Whole-Drive Fill

This test starts with a freshly-erased drive and fills it with 128kB sequential writes at queue depth 32, recording the write speed for each 1GB segment. This test is not representative of any ordinary client/consumer usage pattern, but it does allow us to observe transitions in the drive's behavior as it fills up. This can allow us to estimate the size of any SLC write cache, and get a sense for how much performance remains on the rare occasions where real-world usage keeps writing data after filling the cache.

The SLC write cache in the 1TB SK hynix Gold P31 runs out after just over 100GB of writes. After the SLC cache fills up, the Gold P31's sequential write performance becomes highly variable, ranging from about 1.4 to 2.3 GB/s with little change in character across the entire TLC filling phase. There are no obvious patterns of periodic garbage collection cycles visible at this scale.

Sustained 128kB Sequential Write (Power Efficiency)
Average Throughput for last 16 GB Overall Average Throughput

Despite the variability, the P31's long-term sustained write performance is excellent. It averages out to the best overall write throughput we've measured from a 1TB TLC drive, and in all that variation the performance never drops down to a disappointing level.

Working Set Size

Most mainstream SSDs have enough DRAM to store the entire mapping table that translates logical block addresses into physical flash memory addresses. DRAMless drives only have small buffers to cache a portion of this mapping information. Some NVMe SSDs support the Host Memory Buffer feature and can borrow a piece of the host system's DRAM for this cache rather needing lots of on-controller memory.

When accessing a logical block whose mapping is not cached, the drive needs to read the mapping from the full table stored on the flash memory before it can read the user data stored at that logical block. This adds extra latency to read operations and in the worst case may double random read latency.

We can see the effects of the size of any mapping buffer by performing random reads from different sized portions of the drive. When performing random reads from a small slice of the drive, we expect the mappings to all fit in the cache, and when performing random reads from the entire drive, we expect mostly cache misses.

When performing this test on mainstream drives with a full-sized DRAM cache, we expect performance to be generally constant regardless of the working set size, or for performance to drop only slightly as the working set size increases.

As expected for a drive with a full size DRAM buffer, the P31's random read latency is unaffected by spatial locality: reading across the whole drive is just as fast as reading from a narrow range. And the only other TLC drives that can match this read latency are the two Toshiba/Kioxia SSDs with 96L BiCS4 TLC NAND, but they can't maintain this performance across the entire test.

SK hynix Gold P31 1TB Review AnandTech Storage Bench
Comments Locked

80 Comments

View All Comments

  • MrCommunistGen - Thursday, August 27, 2020 - link

    Truly impressive drive! I'm glad you were able to confirm that the power efficiency numbers held up and weren't the result of some kind of measurement error. When the Platinum P31 comes out I'll probably snag a 2TB model to upgrade my 1TB XG6 in my main ITX rig.

    Interesting that you mentioned the SN520. I needed a 2242 or 2230 SSD for a project I'm working on and was trying to decide between a BG4 and the SN520. I was able to refer to the previous Anandtech 1TB BG4 review but even then, I'm looking at either a 128GB or 256GB drive as that's what's available 2nd hand on eBay and the 1TB drive's performance isn't going to really represent the smaller drives.

    On the other hand I've found exactly nothing on the SN520.

    They're cheap enough so I decided to just buy one of each and test them both and see how they compare.
  • Luminar - Thursday, August 27, 2020 - link

    The 2280 form factor SN500s/SN520s are more common than the 2242s and 2230s.

    I would buy a 2280 SN520 and just Dremel it down to a 2242 form factor. It's been proven to work. As Anandtech wrote in their review, the electronics are only in the first 30mm of the PCB.
  • MrCommunistGen - Thursday, August 27, 2020 - link

    For the heck of it I decided to buy a bunch of different 2230 and 2242 SSDs on eBay to test. I was surprised to find 6 different models, 5 of which are from well-known brands. They all seem to be OEM drives pulled from laptops.

    I skipped any of the no-name, known garbage drives.
  • lilkwarrior - Thursday, August 27, 2020 - link

    Why was the 970 plus in the benchmarks, but not the 970 Pro?
  • MrCommunistGen - Thursday, August 27, 2020 - link

    If you really want to compare the performance of the P31 and the 970 Pro, you can look up the results in the "Bench" section of the website:
    https://www.anandtech.com/bench/product/2627?vs=24...

    To answer your question directly:
    I'd imagine that's because for a lot of client workloads the performance of the 970 Pro is relatively similar to the 970 EVO Plus. The Pro definitely has lower latency, particularly write latency -- but again -- in most client workloads that isn't going to translate into very noticeable differences.

    However, the Pro is SO much more expensive. Realistically, someone shopping for a fast consumer SSD who is looking at the P31 isn't going to also be considering the 970 Pro.

    FWIW, at the 1TB capacity the 970 Pro seems to be selling for ~$320, the 970 EVO Plus for ~$190, and the P31 for $135.

    If the idea is to compare to the highest performing non-volatile storage available today, that's probably what the Optane 905P results are for.
  • PaulHoule - Thursday, August 27, 2020 - link

    It is as if Taiyo Yuden started making writable DVD's under its own name.
  • nirolf - Friday, August 28, 2020 - link

    Ha ha! Good one!
  • jyotaro - Thursday, August 27, 2020 - link

    Any plans for a 500gb review of this product?
  • Billy Tallis - Thursday, August 27, 2020 - link

    Not at the moment. I'll try to get them to cough up a sample of that one when the Platinum P31 is released so I can compare across the full range of capacities, but I don't know how likely it is that they'll agree. This is still a pretty new relationship between us and SK hynix PR.
  • ozzuneoj86 - Thursday, August 27, 2020 - link

    Maybe I missed it, but does the high efficiency of this drive translate to significantly less heat output? Seems like it should. Some kind of thermal test would be useful for choosing a mobile SSD especially.

Log in

Don't have an account? Sign up now