Whole-Drive Fill

This test starts with a freshly-erased drive and fills it with 128kB sequential writes at queue depth 32, recording the write speed for each 1GB segment. This test is not representative of any ordinary client/consumer usage pattern, but it does allow us to observe transitions in the drive's behavior as it fills up. This can allow us to estimate the size of any SLC write cache, and get a sense for how much performance remains on the rare occasions where real-world usage keeps writing data after filling the cache.

The Corsair MP400 opts for the largest possible SLC cache size, allowing for a quarter of the drive's advertised capacity to be written to the cache before performance plummets. This is the same strategy used by the Sabrent Rocket Q and probably all other QLC drives using Phison controllers. The Intel and Crucial QLC drives based on Silicon Motion's SM2263 controller have somewhat smaller variable-sized SLC caches, while the Samsung QLC SATA drives use the same small SLC cache sizes as their TLC counterparts.

Sustained 128kB Sequential Write (Power Efficiency)
Average Throughput for last 16 GB Overall Average Throughput

Overall drive fill performance for the Corsair MP400 is marginally faster than for the other 1TB QLC drives we've tested, but the DRAMless TLC drives and the larger Sabrent Rocket Q are considerably faster. However, all of the budget NVMe drives are clearly much slower for sustained writes than the mainstream and high-end TLC drives.

Working Set Size

The random read latency from the Corsair MP400 is quite similar to what we saw from the 8TB Sabrent Rocket Q, especially for medium to large working sets. The 1TB MP400 doesn't exhibit the same unsteady performance for small working sets as the 8TB Rocket Q, but instead only shows poor performance for the absolute smallest working set size tested.

The performance drop-off when performing random reads across the entire drive is expected and normal, because the Crucial P1 is the only QLC drive in this bunch to include the full 1GB per 1TB of DRAM. Most low-end DRAMless TLC drives will show an even earlier drop in performance, and many of the more mainstream TLC drives that have switched to a lower DRAM ratio will also show the same drop that the Rocket Q and MP400 show.

Introduction AnandTech Storage Bench
POST A COMMENT

75 Comments

View All Comments

  • Spunjji - Monday, December 14, 2020 - link

    This is the only one of the anti-QLC tropes I see routinely rolling around this comment section that I 100% unequivocally agree with.

    A quick scan through a certain UK retailer shows the cheapest 1TB drive is the Intel 665p at £89, with the cheapest (relatively crappy) TLC drive at £96 and a WD Blue at £103. Worse still, at the 2TB level the positions flip and the WD Blue is *cheaper*.

    At the 1TB level, even though I know objectively that the 665p would be fine for my purposes I'd still be tempted to pay the £14 extra for the Blue. If the difference were £25 or more, I wouldn't.
    Reply
  • GeoffreyA - Monday, December 14, 2020 - link

    It's as if margarine were suddenly the same price as butter, or quite close, and the makers, exterting their marketing force, succeeded (almost succeeded) in blurring the distinction between the two.

    It would be nice if somebody did a giant endurance experiment, finding out exactly where these QLC drives stand, like the one Techreport tackled in 2015, writing an horrific number of TBs till the drives "breathed their last." There will be surprises.
    Reply
  • Oxford Guy - Monday, December 14, 2020 - link

    They will succeed. Getting everyone to go to a small form factor was very helpful, along with, apparently, not producing TLC in 1024Gbit dies.

    Margarine was superior to butter back in the day, remember? Superior partially hydrogenated technology. Because they said so.
    Reply
  • Oxford Guy - Monday, December 14, 2020 - link

    Not the best comparison, though – since butter was still widely available for reasonable prices.

    QLC, by contrast, is intended to ruin the economy of scale of TLC. We could find farmers with butter churns pretty easily. Not so easy to find small-scale TLC foundries for the peasantry.
    Reply
  • GeoffreyA - Tuesday, December 15, 2020 - link

    "Margarine was superior to butter back in the day, remember? Superior partially hydrogenated technology. Because they said so."

    Nice one.
    Reply
  • Oxford Guy - Monday, December 14, 2020 - link

    QLC doesn't need "tropes" any more than it needs cheerleading.

    Reality is that 16 voltage states is more problematic than 8 and fewer.
    Reply
  • boredsysadmin - Monday, December 14, 2020 - link

    I 100% agree with @kpb321
    I am also surprised that SK Hynix Gold P31 didn't make the last page budget consumer NVMe SSDs. MP400 1TB costs $114 on amazon, while one of the fastest budget drives in the review, P31 1TB, is currently at $120.60 - Which one you'd buy? I am "puzzled" by QLC drive makers' greed to raise the price per TB on bigger drives. They aren't the first to fill that niche - I expect the opposite - higher drives to be more expensive but priced the same or cheaper per TB.
    Reply
  • Drkrieger01 - Friday, December 11, 2020 - link

    I love how everyone's up in arms at 'Endurance'. Let me give you some insight on just how much 'endurance' you really need. I build a 'high speed storage' server, 16x 840 Evo 1TB, LSI MegaRaid 9750 16i back in 2014. We've been pounding the piss out of this server for *6* years. Not a single drive has reallocated sectors. I believe we've crossed a few petabytes on some of the drives, even after having to flash firmware updates and re-zero the drives due to the decay issue on the 840 Evo's.
    I'm sure these new drives will be just fine for the average user.
    Reply
  • inighthawki - Friday, December 11, 2020 - link

    Let's not forget that most people aren't coming *remotely* close to what you're doing. I bet even most power users would struggle to *consistently* write more than 10-20GB per day average to these drives. Hitting the endurance ratings at these rates would take decades. The SSDs will have been long discarded by that point.

    Even for the exceptional power user writing 100+GB a day, they would need to consistently do that every day for nearly 6 years to hit the endurance cap on the 1TB model. A user like this will also likely replace these drives within that time frame.
    Reply
  • Tomatotech - Friday, December 11, 2020 - link

    I have little to no worry about endurance for the vast majority of users even with QLC, and even that problem will go away with 2TB+ QLC.

    That said, your use is unusual, and the 840s were particularly good drives especially with 1TB. The same usage patterns with 256gb TLC or 512GB QLC might well have seen them wear out. Thankfully we’re past that stage now. However you still had to flash firmware and re-zero the drives, and most people with SSDs nowadays won’t know how to do that.
    Reply

Log in

Don't have an account? Sign up now