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)

Samsung's dominance of this test wasn't being seriously challenged, but the 512GB 860 PRO does show improvement to the average data rate on The Destroyer, putting it up in the range of Samsung's multi-TB SATA drives. It's a small change, but SATA doesn't leave room for big gains.

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

The good average and 99th percentile latency scores of the SanDisk Ultra 3D match or beat the best scores from the Samsung SATA drives. The 512GB 860 PRO shows substantial improvement in 99th percentile latency and more modest gains in average latency, relative to the 850 PRO.

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

The Samsung 860 PROs show the best average read latencies in their respective product classes, but the SanDisk Ultra 3D isn't far behind. For average write latencies, the Ultra 3D takes a clear lead over the Samsung drives, and the Crucial BX300 is ahead of the Samsung drives by a hair.

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

The 99th percentile read and write latencies of the 860 PRO show substantial improvements at 512GB, and smaller improvements among the multi-TB drives.

ATSB - The Destroyer (Power)

While the 4TB 860 PRO performed better on The Destroyer than the 512GB model by every measure, the 512GB model was more power efficient, and sets a new record for its class. The improvements relative to the 850 PRO are remarkable: the old 512GB 850 PRO required 60% more energy to complete The Destroyer than the new 512GB 860 PRO. Samsung has caught up with the modern competitors in terms of energy efficiency.

Introduction AnandTech Storage Bench - Heavy
Comments Locked

64 Comments

View All Comments

  • MayDayComputers - Wednesday, January 24, 2018 - link

    It wakes up in 8 milliseconds. The graph is in nanoseconds.
  • MayDayComputers - Wednesday, January 24, 2018 - link

    Yikes. Actually I was off, it wakes up in 8 nanoseconds, not milliseconds. Even at 8 milliseconds, you would never notice. This is 1000x faster than that.

    “A microsecond is equal to 1000 nanoseconds or 1/1,000 milliseconds. ” -source Wikipedia
  • stux - Saturday, February 17, 2018 - link

    Actually a million times faster.
  • letmepicyou - Wednesday, January 24, 2018 - link

    So...tests were done a while back that showed that 2x 850 EVOs in RAID 0 outperformed a single 850 Pro. The allure of putting 2 500gb EVOs in RAID for $300 and getting better performance than a single PRO 1TB for $450 was a no-brainer IMO, and boy does it scoot. My question is, what do RAID numbers look for the new 860's? Will it still make sense to RAID the EVO's? Or will the 860 series price/perf/space metric slant more towards the single PRO drive?
  • BurntMyBacon - Wednesday, January 24, 2018 - link

    Given that the 860 Pro struggles to distance itself from its EVO counterpart in a straight up comparison, I think its safe to say 2x860EVO will outperform 1x860PRO in the same metrics that 2x850EVO outperforms 1x850PRO. Also, MSRP shows 2x860EVO 500GB costing $340 vs $480 for the 1TB 860PRO. Your price/perf/space metric will not be slanting towards the PRO drive.
  • Luckz - Monday, May 14, 2018 - link

    SATA SSD RAID0 only makes sense if you need 'left to right copying' and can't do NVMe (which is much faster).
  • Roen - Monday, January 29, 2018 - link

    I'd like to see what Enterprise SATA / SAS SSDs the author has in mind that is a better balance of specs and price, especially price.

    The cheapest Seagate Nytro SAS SSD I've found with 1700 / 850 Sustained R/W is > $1500.
  • peevee - Tuesday, January 30, 2018 - link

    There is absolutely no point paying twice as much compared to other drives.
  • JokerzWild - Sunday, February 11, 2018 - link

    Good review, but it seems like there’s a “missing link” in the data presented. You mention in the Introduction that you are using a 512GB 850 PRO from, “the original generation using 32L 3D NAND and LPDDR2 DRAM, rather than the updated model with 48L 3D NAND and LPDDR3.” You also say that you are using the test results from the 2TB 850 Pro review, which I believe still used Samsung’s 2nd-generation (32L) 3D MLC NAND with LPDDR3 in the controller rather than their 3rd-generation (48L) product. I can’t tell which version of the 500GB 850 EVO is being presented for this review, but I suspect it is V1 as well based on its power usage compared to the 4TB 850 EVO (which was offered in V2 only) numbers. In your recent reviews of other current SSDs (e.g. the SanDisk Ultra 3D and the Crucial MX500) it also appears that you are using the 32L/LPDDR2 versions (which I’ll call V1) rather than the 48L/LPDDR3 (V2) iterations of the 850 PRO and EVO.

    Would you please add the data for 850 EVO/PRO V2s to your test results? Alternatively, it would be interesting to see an article that looks at the progress of Samsung’s NAND and controllers in V1 (32L/LPDDR2) and V2 (48L/LPDDR3) of the 850 EVOs and PROs along with the 860s (64L/LPDDR4), preferably in the 1TB configuration since that configuration seems to yield the highest overall performance (at least in the 850s) and is a popular choice. I’ve seen a couple of articles that compared V1 and V2 of the 850 EVO, and it appeared that both performance and power consumption improved in V2. I have yet to seen any comparison of V1 and V2 of the 850 PRO. I seem to also recall reading that Samsung was claiming a 30% reduction in power consumption was one of the benefits in switching from its 2nd-generation to its 3rd-generation NAND, which would wipe out most of the power management gains claimed for the 860 PRO. The changes between V1 and V2 of the 850s appear just as significant as the changes between 850 V2s and the 860s (1 generation DRAM, +16L each), so why not add this data to the mix? Using V1 of the 850s probably overstates the differences between a recent 850 EVO/PRO and an 860 EVO/PRO. Presenting data on V2 of the 850s would also give users of 850 V2s a better idea of what they may be missing. It would also help bargain hunters who want that last little bit of performance from a top performing SATA drive decide if it’s better to buy a marked down 850 or a new 860 for their use cases.
  • Lady Fitzgerald - Tuesday, March 13, 2018 - link

    SATA will be around for a long time, most likely well past 5 years from now. It's plenty fast for storage. Not everyone needs blazing speed for storing music, videos, documents, pictures, etc.

Log in

Don't have an account? Sign up now