Performance Consistency

We've been looking at performance consistency since the Intel SSD DC S3700 review in late 2012 and it has become one of the cornerstones of our SSD reviews. Back in the days many SSD vendors were only focusing on high peak performance, which unfortunately came at the cost of sustained performance. In other words, the drives would push high IOPS in certain synthetic scenarios to provide nice marketing numbers, but as soon as you pushed the drive for more than a few minutes you could easily run into hiccups caused by poor performance consistency. 

Once we started exploring IO consistency, nearly all SSD manufacturers made a move to improve consistency and for the 2015 suite, I haven't made any significant changes to the methodology we use to test IO consistency. The biggest change is the move from VDBench to Iometer 1.1.0 as the benchmarking software and I've also extended the test from 2000 seconds to a full hour to ensure that all drives hit steady-state during the test.

For better readability, I now provide bar graphs with the first one being an average IOPS of the last 400 seconds and the second graph displaying the IOPS divided by standard deviation during the same period. Average IOPS provides a quick look into overall performance, but it can easily hide bad consistency, so looking at standard deviation is necessary for a complete look into consistency.

I'm still providing the same scatter graphs too, of course. However, I decided to dump the logarithmic graphs and go linear-only since logarithmic graphs aren't as accurate and can be hard to interpret for those who aren't familiar with them. I provide two graphs: one that includes the whole duration of the test and another that focuses on the last 400 seconds of the test to get a better scope into steady-state performance.

Steady-State 4KB Random Write Performance

The SMI 2256 manages to pull off decent average IOPS under a sustained random IO workload. It can't challenge the 850 EVO that uses faster 3D V-NAND, but compared to the BX100 with SM2246EN and 16nm MLC the drop in performance isn't massive -- better yet the SMI 2256 is quite a bit faster than SanDisk's TLC drive Ultra II. 

Steady-State 4KB Random Write Consistency

Unfortunately, the performance isn't very consistent, though, but then again the SM2246EN isn't either as the BX100 is only marginally better.

SMI2256 500GB
Default
25% Over-Provisioning

The steady-state behavior of the SM2256 appears to be similar to its predecessor SM2246EN. The baseline performance is fairly low at roughly 2,000 IOPS, but bursts occur frequently and go all the way to up to 25K IOPS, although this only lasts for about a second. 

SMI2256 500GB
Default
25% Over-Provisioning
Introduction, The Drive & The Test AnandTech Storage Bench - The Destroyer
POST A COMMENT

35 Comments

View All Comments

  • anactoraaron - Wednesday, June 17, 2015 - link

    OFF TOPIC: Wow you guys must have had some serious issues with the LG G4 for it to be 2+ weeks past release and still no review.

    I'm guessing you were about to publish the issues you have had with the device (missed taps/unresponsive screen, slow charging/heat, lag/stutter, etc) and basically not recommend anyone purchase it and you were 'advised' by LG to not publish this until they have had a chance to fix those issues, which according to Android Central will be in the next 3-4 weeks. I just can't recall a review of a flagship device that wasn't out within 2 weeks of release (unless it was a Sony device since they won't comp anything to you guys to review).

    I don't see the point of waiting, because even with the issues I personally have (unresponsive screen/lag/stutter) the G4 is still a solid device.

    This isn't a knock on you guys - you do probably the most unbiased and thorough reviews. The delay on this review is just starting to smell a bit to me.
    Reply
  • Death666Angel - Wednesday, June 17, 2015 - link

    "I just can't recall a review of a flagship device that wasn't out within 2 weeks of release"
    Weren't most Note phones, HTC Nexus 9 and Nexus 6 pretty late or only in preview format released? Anandtech is notorious for this stuff when it comes to anything but Apple. :D Though more often than not, eventually we get something.
    Reply
  • Michael Bay - Thursday, June 18, 2015 - link

    Just two weeks?
    I`m still waiting for 960 review.
    Reply
  • Oxford Guy - Thursday, June 18, 2015 - link

    "TLC for everyone?" Maybe if the drives are being given away. The last time I checked MLC isn't that expensive. Reply
  • Oxford Guy - Thursday, June 18, 2015 - link

    Samsung has had "some issues" with TLC, eh? Why not specifically state what those are?

    "Like all TLC SSD designs we have seen, Silicon Motion employs SLC caching in the SM2256 to improve performance and endurance. The size of the SLC cache is configurable by the OEM, but generally the cache size is between 3GB and 12GB depending on the capacity of the drive."

    Actual SLC or some sort of simulated SLC?
    Reply
  • canthearu - Thursday, June 18, 2015 - link

    "Actual SLC or some sort of simulated SLC?"

    Both the same really. As long as it is dedicated space that will never be asked to perform TLC duty.
    Reply
  • Oxford Guy - Saturday, June 20, 2015 - link

    Isn't it rather impossible for triple layer NAND to be single layer NAND? Reply
  • DeepStorage - Sunday, June 21, 2015 - link

    Most flash foundries have at least some models of their MLC, and TLC flash that can dedicate some percentage of the pages per die to use in SLC mode. It doesn't have quite the performance or endurance of pure SLC but it's close. Reply
  • Oxford Guy - Wednesday, July 15, 2015 - link

    " It doesn't have quite the performance or endurance of pure SLC but it's close."

    Any data on that to link to?
    Reply
  • Kristian Vättö - Thursday, June 18, 2015 - link

    "TLC for everyone" meaning that now every OEM can build a TLC drive if they wish.

    SLC cache implementation depends on the NAND. NAND vendors do have special dies with SLC cache functionality (supports a special command set), but so far they haven't sold them to third parties, although I've heard this is about to change. In other cases SLC cache simply means that the controller only writes to the lower pages, which in terms of write performance is about the same as real SLC cache but there is some minor read performance penalty (although I don't know exactly why).
    Reply

Log in

Don't have an account? Sign up now