Performance Over Time & TRIM

SandForce's controllers have always behaved poorly if you pushed them into their worst case scenario. Should you fill a SF-2281 based drive completely with incompressible data then continue to write to the drive with more incompressible data (overwriting some of what you've already written) to fill up the spare area you'll back the controller into a corner that it can't get out of, even with TRIM. It's not a terribly realistic situation since anyone using an SF-2281 SSD as a boot drive will at least have Windows (or some other easily compressible OS) installed, and it's fairly likely that you'll have other things stored on your SSD in addition to large movies/photos. Regardless, it's a corner case that we do have to pay attention to.

I was curious to see if Intel's firmware did anything differently than the standard SandForce build used by other partners. To find out I took an Intel SSD 520 (240GB) and a Kingston HyperX (SF-2281 240GB) and filled both drives with incompressible data. I then ran a 60 minute 4KB random write torture test (QD32), once again, with incompressible data. Normally I'd use HDTach to chart performance over time however HDTach measures performance with highly compressible data so we wouldn't get a good representation of performance here. Instead I ran AS-SSD to get a good idea for incompressible sequential performance in this worst case state. Afterwards, I TRIMed the drives and ran AS-SSD again to see if TRIM could recover the drive's performance.

Intel SSD 520 - Resiliency - AS SSD Sequential Write Speed - 6Gbps
  Clean After Torture After TRIM
Intel SSD 520 240GB 284.5 MB/s 162.9 MB/s 162.9 MB/s
Kingston HyperX 240GB 289.3 MB/s 162.5 MB/s 162.5 MB/s

Surprisingly enough, Cherryville doesn't actually perform any differently than the stock SandForce firmware in this case. SandForce definitely improved the worst case scenario performance with the SF-2281 compared to the first generation controller, and it also improved performance compared to earlier builds of the 2281 firmware, but Intel appears to have not done anything above and beyond here. Based on these results I'd be willing to bet that Intel doesn't have source code access to the SF-2281 firmware otherwise it would've worked on a solution to this corner case. Performance in this worst case scenario isn't terrible but the fact that it's irrecoverable even after a TRIM is what's most troubling. Again, I don't see most end users backing themselves into this corner but it's worth pointing out.

AnandTech Storage Bench 2011 - Light Workload Power Consumption
POST A COMMENT

138 Comments

View All Comments

  • frank1985 - Monday, February 13, 2012 - link

    I am a customer of this guy, and I had to return my $240 Corsair Force 2 drive/paperweight *11 days* after purchase (at which point I traded it in for his old OCZ Vertex2 which has been running without issue).

    Like Iketh, the fact that you can't recover performance with TRIM is a showstopper for me. I'm going to stick with the 320 series drives - they might be slower, but TRIM does what it's supposed to.
    Reply
  • Wolfpup - Monday, July 23, 2012 - link

    The TRIM issue worries me too. Sounds like it's even worse than on Micron/Crucial's M4?

    My 320's been great, though annoyingly the 600GB version is MASSIVELY more than the 480GB 520, or the 512GB M4...possibly because of this issue?
    Reply
  • Ratman6161 - Wednesday, February 08, 2012 - link

    I have an OCZ drive based on the SF22xx. For the first month it was nothing but trouble. Then with the release of the 2.15 firmware, suddenly all the problems went away like magic. It could simply be that OCZ paid the price for being early adopters and released to the market a product that really should have been a Beta version. Now that it has the firmware that it should have had from the beginning, everything is fine.

    Intel has waited until the product is more mature so its no surprise they aren't seeing these problems
    Reply
  • Wolfpup - Monday, July 23, 2012 - link

    Did you read the article? Intel explicitly fixed problems still present in other firmware. Reply
  • ckryan - Tuesday, February 07, 2012 - link

    The Samsung 830 is maybe the finest consumer drive released in all of 2011. There is no shame in going Samsung, and in my mind, it's a fine alternative to almost every drive out there.

    Samsung and Micron probably have to wonder how Intel gets all the press about their NAND being better. As far as I can tell, Samsung's 27nm Toggle in the 830 is pretty damn good. The Micron NAND used in Micron's own drives is fantastic, and of course Intel's top shelf NAND is good too. But only Intel gets the press for having excellent NAND in it's own drives.

    I don't think enough emphasis is placed on the NAND. You can't get a clear picture of NAND quality through benchmarking.
    Reply
  • krumme - Monday, February 06, 2012 - link

    Exit Intel from the controller market. No wrapping in reliability, as rightfully as it is, can change that fact seen from here.
    Anyway, Intel can earn on the memory, as was probably the plan from day 1. We will see how profitability is here in 5 years time.
    Reply
  • tipoo - Monday, February 06, 2012 - link

    I still get a kick out of how a relatively tiny startup company makes better performing parts than the chipzilla that is Intel. Good call working with them. Reply
  • Makaveli - Monday, February 06, 2012 - link

    What I think you are missing is that Intel doesn't because they don't have to or care to. Intel is a huge company with equally big profit margins to maintain for their shareholders. The SSD market is still a small compared to the other things they dabble in. And yes they produced a controller in the past but I don't think the Money and R&D was worth it for them to continue based on the profits needed to keep them interested. We already know they make a ton of money selling nand as is.

    So to sum it up they could easily put out a controller that would be superior but just don't care too at the moment.
    Reply
  • Kevin G - Monday, February 06, 2012 - link

    Intel's plan was to gain control of the entire SSD product: from flash (joint venture with Micron), to the controller and to the firmware. It does appear that their original plan has fallen apart due to delays with their custom controllers which has forced them to use 3rd party controllers.

    I wouldn't say that Intel has stopped investing into developing their own controllers though. Intel can offer a level of integration that no other vendor can offer: direction integration into an x86 SoC. I strongly suspect that the successor to Haskwell and future Atom SoC will come with a direct ONFI connectivity backed by an internal Intel controller.

    At that point in time, I suspect that SSD controllers will become akin to what discrete GPU's are today: a midrange or high end item for consumers.
    Reply
  • NitroWare - Wednesday, February 08, 2012 - link

    People forget Intel was one of the innovators in bringing MATURE Flash to consumer technologies.

    Intel Desktop and OEM motherboards have used their surface mount flash chips for PC BIOS since 1996. This setup provided a recovery mode to get out of BIOS corruption. To this day, excluding motherboards with dual BIOS no motherboard has the same robust recovery system as Intel boards do.

    If the recovery mechanism fails then there is a more serous failure anyway.

    At the time other vendors were still using older DIP socket EEPROM and some still do now , only changing to serial format ROM.

    Even for current ASUS Flagship ROG boards, the best failsafe they offer is a socketed BIOS.

    Intel offered those Turbo Boost Flash cards several years ago for the 1156 motherboards and laptops which did not take off. They also championed Hybrid HDD for portables.

    So I think Intel know what they are doing with solid state, but yes eventually the technology will become commodity
    Reply

Log in

Don't have an account? Sign up now