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 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 ADATA XPG SX950 is the slowest MLC SSD in this group, going by its average data rate on The Destroyer, while the similarly-equipped Crucial BX300 is the second-fastest SATA drive in the half-TB capacity class.

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

ADATA has a latency problem on The Destroyer. The SX950's average latency is much worse than any other MLC SSD, and the 99th percentile latency as bad as the TLC-based SU800, which was already an extreme outlier.

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

Splitting the average latency up by read and write operations, it's clear that the SX950's troubles are mostly on the write side, though the average read latency is also more typical of a TLC SSD than one with 3D MLC.

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

The 99th percentile read latency of the ADATA SX950 is not much worse than other 3D MLC SSDs, but the 99th percentile write latency on The Destroyer is unusually high at over 81ms. It appears that the SX950 is being quite aggressive with its SLC caching, leading to a serious backlog when it is finally forced to perform garbage collection. The BX300 avoids this by using relatively small fixed-size SLC caches.

ATSB - The Destroyer (Power)

Given the mediocre data rate and poor QoS indicating a lot of background work, it's not too surprising to see that the SX950's energy usage on The Destroyer is substantially higher than the Crucial BX300 and most other 3D NAND SSDs. The SX950 does shave 25% off the energy usage of the TLC-based ADATA SU800, but Crucial still does much better with the same controller.

Introduction AnandTech Storage Bench - Heavy
POST A COMMENT

45 Comments

View All Comments

  • Lord of the Bored - Tuesday, October 10, 2017 - link

    Nanu-nanu, as you centaurians say. Reply
  • BrokenCrayons - Tuesday, October 10, 2017 - link

    *beep-borp* I am an alien. I am superior. *borp-borp-beep* Reply
  • svan1971 - Wednesday, October 11, 2017 - link

    I stopped reading after I don't identify as human. To much self loathing from a no doubt educated idiot. Reply
  • svan1971 - Wednesday, October 11, 2017 - link

    LMAO Perfect... Reply
  • Samus - Monday, October 9, 2017 - link

    Even if the SU950 was cheaper than the BX300, I'd still rather have the BX300. I've never had to send a drive in to Crucial for warranty. Ever. Intel and Crucial have a 0% defect rate in my workplaces.

    Samsung had a number of 840 EVO's go sour years ago that resulted in a number of drive being sent in and replaced with new drives with new firmwares. In 2014 an 840 Pro even went bad, although I realize that is quite an anomaly for Samsung. The 840 EVO's were well documented to have issues.

    I've seen a number of ADATA SP500's fail, they just drop and stop detecting at POST. Before Barefoot 3, OCZ drives had all the typical issues Sandforce drives were notorious for having until the SF-2281 launched and firmware matured. Recent OCZ drives, even the ARC100 (the cheapest Barefoot drive) is reasonably reliable. One was mailed in a few months ago for warranty due to Windows detecting SMART errors. The drive didn't fail, and data was cloned to an advance replacement OCZ mailed out next-day. The OCZ warranty process was excellent, but that doesn't help a drive began to fail.

    Two Mushkin Reactors suffered the same issue seemingly years apart, they would randomly not detect, give a BSOD, and so on. The data was cloned to replacement SSD's and the Mushkin drives were RMA'd (which was a complete pain in the ass compared to OCZ with a 2 week turnaround no less) and the drives were fleabayed.

    Granted, even Intel isn't immune to problems. Fortunately I have no SSD535's out in the field. These drives are notorious for self destructing from write amplification wear, and even though a firmware was issues to fix it recently, most of those drives have already killed themselves, and if you have an OEM model like a Lenovo, you can't apply the firmware (and Lenovo - reflecting their typical "quality" support - hasn't issued a firmware update even a year after Intel made it available.)

    Overall, my point is, why would anybody buy a drive from someone other than Intel, Micron/Crucial, or Samsung? It's just a ridiculous gamble and is unlikely to save you money. There are niche drives like the Reactor that is still the cheapest 1TB SSD, so there are exceptions, but what exactly is ADATA bringing to the table that Samsung isn't with the 750, Crucial isn't with the BX300, and Intel isn't with the 600p?
    Reply
  • ddriver - Monday, October 9, 2017 - link

    You poor peasants and your precious money. One's social standing is measured by how much one has spent on hardware, not the actual value of the purchase, and of course, how much RGB LEDs it has.

    Silly ADATA, still haven't figured out how to justify the higher cost of ownership due to the lack of vertical integration. 9 letters - RGB LED FTW. Why is the industry sleeping, we have RGB LEDs on mobos, coolers, ram, mice, keyboards, but not on SSD? Or maybe they are saving that for the next quantum leap in technology that's gonna leave people dazzled.

    What intel brings with the 600p is hard to topple, it sure ain't easy to make an NVME drive that lousy. I also like how certain fairly expressive enterprise intel ssd drives behave when they run out of write cycles. While other vendors drives remain read-only, giving you the possibility to retrieve or use the existing data at your leisure, intel had the ingenious idea that such drives should brick themselves on the next post cycle. Such a great and highly useful feature. Who wouldn't want that?
    Reply
  • Reflex - Tuesday, October 10, 2017 - link

    @samus You poor peasant! You poor poor peasant! Reply
  • Golgatha777 - Wednesday, October 11, 2017 - link

    Anecdotal evidence to support your last paragraph. I have probably 20 or so Intel, Crucial, and Samsung drives (75%+ being Crucial drives) spread around laptops, desktops, and even a couple of game consoles. Not one failure in the bunch. I did have to flash one of my M500 drives due to a post error, but the issue was well documented and a fix was issued within a month of it being reported by Crucial. I do own a couple of Sandisk drives, but I did my research and they use Marvell controllers and Micron RAM, so I felt like those weren't a gamble. Reply
  • Golgatha777 - Wednesday, October 11, 2017 - link

    That should be Sandisk RAM for the Sandisk drives (Ultra IIs), not Micron. Reply
  • leexgx - Monday, October 16, 2017 - link

    but this is a MLC drive so probably outlast most other drives Reply

Log in

Don't have an account? Sign up now