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

Among TLC drives, the ADATA SP550 and OCZ Trion 150 both soundly outperform the PNY CS1311 in sustained random writes. The PNY CS2211 performs well enough for a cheap MLC drive, but is also significantly behind the earlier Phison-based Neutron XT. The gap between the CS2211 and the CS1311 is clear, but there are other MLC and TLC drives blurring that line.

Steady-State 4KB Random Write Consistency

The CS1311's consistency score is decent despite the low overall random write performance. The CS2211 has a very low consistency score, indicating that there's a major difference in firmware between the two models. We would normally expect ongoing garbage collection to be easier for a MLC drive, especially one with as much spare area as a typical TLC drive.

IOPS over time
Default
25% Over-Provisioning

The difference in behavior between the the CS1311 and CS2211 is very clear. The TLC-based CS1311 has the tight regulation of performance we saw on the OCZ Trion 150 while the MLC-based CS2211 behaves like earlier Phison-based drives by having a solid baseline with widely varying performance above that.

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

The 240GB CS1311 and 240GB Trion 150 alternate between the same two performance states, but the Trion 150 spends a far larger portion of its time in the higher performance state. At the 480GB capacity, the Trion 150's slower performance state is the CS1311's faster state.

The CS1311 responds better to overprovisioning than the CS2211, leading to the 480GB CS1311 providing a steady 10k IOPS while the CS2211 varies from 3k to 27k IOPS.

Introduction AnandTech Storage Bench - The Destroyer
POST A COMMENT

44 Comments

View All Comments

  • alexdi - Friday, April 15, 2016 - link

    I read this with only one question in mind: does it beat the 850 Evo? Save for a few ticks in power usage, apparently not. The Evo is perpetually on sale. I've yet to see a compelling reason to opt for anything else on a desktop. Reply
  • ingwe - Friday, April 15, 2016 - link

    Completely agree. I am not tied to it, but I don't see any reason to recommend pretty much anything else. Reply
  • fierywater - Friday, April 15, 2016 - link

    The CS1311 gets marked down from time to time; I picked up my 480GB one for $100 while the Evo 850 500GB usually doesn't get marked down below $130 (and it was $150 everywhere when I picked up the CS1311). It's plenty fast for real world use, especially as a drop-in replacement for an HDD. I think there's a place for drives like it, although that applies less to the CS2211. Reply
  • lilmoe - Friday, April 15, 2016 - link

    For the performance and value you're getting with the EVO, $30 is well worth the extra, and hardly an amount worth saving going for the CS1311, or any other TLC drives at that. In the matter of fact, lots of the current consumer MLC drives don't compare to the value, performance, or in some cases, the endurance and features you're getting with the EVO. Reply
  • ATC9001 - Monday, April 18, 2016 - link

    I agree from most "prosumers" which are frequent to read this article, but for the mainstream user, I don't think its worth it (spending the $30+). Any SSD is better than a HDD, but some garbage bargain bin SSDs aren't worth the cheap price (this being the first exception). I know most people (including myself) think the same thing alexdi posted when reading this....is this going to beat the evo? It doesn't, but at the same time it's not far off from it, and $30 bucks can cause it to break a price plane for some mid range users.

    Each person has there own utility curve or price performance idea, and for me, this is the first drive since the 840 EVO was released in which I would say it's not worth the x dollars to just get the evo!
    Reply
  • Stuka87 - Friday, April 15, 2016 - link

    Having a quality, reliable drive is not reason enough for you? If you want to buy a drive that has a much shorter life span, go for it. But Anybody that cares about data, is not going to by a TLC drive over a MLC drive. Reply
  • lilmoe - Friday, April 15, 2016 - link

    FYI, TLC VNAND has better endurance than most 15nm MLC drives... Reply
  • Impulses - Friday, April 15, 2016 - link

    Define much shorter...

    You're talking in broad strokes about a bunch of different things in the same breath. Life span and reliability aren't necessarily the same thing, unless you need drives to be reliable for 15yrs...

    15yrs ago I was wondering if I'd ever fill my 75GB Deathstar, I'm not sure I'd even keep a drive 5+ years. My 2x 850 EVO have been nothing but reliable since I bought them last year.
    Reply
  • DanoSpumoni - Friday, April 15, 2016 - link

    Same here. the M.2 850 EVO 500GB is my go to SSD right now. They are in all my computers either in M.2 slots or M.2-> SATA sleds. The performance and reliability is hard to beat. I only buy M.2 SSDs now for future compatibility because they last so long I know they'll outlive the computers they inhabit right now. When the 1TB version drops to ~$150 I'll grab some more... Reply
  • Impulses - Friday, April 15, 2016 - link

    I'm sitting on 2x 1TB, bought at like $310 & $330 IIRC, seen them for $260 lately... Waiting for the 2TB to drop lower and I might add one of those. :D Reply

Log in

Don't have an account? Sign up now