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 enterprise SSD heritage of the Intel SSD 750 continues to shine through as it holds on to the lead for steady-state random write performance, but Samsung has mostly caught up with the 960 Pro. This is a huge change from the 950 Pro, which had steady-state performance that was no better than typical SATA SSDs. A few consumer SSDs have offered great steady-state random write performance—most notably OCZ's drives based on the Indilinx Barefoot 3 controller—but the 960 Pro is the first one to reach the level of the Intel SSD 750.

Steady-State 4KB Random Write Consistency

In addition to mostly closing the performance gap, the 960 Pro has a great consistency score that is almost as good as the Intel SSD 750's score. While OCZ's Vector 180 offered remarkably high average performance in its steady state, it was far less consistent than the either the Samsung 960 Pro or the Intel SSD 750 and instead the standard deviation of its steady state performance was more than ten times greater.

IOPS over time
Default
25% Over-Provisioning

After the initial period of very high performance, the 960 Pro enters a steady state with very good short-term consistency but gradual long-term variation in performance. This is more similar in character to the behavior of the Intel SSD 750 than Samsung's earlier SSDs, though it's interesting to note that the 960 Pro is more twice as fast during the initial phase before transitioning to steady state.

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

Focusing on the last 400 seconds of the test shows the 960 Pro's steady state to be essentially flawless, rounding out a full page of what can be considered to be perfect scores for a consumer drive. The performance would even make the 960 Pro a pretty good enterprise SSD, and this is usually not the case for drives with consumer-oriented firmware.

A Note About Drivers AnandTech Storage Bench - The Destroyer
Comments Locked

72 Comments

View All Comments

  • leexgx - Tuesday, October 25, 2016 - link

    the problem with windows 10 when using as a benchmark system is you got to make sure automatic maintenance is disabled and windows update is disabled or it mess the results up (i have 2 laptops and both of them go nuts when screen turns off on win10{fan revved up and lots of SSD activity)

    i would personally stick with windows 7 or 8 as they are more predictable

    if using windows 8 and 10 you need to disable the idle maintenance auto task (set windows update to never check) and windows 10 you have to disable the windows update service as it can mess up benchmark results (or if using windows 10 pro use GPedit to set windows update to ask before downloading, note pressing check or download actually means download and install on windows 10 pro)
  • Badelhas - Tuesday, October 18, 2016 - link

    If I replace my Vertex 3 120Gb Sata3 SSD with this one and use my PC for normal tasks like web browsing and gaming, will I notice any difference? Thats the real question to me.

    Cheers
  • DanNeely - Tuesday, October 18, 2016 - link

    The biggest one will be being able to have all yours games on SSD instead of just 1 or 2. Even a cheap SSD is fast enough that IO rarely is a major bottleneck in day to day consumer use.
  • phobos512 - Tuesday, October 18, 2016 - link

    For the money you will spend, you will not notice a significant difference. If the rest of your system is of the same vintage as the SSD you're replacing, that will be even more true.
  • phobos512 - Tuesday, October 18, 2016 - link

    And here's the evidence.

    https://cdn.arstechnica.net/wp-content/uploads/sit...

    https://cdn.arstechnica.net/wp-content/uploads/sit...

    https://cdn.arstechnica.net/wp-content/uploads/sit...
  • Amoro - Tuesday, October 18, 2016 - link

    There's a typo in the form factor for 960 drives, "Sngle-sided". Also, if the form factor is the same for both drives shouldn't the cell be merged?

    Does this make the 950 Pro obsolete at this point too? At least for the 512GB version.
  • Billy Tallis - Tuesday, October 18, 2016 - link

    Thanks. I fixed the typo, but left the two cells separate and split the PCIe interface so that there's an uninterrupted vertical line separating the old drives from the new.

    Once the 512GB 960 Pro is widely available and once Samsung delivers the drivers for it, there should be no reason to get the 512GB 950 Pro. I do hope to confirm that directly by testing a 512GB 960 Pro against the 950 Pro, but sample supplies have been pretty limited for this launch. The 256GB 950 Pro won't have a direct successor, but if the 960 EVO does what it's supposed to it should offer better real-world performance at a much lower price.
  • TheinsanegamerN - Tuesday, October 18, 2016 - link

    I'd say price would be a big one. If you can get the 950 pro for $100 less then the 960 pro of the same size, unless you need all that speed the 950 pro would be a better deal.
  • Swede(n) - Tuesday, October 18, 2016 - link

    How was the 960 Pro connected during the test?
    On the Asus Z97 mobos M.2 connector that shares bandwidth Sata Express #1?

    If so, is it recommended to unplug any other Sata drive from this Sata port #1 and use a separate Sata port for that device (for not loosing performance under heavy workload where multiple SSD-drives are in use?

    Or was the 960 Pro connected to a PCIE 3.0 via adapter?
    Please explain this and the possible benefits for one or the other, consider a hefty game GPU connected to PCIE 3.0 x16 slot on a similar mobo (Asus Z97 Deluxe).
    Sincerley from Sweden
  • Billy Tallis - Tuesday, October 18, 2016 - link

    The SSD testbed doesn't have a discrete GPU, so all PCIe SSDs are tested in the PCIe 3.0 x16 slot. There's a riser card with the power measurement circuitry between the SSD and the motherboard. M.2 PCIe SSDs are tested in a simple passive PCIe x4 to M.2 adapter card which is plugged in to the power measurement riser card. I'll also be testing the 960 Pro with the Angelbird Wings PX1 adapter and heatsink as I dig deeper into its thermal performance.

Log in

Don't have an account? Sign up now