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)

The Optane SSD 900P doesn't break the record for QD1 random reads, but only because we've also tested the 32GB Optane Memory M.2, which is about two microseconds faster on average for each 4kB read. The Optane SSD 900P is still about 7 times faster than any flash-based SSD.

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

When longer transfers and higher queue depths come into play, the Optane SSD 900P passes the Optane Memory M.2 and remains more than 6 times faster for random reads than any flash-based SSD.

Both Optane devices more or less level off at queue depths of 8 or higher. The Optane SSD 900P saturates at about 1800 MB/s while the Optane Memory tops out around 1300 MB/s. The Samsung 960 PRO 2TB hasn't caught up by QD32, and doesn't surpass the QD1 random read performance of the Optane SSD until the Samsung reaches a queue depth of about 8.

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 burst random write performance of the Optane SSD 900P is slightly higher than the Intel SSD 750 1.2TB, and about 14% faster than Samsung's fastest.

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 higher queue depths in play, the Optane SSD 900P scales up faster than the Intel SSD 750 1.2TB, leaving the Optane SSD with a 7-10% lead over the Samsung 960s and Intel 750.

Samsung's 960 PROs and the larger 960 EVO all trail slightly behind the Optane SSD's random write performance for queue depths 1 to 4, then the Samsung drives level off and leave the Optane SSD with a substantial performance advantage at high queue depths. The Intel 750 is slightly faster at QD1 and QD2, but saturates at an even lower performance level than the Samsung 960s.

AnandTech Storage Bench - Light Sequential Performance
POST A COMMENT

205 Comments

View All Comments

  • Lolimaster - Saturday, October 28, 2017 - link

    The past is the past, few years ago, a 20GB HDD cost $200 so? Reply
  • btb - Saturday, October 28, 2017 - link

    Does the Optane 900P have support for hardware based Bitlocker encryption?

    Currently I have a motherboard with a TPM, and an SSD with Microsoft eDrive/TCG Opal/IEEE 1667 support, and thus support for hardware based(not software) Bitlocker.

    Would the Optane work in a similar manner, if I use it as a boot drive?
    Reply
  • voicequal - Saturday, October 28, 2017 - link

    Mixed reads & writes are a significant weak spot for SSD performance, where a sequential write workload can degrade a sequential read workload and vice versa. It looks like Optane has completely resolved this (no more bathtub curve). It would be interesting to see a mixed sequential test with QD > 1, so that both read & write requests are in the queue. In theory, throughput could be 2x under 50/50 mixed workloads if Optane is fast enough to saturate the full duplex paths, like the PCIe bus, in both directions. Reply
  • evilpaul666 - Saturday, October 28, 2017 - link

    I ordered a 480GB AIC version from the popular online vendor. I was surprised it was actually available. Seems to be bucking a trend this year.

    Hopefully, Intel ironed out the bugs and there won't be crashes until multiple firmware updates over the next year.

    Anecdotally, I've heard good things about improved UX. I'll find out in a few days. It's replacing an Intel 750 400GB from about two years ago.
    Reply
  • Mikewind Dale - Saturday, October 28, 2017 - link

    This is awesome. But what excites me most is using XPoint to replace RAM.

    I wonder, can we get an approximate simulation of what that world could be like, by making a system with a deliberately minuscule amount of RAM, installing a 32 GB Optane module, and setting the Windows page file to be on that Optane module? I'd be interested to see some benchmarks.
    Reply
  • evilpaul666 - Saturday, October 28, 2017 - link

    There was a demo of a system with only 4GB RAM that was supposed to have had good results. Reply
  • "Bullwinkle J Moose" - Saturday, October 28, 2017 - link

    Would System Start-up be any faster?
    ---------------------------------------------------
    Faster that what? Apples to Bannana's?

    A 35 Watt Dualcore Sandy bridge will boot a fresh install of Windows 10 Fall Crapper Edition (Sept 2017) in 5.35 seconds to a Samsung 850 Pro

    or, the same computer will boot a fresh install of Windows XP-SP2 in 3 - 4 seconds (it varies every boot)

    Then, I've seen people bragging on youtube for booting new 90+ Watt Quadcore machines to Windows 10 on an M.2 drive in 17 seconds

    So, wutz your opinion?
    How fast is fast ?
    Reply
  • cheshirster - Sunday, October 29, 2017 - link

    Those prices are FAKE. Reply
  • CaedenV - Sunday, October 29, 2017 - link

    Come on Intel! Storage isn't what this tech is made for! This was supposed to have faster throughput and act as a RAM replacement, not SSD replacement! Being able to replace RAM and storage with something that is slightly slower than RAM, but the capacity of a large SSD would have huge benefits. Imagine 'launching' a program and all that needs to be done is to flip a flag from inactive to active and your whole program is up and running. No loading from the HDD/SSD into RAM, just activate a section of memory and update the windows registry keys if needed. Having direct HDD/SSD access to the CPU without needing to load into RAM first. These would be huge advantages. But instead Intel saw that it wasnt going to be good enough for that so they released what they had as a way to cash in and make up for all of the wasted R&D on this tech over the years.

    Granted; it is not ALL bad. For consumers this would be like burning money. But for business use this is amazing tech. At my work we have a huge document management system with some ~6 million documents in it, and 200+ users running searches on them all the time. On HDD these searches would take just over a minute. We recently moved the search cache to SSDs which dropped the search time down to ~10-20 sec. With Optane we could lower it to near instant search times. Not going to do it any time soon, but there is absolutely a market for this kind of tech in the business IT world. I just don't understand why Intel is marketing it to gamers.
    Reply
  • Reflex - Sunday, October 29, 2017 - link

    To be fair, the software ecosystem is a decade or more behind the concept of unified memory. Even if this was a capable RAM replacement today, nothing could take advantage of it, and wouldn't be able to for a very long time. Reply

Log in

Don't have an account? Sign up now