Sign in with
Sign up | Sign in

TRIM Testing: Our Suite Evolves Yet Again

SanDisk X210 256 And 512 GB: Enthusiast Speed; OEM Reliability
By

Finally, I want to introduce a new test I've been working on: using JEDEC's 218A consumer workload trace to create a TRIM test. It's not a neatly-packaged little utility you can run at home. Rather, this is a test scripted in ULINK's DriveMaster 2012 software and hardware suite.

DriveMaster is used by most SSD manufacturers to create and perform specific metrics. It's currently the only commercial product that can create the scenarios needed to validate TCG Opal 2 security, but it's almost unlimited in potential applications. There are various hardware components associated with the platform, such as a SATA/SAS power hub that allows the benchmarked drive to be power-cycled independently of the platform. Much of the benefit tied to a solution like DriveMaster is its ability to diagnose bugs, ensure compatibility, and issue low-level commands. In short, it's very handy for the companies actually building SSDs. And if off-the-shelf scripts don't do it for you, make your own. There's a steep learning curve, but the C-like environment and command documentation gives you a fighting chance.

This product also gives us some new ways to explore performance. Testing the TRIM command is just the first example of how we'll be using ULINK's contribution to the Tom's Hardware benchmark suite.

DriveMaster's test window, with script, log, and register info.DriveMaster's test window, with script, log, and register info.

The suite ships with some built-in scripts, but also contains its own scripting language for extensibility and customization. This particular test uses JEDEC's published master trace of consumer I/O activity (similar to our Tom's Hardware Storage Bench trace). The read commands are removed from the trace, leaving write, flush, and TRIM commands. After secure erasure and writing preparatory data, the test commences. The trace is played against the drive four times using NCQ with and without TRIM, and DMA with and without TRIM. IOPS are measured and averaged every 100,000 commands. 

On a 256 GB drive, each iteration writes close to 800 GB of data, so running the JEDEC TRIM test suite once on a 256 GB SSD generates almost 3.2 TB of mostly random writes (it's 75% random and 25% sequential). By the end of each run, over 37 million write commands are issued. If that sounds like a lot of storage traffic, it is.

The first two tests employ DMA to access the storage, while the last two use Native Command Queuing. Since most folks don't use DMA with SSDs (aside from some legacy or industrial applications) we don't concern ourselves with those. It can take up to 96 hours to run one drive through all four runs, though faster drives can cut the time in half, roughly. Because so much information is being written to an already-full SSD (the drive is filled before each test, and then close to 800 GB are written per iteration), SSDs that perform better under heavy load fare best. Without TRIM, on-the-fly garbage collection becomes a big contributor to high IOPS. With TRIM, 13% of space gets TRIM'ed, leaving more room for the controller to use for maintenance operations.

TRIM Testing

Here's the chart derived from our DriveMaster JEDEC TRIM test data. We have the 256 GB SanDisk X210, Samsung's venerable 840 Pro at 256 GB, and Crucial's more mainstream M500 (240 GB). Each device's NCQ-based test is plotted. The solid line represents average IOPS every 100,000 commands, but without TRIM. The hashed line represents performance every 100,000 with TRIM. In each case, the workload is mixed in with tons of small, random writes.

It seems logical that adding TRIM is helpful (depending on when and how a drive prefers to incorporate TRIM functionality). But that's not quite what we see. Crucial's 240 GB M500 doesn't show much gain from the addition of TRIM; both runs hover under 2000 IOPS. Samsung's 840 Pro enjoys substantial gains as as the test drags on. By the end, the 840 Pro is 50% faster with TRIM.

SanDisk's 256 GB X210 is almost as quick as the 840 Pro and M500 combined, though. In this trace-based benchmark, it appears that nCache is in its element, and even without additional over-provisioning, the X210 is smoking-fast. But using TRIM seems detrimental to performance. Either the X210 is tuned to excel in environments that don't support TRIM, or it's fast enough that the overhead associated with TRIM hurts more than it helps. Either way, SanDisk crushes this test. Interestingly, the X210 and M500 share Marvell's storage processor. It's the difference in NAND and firmware that yield the gap we're measuring.

Display all 22 comments.
This thread is closed for comments
  • 1 Hide
    vertexx , January 15, 2014 3:15 AM
    I logged onto Tom's this morning totally expecting a Kaveri review. What's up?
  • 2 Hide
    TeraMedia , January 15, 2014 9:00 AM
    Is the warranty 5 years or 3? Last page says one thing, an early page says another.Decent review, decent drives. Has THG considered doing something similar to what the car mags do, where they take certain products and use them for a year? It would be great to capture that kind of longer-term info on certain types of products, especially the kind that wear out (ODDs, fans, cases, HDDs, SSDs, etc.).
  • 1 Hide
    Quarkzquarkz , January 15, 2014 9:05 AM
    What about Samsung SSD pro 512GB? I bought 2 of these and on that chart is only 128 and 256GB
  • -1 Hide
    vmem , January 15, 2014 9:07 AM
    @vertexxthere isn't anything particularly exciting about Kaveri going by Anand's review. I shall want for the A10 version with higher clocks
  • 2 Hide
    smeezekitty , January 15, 2014 11:05 AM
    MLC with 5k write endurance!And affordable and fast?We may very well have a new solid contender in the SSD world
  • 1 Hide
    eklipz330 , January 15, 2014 3:47 PM
    anandtech has one? sweeeet, later tom's!
  • 3 Hide
    RedJaron , January 15, 2014 3:52 PM
    I agree with Chris. I don't need the fastest bench speeds in a SSD. Most models now are very fast and the user won't see the performance difference. I want reliability and longevity. Looks like this is a smart choice for any new builder.
  • 2 Hide
    Phillip Wager , January 15, 2014 9:54 PM
    sweet finally another company that can compete with intel's 5 year warrenty!
  • 2 Hide
    jake_westmorley , January 16, 2014 2:31 AM
    Can we PLEASE have some normal graphs for once? The graph on page 5 in stupid 3D is so bad it's comical. The "perspective" effect completely screws with the data. This has zero added value and is almost as bad as still using clipart.
  • -1 Hide
    Duff165 , January 16, 2014 2:35 AM
    I find it hard to believe that the author has had "literally dozens of SSd's die" on him over the years. This would suggest that many systems have contributed to the demise of many of the SSD's being used, which seems somewhat outlandish. Just the cost factor involved in the purchase of so many SSD's and then having over a dozen of them fail, supposedly also from various companies, since if they were all from the same company it would not really be conducive to good sales. One, or maybe two I could live with, but dozens? No.
  • 1 Hide
    jabliese , January 16, 2014 9:18 AM
    @Duff165 I would respond to your ignorance, but I really, really dislike this "new" comment system that Tom's has foisted on us.
  • 0 Hide
    smeezekitty , January 16, 2014 11:02 AM
    Quote:
    @Duff165 I would respond to your ignorance, but I really, really dislike this "new" comment system that Tom's has foisted on us.


    Agreed. Now it is truly screwed more than ever.
    WTH is will the continuously scrolled unrelated articles, dropped newlines and just plain broken comments.
  • 2 Hide
    RedJaron , January 16, 2014 11:23 AM
    Quote:
    I find it hard to believe that the author has had "literally dozens of SSd's die" on him over the years. ... One, or maybe two I could live with, but dozens? No.

    For the typical user that only has a few systems at the same time, yes. For someone who works with dozens of systems, the chance for one of them to fail goes up. Constantly building, testing, and reconfiguring machines means you're constantly imaging and re-imaging drives, which wears them out much faster than normal. Perhaps it was a little exaggerated, but it's definitely possible
  • 0 Hide
    tripleX , January 16, 2014 7:05 PM
    Lol. All of the work that went into making page 5 seem like some super-complicated test. That is just HDTune, which can be ran with the click of a button and hardly requires and entire page and horrible graph for explanation.
  • 1 Hide
    stewieco , January 17, 2014 9:08 AM
    Question: Can this drive be used on a personal laptop??
  • 2 Hide
    smeezekitty , January 18, 2014 2:53 PM
    Quote:
    Question: Can this drive be used on a personal laptop??


    I don't see why not
  • 0 Hide
    LFDRTechPro , January 20, 2014 5:14 PM
    There's actually a pretty huge difference between MLC-SLC translator based emulation and the way that tlc can behave electronically(physically) as SLC independently at the bottom of the translation layer. MLC-emulation has a partially positive and partially negative effect on translator stability, but with a robust controller(Marvell), the net effect on reliability is positive.With TLC as SLC, the effect is nearly perfectly proportionate to having SLC physically present.In fact, it is possible(with manufacturer tools) to take just about any TLC device like a 64GB usb flash drive and make it electronically indistinguishable(nearly) to a ~22Gb TLC flash drive in respect to performance, endurance, etc...MLC emulation is much more nuanced.
  • -1 Hide
    mwearl , January 20, 2014 6:52 PM
    @Duff165. It's very believable that the author could have had dozens of SSD failures, especially if he used some brands like Mushkin. We were building systems using their 120GB SSDs and had nearly 75% failure rate within the first 6 months. Since then, we have used primarily Kingston with a few Intel and Crucial SSDs and haven't had a single failure since.
  • -1 Hide
    mwearl , January 20, 2014 6:54 PM
    @Duff165. It's very believable that the author could have had dozens of SSD failures, especially if he used some brands like Mushkin. We were building systems using their 120GB SSDs and had nearly 75% failure rate within the first 6 months. Since then, we have used primarily Kingston with a few Intel and Crucial SSDs and haven't had a single failure since.
  • 2 Hide
    gwolfman , January 21, 2014 10:31 AM
    Since the major difference between the X210 and the Extreme II is the firmware, can you please test the Extreme II on the new DriveMaster TRIM test and see if they behave the same? SanDisk just turned into a new favorable option for me for my next drive.
Display more comments