AnandTech Storage Bench - The Destroyer

The Destroyer is an extremely long test replicating the access patterns of very IO-intensive desktop usage. A detailed breakdown can be found in this article. Like real-world usage, the drives do get the occasional break that allows for some background garbage collection and flushing caches, but those idle times are limited to 25ms so that it doesn't take all week to run the test. These AnandTech Storage Bench (ATSB) tests do not involve running the actual applications that generated the workloads, so the scores are relatively insensitive to changes in CPU performance and RAM from our new testbed, but the jump to a newer version of Windows and the newer storage drivers can have an impact.

We quantify performance on this test by reporting the drive's average data throughput, the average and 99th percentile latency of the I/O operations, and the total energy used by the drive over the course of the test.

ATSB - The Destroyer (Data Rate)

The Toshiba XG5's average data rate on The Destroyer isn't fast enough to compete with Samsung's drives, but it beats some of the slower MLC-based PCIe SSDs and is much faster than the other (non-Samsung) TLC SSDs.

ATSB - The Destroyer (Average Latency)ATSB - The Destroyer (99th Percentile Latency)

The average latency of the XG5 on The Destroyer is slower than the top NVMe SSDs including the Samsung 960 EVO, but the XG5's performance is still decent. The TLC-based WD Black and Intel 600p have average latencies that are several times higher. The 99th percentile latency of the Toshiba XG5 is very good, with a third-place score that beats all of Samsung's SSDs.

ATSB - The Destroyer (Average Read Latency)ATSB - The Destroyer (Average Write Latency)

The drive rankings for average read and write latencies are very similar, with the Toshiba XG5 scoring in the middle of the pack on both measures. The overall spread of scores is very different: while the XG5's average read latency is only 25% slower than the fastest SSD and is essentially tied with the Samsung 960 EVO, its average write latency is almost 3.5 times worse than the fastest drive.

ATSB - The Destroyer (99th Percentile Read Latency)ATSB - The Destroyer (99th Percentile Write Latency)

The 99th percentile latency scores of the XG5 are great for a TLC SSD. The write latency places the XG5 in the top tier of drives, while its read score is better than average among a widely varying bunch. The Intel SSD 750 is the only drive that offers better 99th percentile latency for both read and write operations.

ATSB - The Destroyer (Power)

Only a handful of drives have ever completed The Destroyer while using less energy than the Toshiba XG5. Most of those were SM2246EN-based SATA drives, and a few other SATA SSDs. The Crucial MX300 and Intel 545s are the only other TLC SSDs with this kind of efficiency. None of the other NVMe SSDs come close; the fastest competitors sacrifice power efficiency to obtain that performance, and the slower NVMe SSDs waste too much power operating a PCIe link that they can't fully utilize.

Introduction AnandTech Storage Bench - Heavy


View All Comments

  • Tros - Thursday, August 3, 2017 - link

    It's so weird. Almost like it was published without utility and just the timing necessary to boot ThreadRipper out of the spotlight on the front page. Reply
  • Billy Tallis - Friday, August 4, 2017 - link

    It was purely coincidental that this review and the TR unboxing news were published on the same day. I didn't even know about the TR unboxing embargo until a few hours before these articles went up. But we did deliberately position this article above the TR news, because we believe substantive reviews are more important than unboxing pictures. Reply
  • Sivar - Thursday, August 3, 2017 - link

    The technical detail and comprehensiveness of this SSD review is impressive.
    It must have taken man weeks to put this all together.
  • Billy Tallis - Thursday, August 3, 2017 - link

    It did take quite a while to put together the new test suite, and re-test older drives for comparison data. The Intel 750 is a particular nuisance since I only have equipment to measure one power rail at a time and the 750 uses both the 12V and 3.3V lines.

    Now that I have the new test suite (mostly) complete and automated as much as possible, I hope to be able to churn out reviews more quickly. There will be another SSD review going up next week while I'm at Flash Memory Summit.
  • Ratman6161 - Monday, August 7, 2017 - link

    AT is my go to place for SSD reviews. I like the consistent testing procedures over time which really helps to do apples to apples comparisons. One thing I' would like to see though is more comparisons of different sized drives in the same brand/model family.
    For example I was recently shopping for an NVMe drive and Samsung has been my go to brand for a few years now for high-mid range with crucial for my more run of the mill drives. So of course I was considering the 960 EVO. The trouble was I didn't need or want to pay the price for a 1 TB model (have a 750GB MX300 for bigish storage)...the 256GB was more in my price range. But you only usually have results for the 1 TB model. Digging around on other sites I was able to determine that the 256GB model took a huge performance hit compared to the larger sizes. In the past, Samsung drives used to seem like they took less of a hit when down sizing capacity, which was another plus for the brand. I ended up with a 512 GB Evo as a good balance between price and performance...but there were a lot of guesswork and assumptions involved with that choice.
    I can't be the only one out there that thinks this way?
  • Billy Tallis - Monday, August 7, 2017 - link

    We always try to get as many different capacities as possible. Some vendors are more willing than others to sample one of each size. Lately with the flash shortage, most vendors have been hesitant to send the full range of samples, and even after the product launches their PR teams have pretty limited sample availability.

    When I'm reviewing a single drive, I try to include primarily other drives from the same capacity class. If you want to compare eg. a 250 GB NVMe drive against a 512GB SATA drive, that's what our Bench database is for. (Though I haven't posted the 2017 results to Bench yet, and won't have time for that until after Flash Memory Summit.)
  • Bullwinkle J Moose - Thursday, August 3, 2017 - link

    NewEgg insiders could buy the Samsung 850 Pro this week for $109 (256GB)

    A 10 year warranty on the Worlds Fastest consumer SATA drive turned out to be less expensive than a slower TLC drive with 3 - 5 year warranty then adding an extended warranty

    Or you could just save $10 and get a slower TLC drive with a crappy warranty

    Your choice
  • Kwarkon - Saturday, August 12, 2017 - link

    Hi, you state that you have L1.2 enabled in your Idle powertest but clearly it is not working on your setup.
    I guess you use PCIe vertical to m.2 adapter and because of that you cannot go lower than L1.
  • SanX - Monday, August 14, 2017 - link

    A year later and still 3x slower then Samsung 960 ? Why this company still exists? Reply

Log in

Don't have an account? Sign up now