Performance Consistency

Performance consistency tells us a lot about the architecture of these SSDs and how they handle internal defragmentation. The reason we don’t have consistent IO latency with SSD 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 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 buttons 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.

  Plextor M6S Plextor M5M Crucial M550 SanDisk Extreme II Samsung SSD 840 EVO mSATA
Default
25% Spare Area

Ouch, this doesn't look too good. The IOPS is constantly dropping below 1000, which is something I'm not expecting to see anymore. Even the M550 can keep the IOPS at ~4000 minimum, so the M6S certainly doesn't do well here. With added over-provisioning (OP) the performance does look a lot better and the minimum IOPS jumps to ~10K but it's still a downgrade from the M5M. Given the change to a lighter controller this is perhaps expected, but I'm still worried about the consistency without additional OP. I would make sure to leave ~10-15% of empty space with the M6S and M6M to avoid running into inconsistent performance.

  Plextor M6S Plextor M5M Crucial M550 SanDisk Extreme II Samsung SSD 840 EVO mSATA
Default
25% Spare Area

  Plextor M6S Plextor M5M Crucial M550 SanDisk Extreme II Samsung SSD 840 EVO mSATA
Default
25% Spare Area

 

The Math of Marketing: Not All 19nm NAND Is the Same AnandTech Storage Bench 2013
Comments Locked

30 Comments

View All Comments

  • Kristian Vättö - Friday, April 11, 2014 - link

    Yes, there will be a review of the M6e. There has been some issues with testing and hence it's taking this long but it'll be my first priority as soon as we are able to overcome the issues.
  • n3cw4rr10r - Thursday, April 10, 2014 - link

    The prices for SSDs are still high imo. I am surprised they are still holding @ approx $1/gb in most cases.
  • philipma1957 - Thursday, April 10, 2014 - link

    samsung evo is far below $1 a gb and crucial M550's are far below $1 a gb
  • ssj3gohan - Thursday, April 10, 2014 - link

    Finally actual proper idle power consumption results! It only took you a bit more than 5 years!

    HIPM and DIPM have been supported on all desktop systems - all desktop OSes as well as all hardware platforms - since the Athlon 64 X2 and first generation Core 2 (i945 chipset) times. I have been measuring SSD power consumption since they first came out - and yes, ever since the first SSDs, most of them supported the full gamut of power management - and have always been baffled that no review site on the internet has ever published accurate idle power consumption data. You are - aside from a few blogs I have come across - the first one. Congratulations!

    Now, it's true that the first instances of especially DIPM were frought with latency issues and for a long time DIPM was turned off by default on Linux kernels (although Windows always forces it on in anything but the 'performance' power profile since Vista). Also, and this is probably the biggest issue, lots of BIOSes do not properly report SATA LPM or have it off by default. As it is usually fairly well hidden away it is not something many people think to turn on.

    SATA ALPM is not at all a mobile feature. It has never been. Since support was baked in it has always been available to both laptop and desktop users. And with SSDs spending >>99% in idle, it represents easy and harmless power savings both in the SSD as well as the SATA I/O on the motherboard.
  • zodiacsoulmate - Thursday, April 10, 2014 - link

    wow nice info!
  • chubbypanda - Friday, April 11, 2014 - link

    That's right. I always wondered why SSD reviewers at AnandTech claim it's mobile only. It's available at least on Denlow platform for sure.

    By the way, HIPM/DIPM feature itself is easy to control in Windows with simple registry update:
    https://communities.intel.com/message/225489
  • Kristian Vättö - Friday, April 11, 2014 - link

    "Also, and this is probably the biggest issue, lots of BIOSes do not properly report SATA LPM or have it off by default."

    That is the issue I had previously. The motherboard in my old testbed doesn't have the option to enable link power management and hence I couldn't perform those tests.
  • Ethos Evoss - Thursday, April 10, 2014 - link

    Don't understand your tests ...
    It is all bollocks.. so few months back M5M was best and now it is on bottom ? that is all BULLCRAP
  • DanNeely - Thursday, April 10, 2014 - link

    The M5M was reviewed a year ago. It's performance was underwhelming compared to sata drives then; but at the time getting an mSata drive at retail was much easier said than done; and at the time it was still a reasonable buy for its money. Since then much better mSata drives have came out and Plextor is trying to charge a premium device price while providing performance much worse than that of competing products that are significantly cheaper.

    http://anandtech.com/show/6722/plextor-m5m-256gb-m...
  • Ethos Evoss - Thursday, April 10, 2014 - link

    Then M6M is NOT big step from M5M .. I have 2x M5M 256GB and they flies
    I am not replacing then .. not worth ..

Log in

Don't have an account? Sign up now