Sequential Read Performance

Our first test of sequential read performance uses short bursts of 128MB, issued as 128kB operations with no queuing. The test averages performance across eight bursts for a total of 1GB of data transferred from a drive containing 16GB of data. Between each burst the drive is given enough idle time to keep the overall duty cycle at 20%.

Burst 128kB Sequential Read (Queue Depth 1)

The burst sequential read performance of the Samsung 860 QVO is generally competitive with mainstream TLC SATA SSDs and is well ahead of the DRAMless Toshiba TR200. The 1TB 860 QVO's score is a bit lower when the drive isn't full because of the timing of the tests: the drive was still flushing the SLC cache in the background when the read test started.

Our test of sustained sequential reads uses queue depths from 1 to 32, with the performance and power scores computed as the average of QD1, QD2 and QD4. Each queue depth is tested for up to one minute or 32GB transferred, from a drive containing 64GB of data. This test is run twice: once with the drive prepared by sequentially writing the test data, and again after the random write test has mixed things up, causing fragmentation inside the SSD that isn't visible to the OS. These two scores represent the two extremes of how the drive would perform under real-world usage, where wear leveling and modifications to some existing data will create some internal fragmentation that degrades performance, but usually not to the extent shown here.

Sustained 128kB Sequential Read

On the longer sequential read test, the 860 QVO continues to get reasonably close to the SATA speed limit when reading data that is contiguous on the flash itself. Where internal fragmentation is caused by writing to the drive randomly, the QVO's read speed suffers much more than for the TLC drives, and the 1TB 860 QVO ends up slightly slower than a mechanical hard drive.

Sustained 128kB Sequential Read (Power Efficiency)
Power Efficiency in MB/s/W Average Power in W

The power efficiency of the 860 QVO is only a little bit lower than the TLC drives for the contiguous data case. When dealing with fragmented data, the QVO is slightly more efficient than the Intel/Micron NVMe QLC drives despite being a bit slower.

The queue depth scaling behavior for the 860 QVO is very typical, with QD1 not quite saturating the SATA link but all higher queue depths hitting close to full speed. The one exception is a slight decrease from the 1TB drive during the final QD32 phase.

Aside from the mild QD32 drop in performance, the sequential read behavior of the 860 QVO doesn't fall outside the normal ranges we've come to expect from TLC drives.

Sequential Write Performance

Our test of sequential write burst performance is structured identically to the sequential read burst performance test save for the direction of the data transfer. Each burst writes 128MB as 128kB operations issued at QD1, for a total of 1GB of data written to a drive containing 16GB of data.

Burst 128kB Sequential Write (Queue Depth 1)

The Samsung 860 QVO handles the burst sequential write test fine when the drive is mostly empty and there's plenty of room in the SLC cache. When the drive is full, the 1TB model's speed suffers somewhat, but is still much faster than the mechanical hard drive or the DRAMless TLC drive.

Our test of sustained sequential writes is structured identically to our sustained sequential read test, save for the direction of the data transfers. Queue depths range from 1 to 32 and each queue depth is tested for up to one minute or 32GB, followed by up to one minute of idle time for the drive to cool off and perform garbage collection. The test is confined to a 64GB span of the drive.

Sustained 128kB Sequential Write

On the longer sequential write test, the SLC cache of the 1TB 860 QVO is not quite enough even when the drive is mostly empty, so it ends up in last place. The 4TB model's SLC cache keeps up with this test and it is as fast as any SATA drive.

Sustained 128kB Sequential Write (Power Efficiency)
Power Efficiency in MB/s/W Average Power in W

The 860 QVO is a bit more power-hungry than the 860 EVO, so the 4TB QVO only takes third place for efficiency among the SATA drives in this bunch. The 1TB QVO has similar efficiency to the faster but more power-hungry 1TB QLC NVMe drives from Intel and Micron.

The 1TB 860 QVO is mostly slow and steady during the sequential write test, while the 4TB model's performance is as good as any other SATA drive.

The 1TB 860 QVO's sequential write behavior sticks out clearly as far slower than typical, but it's not unprecedented: there have been TLC drives this slow, but most of them were much smaller than 1TB. The 4TB model blends in with the crowd much better.

Random Performance Mixed Read/Write Performance
Comments Locked

109 Comments

View All Comments

  • 0ldman79 - Saturday, December 15, 2018 - link

    I found some answers to that question on the Bench.

    https://www.anandtech.com/bench/product/2339?vs=22...

    Load time, file copy times, etc, I guess more day to day testing would be appreciated. The spinner does a decent job on the BAPCo tests.

    I'm wondering if the QLC will really improve my load times vs my 2TB mechanical drive that sustains nearly 200MBps read speeds in practice.
  • ewitte - Thursday, March 28, 2019 - link

    Who in their right mind would purchase a QVO drive when the EVO is similarly priced? These drives perform poorly and have horrible longevity. 160MB/s writes after the cache!!!! I will not even touch the smaller EVO drives because the write speed is so low (500GB is ok but 1TB is so affordable now it is my minimum).
  • Sanmayce - Friday, March 29, 2019 - link

    Anand, please consider debunking the myth of data retention being limited to a few months.

    Really, where this myth originates from, f.e. if one buys QVO for backups what is the safe-time-between-losing-data?

    My proposal is to fill the drive with 900GB 7z archive and test the integrity after, say, 6 months.

    I myself am planning to buy the 1TB QVO just to find how durable it is under superheavy REAL-WORLD random read/writes - QD1. By the way, you are using 16GB spans in your tests, but the SLC is bigger than that, don't you see a problem?
  • bobhumplick - Thursday, May 16, 2019 - link

    so if the intel 660p and the smasung qvo are the same price which one do you get?
  • praveenvj - Friday, July 26, 2019 - link

    Now that price has dropped to $400 for 4TB, does this make sense compared to EVO for daily driver PC?
  • southleft - Monday, August 19, 2019 - link

    Here we are about 9 months later - August 2019 - and the price of Intel's 660p 1TB model has dropped by 50% when it's on sale. So, for $85 - $95 you can get an NVMe drive with, say, 750GB of usable capacity and it will blow all of these SATA SSDs out of the water. In other words, if you don't fill the drive over about 75% full you'll have a smokin' fast rig.
  • problemchild - Wednesday, October 23, 2019 - link

    As a professional many comments here disregard what these drives are intended for and for obvius reasons you wouldn't want to buy them for a server array or high availability applications. Samsung has entered a product which is now competing with lower priced product \ brands. For the average consumer looking for a Samsung branded drive to be used as a basic storage drive for documents, photos, music and game storage it makes a lot of sense. Admittedly IMHO buying a 2TB SSD for general storage use at the $200 mark is extreme in comparison to $80 for a spinner but for some users but for gamers it may be worth the price to use as game storage drive.
  • Scour - Sunday, June 28, 2020 - link

    Question: If the SLC-cache is full, the speed drops, that´s clear. But what happened if the SSD have a power loss directly after the data-writing ends? Is all Data still accessable on the SSD (because data still have to be copied from the SLC-cache to the QLC)?
  • leexgx - Thursday, August 6, 2020 - link

    did they ever issue a firmware update with the 4TB QVO, EVO and Pro samsung ssds been problematic sometimes not working

Log in

Don't have an account? Sign up now