested in a PCIe drive.
  • dj_aris - Thursday, February 2, 2017 - link

    Sorry, but since all SATA 3 drives top at theoretical 600MBs but SSD speeds are WAY faster than that, what is the point in testing them?
  • Death666Angel - Thursday, February 2, 2017 - link

    It might just be my eyes, but I see a lot of instances where the drives don't actually achieve the 600MB/s data cap of the SATA3 protocol....
  • dj_aris - Thursday, February 2, 2017 - link

    Indeed, but who cares if it's 600 or 500 or even 300, when you can reach 2000+ using modern pcie storage (with similiar prices). SATA is dead, obviously will not be developed anymore, SATA express drives don't even exist. I mean we don't test DVD drives or floppies, right?
  • doggface - Thursday, February 2, 2017 - link

    Sata is dead? Seriously?
    Also, similar prices?

    So many things wrong with this post.

    Sata SSDs will continue to thrive for years to come. No doubt about that, and while they do, we will need reviews on their performance to help differentiate the good from bad. Especially as the industry continues to peddle TLC and even possibly QLC drives.
  • vladx - Thursday, February 2, 2017 - link

    Upcoming QLC NAND is not something to scowl at. SSD drives using this technology will not be targeted to compete with MLC or even TLC instead they are planned for archival purposes: https://www.pcper.com/news/Storage/FMS-2015-Toshib...
  • Billy Tallis - Thursday, February 2, 2017 - link

    SATA isn't dead, but it is a dead end in the same sense that VGA ports are.
  • extide - Thursday, February 2, 2017 - link

    Not everyone's system is compatible with and/or cannot boot from NVME drives. There is definitely still a need to have these data points.
  • lopri - Thursday, February 2, 2017 - link

    Two of my boards refuse to accept any M.2 drive as a boot drive, leaving me no choice but to go with SATA. Plus SATA is nowhere near dead because it is still much more affordable. Plus there is no convincing argument as to why backups have to be saved in SSDs.

Log in

Don't have an account? Sign up now