Random Read Performance

Our first test of random read performance uses very short bursts of operations issued one at a time with no queuing. The drives are given enough idle time between bursts to yield an overall duty cycle of 20%, so thermal throttling is impossible. Each burst consists of a total of 32MB of 4kB random reads, from a 16GB span of the disk. The total data read is 1GB.

Burst 4kB Random Read (Queue Depth 1)

Random reads at queue depth 1 are where Intel's Optane products shine. Compared to the fastest NVMe SSDs using MLC NAND flash, the Optane SSDs aren't quite an order of magnitude faster, but only because the latency of the NVMe protocol over PCIe becomes the bottleneck. Intel's tiny Optane Memory M.2 cache drive is slightly faster in this one benchmark, but the difference hardly matters.

Our sustained random read performance is similar to the random read test from our 2015 test suite: queue depths from 1 to 32 are tested, and the average performance and power efficiency across QD1, QD2 and QD4 are reported as the primary scores. Each queue depth is tested for one minute or 32GB of data transferred, whichever is shorter. After each queue depth is tested, the drive is given up to one minute to cool off so that the higher queue depths are unlikely to be affected by accumulated heat build-up. The individual read operations are again 4kB, and cover a 64GB span of the drive.

Sustained 4kB Random Read

Adding some higher queue depths to the average shows a small speed advantage for the 480GB Optane SSD over the 280GB model, and the Optane Memory M.2 starting to fall behind the larger Optane SSDs. The NAND flash-based SSDs also pick up speed as queue depths grow, but they need to go far beyond QD4 to catch up.

Sustained 4kB Random Read (Power Efficiency)

Given how thoroughly the Optane SSDs have shattered the record for random read performance, it's not too surprising to see them at the top of the charts for power efficiency when performing random reads. The 480GB Optane SSD is a bit less efficient than the smaller model because it has to power significantly more 3D XPoint memory chips with only a small performance boost to show for it. Compared to the flash-based SSDs, the Optane SSDs are only about 2.5 times more efficient, despite being about 7 times faster. The performance doesn't come for free.

At low queue depths the two Optane SSDs offer nearly the same random read performance. When they both reach saturation at QD8, the 480GB model has slightly higher performance, and is drawing about 0.85W more power—a 13% power increase for a 7% performance boost.

Random Write Performance

Our test of random write burst performance is structured similarly to the random read burst test, but each burst is only 4MB and the total test length is 128MB. The 4kB random write operations are distributed over a 16GB span of the drive, and the operations are issued one at a time with no queuing.

Burst 4kB Random Write (Queue Depth 1)

The random write performance at queue depth 1 of the Optane SSDs is great, but not record-setting. Flash-based SSDs can cache write operations in their DRAM and report the command as complete before the data has actually made it to the flash memory. This means that for most flash-based SSDs the burst random write speed is more of a controller benchmark than a test of the storage itself. The Optane SSDs don't have large DRAM caches on the drive and are actually writing to the 3D XPoint memory almost as quickly as the Intel SSD 750 can stash the writes in its DRAM.

As with the sustained random read test, our sustained 4kB random write test runs for up to one minute or 32GB per queue depth, covering a 64GB span of the drive and giving the drive up to 1 minute of idle time between queue depths to allow for write caches to be flushed and for the drive to cool down.

Sustained 4kB Random Write

With larger queue depths and test durations long enough to defeat any DRAM-based write caching and many SLC write caches, the Optane SSDs rise to the top. With this second round of testing, the 280GB Optane SSD performed slightly worse than the first run, but it's still essentially tied with the fastest flash-based SSDs. The 480GB model is a tiny bit faster than even the previous record from the 280GB model, putting it about 8% faster than the Samsung 960 PRO.

Sustained 4kB Random Write (Power Efficiency)

Without a huge performance lead, the high power consumption of the Optane SSDs takes a toll on their efficiency scores for random writes. They are ahead of early NVMe SSDs and on par with the fastest SATA SSDs, but the best current flash-based NVMe SSDs are substantially more efficient. The Toshiba XG5 prioritized efficiency over peak performance and ends up offering more than twice the power efficiency of the Optane SSDs, while the Samsung 960 EVO has a mere 77% efficiency advantage at essentially the same level of performance.

As with random reads, the performance and power consumption gap between the two Optane SSD 900p capacities widens at higher queue depths. With power consumption starting at 5W and climbing to over 10W for the larger model, the Optane SSDs are in a completely different league from M.2 NVMe SSDs, which mostly top out around 4.5W.

AnandTech Storage Bench - Light Sequential Performance
POST A COMMENT

69 Comments

View All Comments

  • nathanddrews - Friday, December 15, 2017 - link

    Is there any indication that performance would degrade with a lesser CPU or improve with a faster CPU? Just curious if we should expect CPU bottlenecks from these hyper-speed SSDs. Reply
  • Billy Tallis - Friday, December 15, 2017 - link

    Yes, CPU speed can definitely affect latency and consistency and maximum achievable IOPS. Some of those limits were hit with the P4800X testing on a server with lower per-core performance than this consumer testbed.

    At the moment, my inclination is to leave the consumer test suite single-threaded, because consumer workloads don't actually hit the queue depths necessary to go beyond the I/O capabilities of a single CPU core. I don't care too much if the high QD range on some of the graphs doesn't quite reach the theoretical limit of the drive, because that's not the part of the graph we should be paying attention to. (For consumer drives.)

    At low queue depths, interrupt servicing latency can be helped a bit by a faster CPU. But for most consumers, switching from Windows to Linux will do a lot more to help reduce their storage latency. And saving a few more microseconds only matters on Optane; mainstream products won't be this fast for quite a while.
    Reply
  • ddrіver - Friday, December 15, 2017 - link

    I think I'll have to reevaluate my position on Optane. Seems like a solid product. Guess I misread the signs. Reply
  • ddriver - Friday, December 15, 2017 - link

    How can you tell a fake "ddriver" apart? Well, the real one would never mislabel the product hypetane ;)

    It is what it is, and still nowhere nearly "1000x" better as intel claimed. It has its niche strengths, but those offer no tangible benefit to 99.9999% of the consumers out there. I don't recall ever claiming that it will suck, my claim has always been that it will epically fail to live up to the hype, which testing thoroughly confirms. It is not even 10% of "1000x better". So hypetane it is.

    SLC flash can easily match and even bet it in most performance metrics. Unfortunately the industry is not even trying, even mlc is now considered "ultra high end enterprise". Which is understandable, as the workloads that could actually benefit from higher performance are very few and far in between, and for 99% of them using ram is the more applicable and still tremendously better performing solution.

    It is definitely not a bad product on its own. And I would not refer to it as "hypetane" if only intel hadn't shamelessly lied about it on such a preposterous scale.
    Reply
  • tuxRoller - Friday, December 15, 2017 - link

    Interface speeds≠ media speeds

    When will you learn?

    Also, still need a reference for those SLC numbers.
    Reply
  • LordanSS - Saturday, December 16, 2017 - link

    Indeed, Interface speeds are not "media" speeds. I never expected it to work according to Intel's "1000x" claims, but was hoping for a more in-line 20x better from what we currently have, considering first generation product and all.

    And it doesn't even do that. Sorry, for this one time, I am (partially) siding with ddriver.
    Reply
  • lmcd - Saturday, December 16, 2017 - link

    20X better overall is entirely unrealistic. Certain attributes are 20X better. That is all you can really expect when so many things (form factor, power usage, interface protocols, physical interface, etc) are retained from the previous generation. Reply
  • LordanSS - Sunday, December 17, 2017 - link

    20x is 2% (TWO PERCENT) of what Intel claimed when they disclosed XPoint. It's not realistic?

    Intel shoul have kept their mouth shut back then, just like Micron has done so until now. If they can't even do 2% of what they claimed, they're the unrealistic ones.
    Reply
  • tuxRoller - Sunday, December 17, 2017 - link

    Were those claims that Intel made in reference to xpoint the tech, or optane the first gen product? Reply
  • tuxRoller - Saturday, December 16, 2017 - link

    I'm not sure why you are quoting media, but you are absolutely welcome to be disappointed.
    You'll notice I don't care whether anyone thinks Intel over hyped their product only that we still don't know what the actual xpoint (the media, or "tread + xfer + Misc" times, as Handy refers to it) response times look like.
    If AT doesn't allow the below link, search for "xpoint presence in slow lane explained" on the register, or Google "Why XPoint SSDs won't meet original speed claims: A guide"

    https://www.theregister.co.uk/2016/09/29/xpoint_pr...
    Reply

Log in

Don't have an account? Sign up now