TRIM Performance

In practice, SandForce based drives running a mainstream client workload do very well and typically boast low average write amplification. However if subjected to a workload composed entirely of incompressible writes (e.g. tons of compressed images, videos and music) you can back the controller into a corner.

To simulate this I filled the drive with incompressible data, ran a 4KB (100% LBA space, QD32) random write test with incompressible data for an hour and a half, and then ran AS-SSD (another incompressible data test) to see how low performance could get:

Kingston HyperX 3K - Resiliency - AS SSD Sequential Write Speed - 6Gbps
  Clean After 1.5hr Torture After TRIM
Kingston HyperX 3K 240GB 312.4 MB/s 103.9 MB/s 245.8 MB/s

I usually run this test for only 20 minutes but after seeing an unusually resilient performance by the 240GB drives I decided to extend the test period to a full 90 minutes. Performance does drop pretty far at that point, down to 103MB/s. TRIMing the drive does restore some performance but not all. If you have a workload that uses a lot of incompressible data (e.g. JPGs, H.264 videos, software encrypted data, highly random datasets etc...) then SandForce just isn't for you.

Power Consumption

SandForce boasts fairly low power consumption, particularly at idle. Even with incompressible data the HyperX 3K's power draw is competitive:

Drive Power Consumption - Idle

Drive Power Consumption - Sequential Write

Drive Power Consumption - Random Write


AnandTech Storage Bench 2011 - Light Workload


View All Comments

  • Coup27 - Tuesday, April 10, 2012 - link

    I've had this issue across multiple platforms though. W7 with IE9 and FF11. XP with IE8, and now my work machine which is W7 with IE9. I regularly run CCleaner so I really don't think it's been my end, sorry. Reply
  • InsaneScientist - Tuesday, April 10, 2012 - link

    I'm not a web programmer, so I can't help much with why it's happening, but, for what it's worth, I had some extra time so I tried to reproduce your problem.

    I tried every browser/os combination available to me at my house, which would be:
    Windows 8 CP x86 (to go) with IE10, Chrome 18, & FF 11
    Windows 8 CP x64 with IE10, IE10x64, Chrome 18, FF11, FF Aurora 13, & Safari 5.1.5
    Windows 7 x86 with IE9, Chrome 18, FF 11
    Windows 7 x64 with IE9, Chrome 18, FF 11, Safari 5.1.5
    Windows Vista x86 with IE9, Chrome 18, & FF 11
    Windows XP x86 (Virtual Machine) with IE8 and FF11
    Windows Thin PC x86 with IE9 & FF 11
    Safari in iOS 4.2 and 5.1
    The stock browser in Android 2.3.7 and 4.0.4

    I couldn't reproduce the problem at all... So it's definitely not something inherent to the website, because it works for some people and doesn't work for others... but I'm at a loss as to what that could be. It also doesn't seem like it's unique to you though (as seen in the comment below yours. Very strange...
  • bji - Tuesday, April 10, 2012 - link

    It's an intermittent problem; just because you couldn't reproduce it the time or times that you tried, doesn't mean that it's not inherent in the site. Unless by inherent you mean "happens every time", but we already established that that is not the case.

    The server(s) that serve up some part of the content of the site are occasionally generating something bad. Maybe there are multiple web servers behind a load balancer and one of them as a corrupted style sheet page or something that is occasionally being handed out?
  • bji - Tuesday, April 10, 2012 - link

    That's EXACTLY what I was seeing occasionally.

    Something is busted on your site; don't blame it on our browsers.
  • AssBall - Tuesday, April 10, 2012 - link

    An intermittent problem that you have and can't reproduce or analyze. I blame the user, not Anandtech or your browser. Reply
  • Coup27 - Wednesday, April 11, 2012 - link

    So I turn on my machine, load AT, and it does that? And thats somehow "the user"?

    Well done.
  • bji - Wednesday, April 11, 2012 - link

    If it were just me, maybe that would be true. But multiple people have reported the exact same symptoms. Reply
  • Jaraxal - Tuesday, April 10, 2012 - link

    I'd suggest a small modification to alert readers that they've reach the end of the article. As it is, it definitely appears as though it's missing "something" when the article just ends with a chart. Reply
  • KZ0 - Tuesday, April 10, 2012 - link

    I love it. Keep up the good work! Reply
  • aguilpa1 - Tuesday, April 10, 2012 - link

    That's cool but if there was a,... and in overview and conclusion notation (at the beginning) with a read next few pages for results of testing, information and benchmarks,... noted somewhere it would make more sense, don't ya think? Reply

Log in

Don't have an account? Sign up now