Sign in with
Sign up | Sign in

Results: Tom's Hardware Storage Bench v1.0

Samsung 850 Pro SSD Review: 3D Vertical NAND Hits Desktop Storage
By

After winning an award last year for its 840 EVO, Samsung is ready to follow up with another high-end offering. The company's 850 Pro SSD merges the EVO's familiar MEX controller with 3D V-NAND. Does the combination justify an upgrade, or should you wait?

Storage Bench v1.0 (Background Info)

Our Storage Bench incorporates all of the I/O from a trace recorded over two weeks. The process of replaying this sequence to capture performance gives us a bunch of numbers that aren't really intuitive at first glance. Most idle time gets expunged, leaving only the time that each benchmarked drive is actually busy working on host commands. So, by taking the ratio of that busy time and the the amount of data exchanged during the trace, we arrive at an average data rate (in MB/s) metric we can use to compare drives.

It's not quite a perfect system. The original trace captures the TRIM command in transit, but since the trace is played on a drive without a file system, TRIM wouldn't work even if it were sent during the trace replay (which, sadly, it isn't). Still, trace testing is a great way to capture periods of actual storage activity, a great companion to synthetic testing like Iometer.

Incompressible Data and Storage Bench v1.0

Also worth noting is the fact that our trace testing pushes incompressible data through the system's buffers to the drive getting benchmarked. So, when the trace replay plays back write activity, it's writing largely incompressible data. If we run our storage bench on a SandForce-based SSD, we can monitor the SMART attributes for a bit more insight.

Mushkin Chronos Deluxe 120 GB
SMART Attributes
RAW Value Increase
#242 Host Reads (in GB)
84 GB
#241 Host Writes (in GB)
142 GB
#233 Compressed NAND Writes (in GB)
149 GB

Host reads are greatly outstripped by host writes to be sure. That's all baked into the trace. But with SandForce's inline deduplication/compression, you'd expect that the amount of information written to flash would be less than the host writes (unless the data is mostly incompressible, of course). For every 1 GB the host asked to be written, Mushkin's drive is forced to write 1.05 GB.

If our trace replay was just writing easy-to-compress zeros out of the buffer, we'd see writes to NAND as a fraction of host writes. This puts the tested drives on a more equal footing, regardless of the controller's ability to compress data on the fly.

Average Data Rate

The Storage Bench trace generates more than 140 GB worth of writes during testing. Obviously, this tends to penalize drives smaller than 180 GB and reward those with more than 256 GB of capacity.

Again, I added results from Samsung's 512 GB XP941 over two and four PCIe lanes. Also making an appearance is the 250 GB 840 EVO with RAPID enabled. Those results are in orange, helping put the 850 Pro's performance into context.

The 128 GB 850 Pro sweeps past even the PCIe-based Samsung XP941 over two second-gen PCIe lanes. The larger models are faster still. Only Samsung's XP941 communicating across four PCIe lanes bests the 850 Pros.

Service Times

Beyond the average data rate reported on the previous page, there's even more information we can collect from Tom's Hardware's Storage Bench. For instance, mean (average) service times show what responsiveness is like on an average I/O during the trace.

Write service time is simply the total time it takes an input or output operation to be issued by the host operating system, travel to the storage subsystem, commit to the storage device, and have the drive acknowledge the operation. Read service is similar. The operating system asks the storage device for data stored in a certain location, the SSD reads that information, and then it's sent to the host. Modern computers are fast and SSDs are zippy, but there's still a significant amount of latency involved in a storage transaction.

Mean Read Service Time

Whatever medieval magic that animates the 850 Pros pushes read service times into previously-unseen territory. Without more detail on this drive's tweaks, I have to assume the subtle improvement over Samsung's previous-gen offerings comes from V-NAND.

Mean Write Service Time

Crucial's 1 TB M550 splits the smaller 850 Pros. But the two larger Samsung SSDs break new ground otherwise.

Display all 43 comments.
  • 0 Hide
    iknowhowtofixit , June 30, 2014 8:19 PM
    Hmm, what's next for SATA? SATA4? SATA 3.2 (SATA Express) doesn't look like a long term solution. PCI-E is fine, but SATA is still so convenient.
  • 2 Hide
    MoulaZX , June 30, 2014 8:34 PM
    I 'just' ordered 2x Samsung EVO 120GB a few hours ago, then I stumbled onto this article. Damn it! Damn it! Damn it! Every freaking time I run into this, be it Storage, CPU, or GPU.... -_-
  • 2 Hide
    cryan , June 30, 2014 8:41 PM
    Quote:
    I 'just' ordered 2x Samsung EVO 120GB a few hours ago, then I stumbled onto this article. Damn it! Damn it! Damn it! Every freaking time I run into this, be it Storage, CPU, or GPU.... -_-


    I don't know if this really changes anything for you. Two EVOs are still going to be better than one 850 Pro in [most] every way. But I understand the sentiment!

    Christopher Ryan
  • 5 Hide
    lp231 , June 30, 2014 9:05 PM
    Quote:
    I 'just' ordered 2x Samsung EVO 120GB a few hours ago, then I stumbled onto this article. Damn it! Damn it! Damn it! Every freaking time I run into this, be it Storage, CPU, or GPU.... -_-

    You just ordered a few hours ago. Just cancel your order if you really want this 850 Pro.
  • 6 Hide
    tomfreak , June 30, 2014 9:12 PM
    10yrs warranty, may be finally I have a reason to buy SSD. lol
  • 7 Hide
    g-unit1111 , June 30, 2014 9:42 PM
    Quote:
    10yrs warranty, may be finally I have a reason to buy SSD. lol


    I can guarantee that in 10 years you won't own that drive anymore. :lol: 
  • 3 Hide
    10tacle , June 30, 2014 9:52 PM
    I still have several 8-10 year old drives laying around between 80GB-150GB. I mostly use them as external drives for backing up USB thumb drives and other files that aren't large volume.
  • 1 Hide
    helper800 , June 30, 2014 9:54 PM
    Hoping for some SATA 12gbs (or more) transfer speeds in the coming years.
  • 0 Hide
    razor512 , June 30, 2014 9:56 PM
    Will overclocking the bus that the sata controller is on impact the performance?

    Can you test on an AMD platform which makes it easier to over clock that bus and some of the connected components?
  • -1 Hide
    BestJinjo , July 1, 2014 2:05 AM
    Looking forward to future generations of 3D Vertical Nand on M.2 / M.2 Ultra interface. Too bad SATA 3 is all maxed out and the next generation standards are not yet mainstream for the masses which is holding back SSD performance. As far as this drive goes, it's only slightly faster than MX100 but costs double. I don't think it's worth it. MX100 512GB sounds like a perfect stop-gap until M.2/SATAe drives arrive with 1-1.5TB/sec throughput. Perhaps Samsung will give us 95% of the performance for a fraction of the price in the 850 EVO.
  • 0 Hide
    MoulaZX , July 1, 2014 4:49 AM
    Quote:
    Quote:
    I 'just' ordered 2x Samsung EVO 120GB a few hours ago, then I stumbled onto this article. Damn it! Damn it! Damn it! Every freaking time I run into this, be it Storage, CPU, or GPU.... -_-


    I don't know if this really changes anything for you. Two EVOs are still going to be better than one 850 Pro in [most] every way. But I understand the sentiment!

    Christopher Ryan


    Not quite. One is for my Desktop, the other is for my Father's Desktop.

    For my Desktop, I'll be stepping up from 2x OCZ Vertex 2 60GB in RAID 0. Hope it'll be worth it...
  • 0 Hide
    MoulaZX , July 1, 2014 5:00 AM
    Quote:
    Quote:
    I 'just' ordered 2x Samsung EVO 120GB a few hours ago, then I stumbled onto this article. Damn it! Damn it! Damn it! Every freaking time I run into this, be it Storage, CPU, or GPU.... -_-


    I don't know if this really changes anything for you. Two EVOs are still going to be better than one 850 Pro in [most] every way. But I understand the sentiment!

    Christopher Ryan


    Not quite. One is for my Desktop, the other is for my Father's Desktop.

    For my Desktop, I'll be stepping up from 2x OCZ Vertex 2 60GB in RAID 0. Hope it'll be worth it...
  • 1 Hide
    cpy , July 1, 2014 5:19 AM
    Why does tom hardware looks like website from 1990? This page is narrow as hell!
  • 2 Hide
    Flying-Q , July 1, 2014 5:40 AM
    Quote:
    Why does tom hardware looks like website from 1990? This page is narrow as hell!

    That would be to give more room for the sidebar adverts.
  • 0 Hide
    Novuake , July 1, 2014 6:02 AM
    Impressive, didn't think some of these numbers were possible via SATA3.

    But the limit has now been reached in almost every way.

    Which solution will stick? Anyone care to guess?
  • 0 Hide
    crawlgsx , July 1, 2014 6:14 AM
    Eh, unimpressed. Sure its a nice bump but when I look at how long the 840 Pro has been on the market and compare it to the increase to the soon coming 850 Pro, to say the least it doesn't make me want to run out and replace my 840's.
  • 3 Hide
    cryan , July 1, 2014 7:44 AM
    Quote:
    Hmm, what's next for SATA? SATA4? SATA 3.2 (SATA Express) doesn't look like a long term solution. PCI-E is fine, but SATA is still so convenient.


    I disagree. SATA Express melds the convenience of SATA and pcie performance. With NVMe and Gen 3 PCIe, I think there WILL be much to like.

    Christopher Ryan
  • 0 Hide
    RedJaron , July 1, 2014 8:21 AM
    That much speed for that little power draw? Impressive.
  • -1 Hide
    arneberg , July 1, 2014 8:21 AM
    SATA Express haw to wide Cables as bad as the old IDE Cables M.2 or PCIE cards
  • 1 Hide
    xenol , July 1, 2014 8:23 AM
    Part of me doesn't care about bandwidth. I've had Windows 7 load up much faster (as in the throbber didn't get to finish) on a laptop with a SATA 3Gbps interface than a desktop using the 6Gbps interface. And most of the time file accesses are small, but many.

    I'd rather start seeing IOPS crank up to RAM levels.
Display more comments
React To This Article