Performance Consistency

Performance consistency tells us a lot about the architecture of these SSDs and how they handle internal fragmentation. The reason we do not have consistent IO latency with SSDs 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 (Logical Block Addresses) 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 dropdown selections 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.

Micron M600 256GB
Default
25% Over-Provisioning

Performance consistency is not very good, but that was expected as we are dealing with a DRAM-less design. With insufficient cache for the NAND mapping table, it is very hard to manage the table efficiently in the steady-state environment, which results in decreased performance. The good news is that Chrome OS is extremely light and a scenario like this is basically impossible, so the data here is merely an interesting piece of information.

Micron M600 256GB
Default
25% Over-Provisioning

 

Micron M600 256GB
Default
25% Over-Provisioning

TRIM Validation

To test TRIM, I filled the drive with sequential 128KB data and proceeded with a 30-minute random 4KB write (QD32) workload to put the drive into steady-state. After that I TRIM'ed the drive by issuing a quick format in Windows and ran HD Tach to produce the graph below.

It seems that TRIM is working, although there is an odd drop in read performance after the ~140GB mark.

Managing Storage in Chrome OS AnandTech Storage Bench 2013
POST A COMMENT

67 Comments

View All Comments

  • hojnikb - Tuesday, October 21, 2014 - link

    Given the space constraints of the 2242, wouldn't make more sense to go with something like sandforce ?
    This is already designed with dramless in mind, so it would perform better. And at this point, these old controllers must be dirt cheap.

    Also one more thing. Given how many cheap laptops use eMMC instead of real ssds, would it be possible to test that aswell ? As i'm aware, eMMC solutions are usually not that fast, but i do wonder how slow they really are.
    Reply
  • III-V - Thursday, October 23, 2014 - link

    eMMC 5.0 should be very fast, and should be surfacing in devices this year. Reply
  • noelbonner - Tuesday, November 11, 2014 - link

    I'd go for one of the top laptops on the market instead (like the rankings at http://tinyurl.com/msegfz9 for example). Reply
  • duploxxx - Tuesday, October 21, 2014 - link

    Anandtech quote:

    While the whole netbook boom kind of died with the introduction of tablets, Chromebooks have been gaining more and more traction recently. The original Windows netbooks failed to provide a smooth user experience due to the lack of operating system optimization, and Windows was simply way too heavy to be run with such limited resources

    requires a BIG correction, AMD Brazos was and still is more then fine to run these netbook designs, a 7.2k or better SSD HD gives a very good daily usage of that device. It are the horrible ATOM all over the world thx to Intel and OEM designs that screwed the netbook usecases......
    just like first generation ATOM for tablet is useless and made th windows tablet flop.
    Reply
  • lilmoe - Tuesday, October 21, 2014 - link

    Exactly. Hardware played a larger role in the failure of netbooks. It was simply too slow, and low power processors weren't "there" yet.
    Microsoft should make a huge comeback with Windows 10 in netbook form factor. Interesting will be the price and capabilities of these devices. Good performing $150-$250 Windows 10 netbooks will eat Chromebooks' lunch and make it seem they never really existed.
    Reply
  • titaniumalloy - Tuesday, October 21, 2014 - link

    I think the biggest problem is that they had stupid slow 320GB HDD. If they had used 16GB SSD, the machines would have been good. Of course, SSD were pricey 4 years ago. I believe if Microsoft come back into the netbook arena, they would perform well. Microsoft is usually way to ahead of its time or too far behind. Reply
  • hojnikb - Tuesday, October 21, 2014 - link

    Dont forget. Very first netbooks were fitted with ssds (very small ones though). Reply
  • andrewaggb - Tuesday, October 21, 2014 - link

    I owned 2 netbooks, one with an atom processor and a 1024x600 res screen, one with an amd c50 and 1280x720 screen.

    I hated them both. The keyboards were small and cramped, the screen resolution on both needed to be 768 minimum height. The hard disks were painfully slow, it took forever to get anything loaded. The amd c50 cpu was awful. It was cpu bound by all sorts of things and not nearly enough applications were gpu accelerated.

    Personally, I think the 11-13" laptop range is a much better size, with at least a 1366x768 resolution and an ssd. We have those today, but not for $200.

    I'd definitely rather have windows than a chromebook, but microsoft needs to get their windows updates streamlined and smaller...
    Reply
  • jabber - Wednesday, October 22, 2014 - link

    I've tried some of the new ultra cheap Windows 8.1 laptops with AMD E1 CPUs etc. and they are terrible. The CPU is so underpowered it runs at 100% the whole time. Poor HDD performance etc. Miserable experience. The problem is these are the new cheap laptops folks are buying. Makes you really appreciate a lowly Athlon or Celeron CPU, at least they worked. In comparison a Chromebook works far better for the average Joe. Reply
  • abianand - Thursday, October 23, 2014 - link

    Let us hope the laptop manufacturers stop pairing Windows 8.1 only with slow CPUs.

    Richland and Kaveri have many design wins and have many (I say many, keeping in mind I'm talking about AMD CPUs here) laptops in the US market but not in many other countries.
    Reply

Log in

Don't have an account? Sign up now