Burst IO Performance

Our burst IO tests operate at queue depth 1 and perform several short data transfers interspersed with idle time. The random read and write tests consist of 32 bursts of up to 64MB each. The sequential read and write tests use eight bursts of up to 128MB each. For more details, please see the overview of our 2021 Consumer SSD Benchmark Suite.

QD1 Burst IO Performance
Random Read Random Write
Sequential Read Sequential Write

Given that the Samsung SSD 980 is a DRAMless drive with aggressive SLC caching, it's no surprise to see that most of the burst IO scores show a big difference between testing a small slice of a mostly-empty drive and testing across an 80% full drive. The test runs with the NVMe Host Memory Buffer feature disabled also show how important that is to enabling random IO performance. In the best cases, the SSD 980 even manages to deliver better QD1 random read performance than the 980 PRO. However, when testing against a large data set (or with HMB disabled), the SSD 980's performance only remains competitive for sequential reads. For the other IO patterns, its performance drops far more than typical for entry-level drives.

Sustained IO Performance

Our sustained IO tests exercise a range of queue depths and transfer more data than the burst IO tests, but still have limits to keep the duration somewhat realistic. The primary scores we report are focused on the low queue depths that make up the bulk of consumer storage workloads. For more details, please see the overview of our 2021 Consumer SSD Benchmark Suite.

Sustained IO Performance
Random Read Throughput Power Efficiency
Random Write Throughput Power Efficiency
Sequential Read Throughput Power Efficiency
Sequential Write Throughput Power Efficiency

When testing under favorable conditions with a small data set, the SSD 980's performance is at the very least competitive with other entry-level NVMe SSDs and sometimes against mainstream drives as well, but it falls apart too easily. Performance on the sustained IO tests shows similar behavior to the burst IO tests with generally quite large performance drops on the SSD 980 when testing is not constrained to a small data set. Bringing in some moderately higher queue depths has helped the sequential write performance catch up somewhat when testing against an 80% full drive.

Random Read
Random Write
Sequential Read
Sequential Write

When running the sustained IO tests against a 32GB range of the drives, the Samsung SSD 980 shows good performance scaling with increasing queue depths and can eventually achieve quite high throughput. Random reads take a long time to reach full speed at QD64, while random writes and sequential reads saturate around QD8 and sequential writes saturate at QD2. But testing on a wide range of a mostly-full drive reveals lots of problems: random reads and writes are very slow even at extreme queue depths, and the sequential write test frequently overruns the SLC cache.

Random Read Latency

This test illustrates how drives with higher throughput don't always offer better IO latency and Quality of Service (QoS), and that latency often gets much worse when a drive is pushed to its limits. This test is more intense than real-world consumer workloads and the results can be a bit noisy, but large differences that show up clearly on a log scale plot are meaningful. For more details, please see the overview of our 2021 Consumer SSD Benchmark Suite.

This random read latency test paints a pretty bleak picture for the Samsung SSD 980: every other drive in this bunch has lower latency and higher throughput, and the 500GB 980 actually does slightly better than the 1TB. This test is covering 80% of the drive so HMB is no help here, but that applies to the other DRAMless NVMe drives as well.

Trace Tests: AnandTech Storage Bench and PCMark 10 Advanced Synthetic Tests: Block Sizes and Cache Size Effects
Comments Locked

54 Comments

View All Comments

  • XacTactX - Tuesday, March 9, 2021 - link

    Next time I have to proof read. The performance is very good but it drops off when the drive is full of data
  • DigitalFreak - Tuesday, March 9, 2021 - link

    @Billy All those names are related to space. Polaris and Photon are obvious. Phoenix was a Mars robot. Elpis was the Greek spirit of hope, but is also an asteroid. Maru = Kobayashi Maru? Pablo = Pablo Gabriel de León?
    Maybe someone at Samsung is good with obscure facts :-)
  • Tomatotech - Tuesday, March 9, 2021 - link

    An OK-ish drive especially if it comes down slightly to a better price. Middle of the pack on everything, which is excellent for an entry level drive.

    The one thing I really don't like is the name. The Samsung 980 Pro is one of the best NVME drives ever made. This is ... not. The internet is full of praise for the Samsung 980, often without specifying the 'Pro' because there's been only one model. How many people are going to see that praise, whether it says 'Pro' or not, and assume it's referring to this drive?

    This feels like a sneaky misleading cash-grab by Samsung Marketing for the sake of allowing PC adverts to plaster 'Contains Samsung 980 SSD!' on them.

    To be honest the entry-level 980 doesn't deserve this, it's a decent drive that doesn't need cheating to promote it. If it was labelled 'Samsung SSD 975' that would be perfectly fine.
  • Oxford Guy - Tuesday, March 9, 2021 - link

    Samsung is a particularly slimy company. First, there were the fantasy power claims for the 830 drives, claims people took at face value to promote them in forums. (Samsung, I also recall, was caught paying people to astroturf in forums.) Anandtech's power tests showed that the 830 drives used a lot more power.

    Then, when the planar 840 drives were exposed as being a faulty design, the company refused to replace the drives. Instead, it issued a kludgy work-around — causing the drive to re-write data as many times as necessary to continue to work around the problem (voltage drift from planar TLC I assume). The steady state performance of the 128 GB 840 was so bad HardOCP said it had worse performance than a laptop hard disk, at least in some areas.

    Then, we have the company's decision to use (unfortunately legal, probably) fraud, labeling TLC and QLC drives 'MLC' — as an intentional bait and switch since the industry standard for naming has been MLC for 2 layer flash for many years — many years before Samsung started cheating with the naming. Calling QLC drives MLC is so egregious it shouldn't be legal.

    Now, we have what you pointed out. Samsung loves to trick people with naming lately. Making up power consumption figures is old school, apparently.

    It's anecdotal, but someone in my family bought a 1 TB 860 QLC and it has never been reliable. It even causes the entire machine to slow and have booting problems. The Inland NVME TLC drive I had him replace the Samsung with doesn't have those issues nor does the Inland TLC SATA drive that has been in the machine from the start. He did an RMA and the replacement (which probably was the original drive) had the same problem.
  • Oxford Guy - Tuesday, March 9, 2021 - link

    Also, it's unacceptable to sell expensive televisions with undefeatable ads. That other companies do it also is no excuse.
  • UltraTech79 - Monday, March 29, 2021 - link

    Dude this naming scheme has been around for a decade. If someone is too stupid to see PRO thats on them, not Samsung.
  • Samus - Tuesday, March 9, 2021 - link

    I know this is anecdotal but I can't help but mention the third Samsung SSD in as many months sitting on my bench that failed right out of its 5-year warranty, in this case an 850 EVO. Randomly drops off the bus. These drives have serious longevity concerns and it seems to be a firmware\controller issue, as every drive uses the MEX controller (the 840's and the 850 EVO 1TB) and while I realize that controller hasn't been used in recent products, it's astonishing coincidence they are failing out of warranty obviously doesn't result in repeat customers for Samsung.

    Meanwhile every Crucial drive I've seen in the last decade from the C300, M4, M500\550, MX and BX series have reliability beyond Crucial's metrics and I'm talking at least 200 installed drives throughout our company.
  • DigitalFreak - Tuesday, March 9, 2021 - link

    Both of my 860 EVOs (500GB & 4TB) have been solid. I agree that their pre-860 series drives were not that great.
  • Samus - Tuesday, March 9, 2021 - link

    I agree the newer drives seem reliable, yet the 840 Pro and 850's seemed reliable for the longest time too while recently reports across the web have shown them beginning to fail in the last few years, and Samsung has the most annoying, anti-customer tech support in the industry. Their storage support is outsourced to a 3rd party (storage tech solutions) and with ridiculous warranty qualifications (good luck finding that proof of purchase from 4 years ago) and weeks-long turnarounds. Meanwhile Intel, Toshiba\OCZ and presumably Crucial (we've never had to RMA a Crucial SSD, and most are Micron OEM drives anyway) have fairly normal RMA process from what I've heard. My experience RMA'ing a Intel SSD730 was met with a replacement of a larger capacity 545s and my experience RMA'ing a OCZ ARC100 was met with a free ADVANCED replacement of a same capacity Trion 150. The one Samsung SSD I replaced a few years ago (3 months before the warranty was set to expire) was an 840 EVO and they replaced it with AN 840 EVO. We all know the 840 EVO is one of the worst, if not the worst, SSD Samsung has ever made, you'd think they want to cycle them out of existence yet they seemingly have a stockpile of them somewhere to offload to customers and have them fail again from their write amplification issues.

    I realize my experience is mostly anecdotal but there are plenty of references to customers with Samsung failures, and like Apple, it's hard to actually find negative reviews due to blind loyalty, and the fact the majority of Samsung's SSD business is (was) in OEM's so customers likely didn't even know the failures were Samsung-related in the systems that failed. Probably explains why Toshiba and Hynix, along with the OEM regulars Intel and Micron have erroded their OEM wins and you see fewer and fewer Samsung SSD's in modern PC's. Admittedly this can be attributed to Samsung's premium pricing over every other competitor in an industry plagued with tiny margins.
  • Tomatotech - Tuesday, March 9, 2021 - link

    Interesting reports. I've installed rather a lot of Samsung 2.5" SATA SSDs (various models, purchased used for cheap) for various charities and non-profits and never had any fail yet. Had an inherited Kingston 480GB SATA 2.5" fail - they sent me a replacement. A 1TB Crucial P1 NVMe m.2 (purchased new) started giving problems in one of my Macbook Airs - freezing and crashes - had to move it to my PC desktop where it's worked fine ever since.

    My main m.2 are a pair of ADATA XPG SX8200 1TB nvme, also purchased used, which have given magnificent service in everything I've put them in - PCs and Apple Airs and MacBooks Pro.

Log in

Don't have an account? Sign up now