Samsung 840 Pro vs. OCZ Vector

Discussion in 'Memory and Storage' started by christer12, Dec 13, 2012.

  1. kleinkinstein

    kleinkinstein Senior member

    Joined:
    Aug 16, 2012
    Messages:
    823
    Likes Received:
    0
    My post follows TB's, got it. Did you not link click?

    Either way... The new firmware. Was released. Today. You good now?
     
  2. jwilliams4200

    jwilliams4200 Senior member

    Joined:
    Apr 10, 2009
    Messages:
    532
    Likes Received:
    0
    Did you not read the post I wrote in the thread you linked to?

    Can you compose a coherent sentence that clearly expresses what you want to say?
     
  3. JellyRoll

    JellyRoll Member

    Joined:
    Nov 30, 2012
    Messages:
    64
    Likes Received:
    0
    #28 JellyRoll, Dec 15, 2012
    Last edited: Dec 15, 2012
  4. jwilliams4200

    jwilliams4200 Senior member

    Joined:
    Apr 10, 2009
    Messages:
    532
    Likes Received:
    0
    I'm not sure if you are joking and forgot the smiley. I'll assume you are serious.

    The sentence of his I was referring to was, "It dropped" and then a link.
     
  5. JellyRoll

    JellyRoll Member

    Joined:
    Nov 30, 2012
    Messages:
    64
    Likes Received:
    0
    The firmware. It dropped.
    Give the guy points for brevity.

    :)
     
    #30 JellyRoll, Dec 15, 2012
    Last edited: Dec 15, 2012
  6. rob.L

    rob.L Junior Member

    Joined:
    Dec 15, 2012
    Messages:
    12
    Likes Received:
    0

    In real-world testing where there are both mixed read/writes, the Vector clearly dominates. Take a look at these links below.


    http://www.tomshardware.com/reviews/vector-ssd-review,3358-7.html


    http://www.pcper.com/reviews/Storag...inx-zeroes-competition/IOMeter-v20060727-IOps
     
  7. jwilliams4200

    jwilliams4200 Senior member

    Joined:
    Apr 10, 2009
    Messages:
    532
    Likes Received:
    0
    IOMeter in your second link is hardly "real-world testing".
     
  8. rob.L

    rob.L Junior Member

    Joined:
    Dec 15, 2012
    Messages:
    12
    Likes Received:
    0
    maybe not so much the 2nd link but the point is the mixed workloaf results which is what happens in the real world. Most benchmarks the 840 Pro win are the non- real- world scenario such as 100% reads.


    Quote from that 2nd link:


    "Regarding why we use this test as opposed to single-tasker tests like 4KB random reads or 4KB random writes, well, computers are just not single taskers. Writes take place at the same time as reads. We call this mixed-mode testing, and while a given SSD comes with side-of-box specs that boast what it can do while being a uni-tasker, the tests above tend to paint a very different picture.

    This test is one of the most grueling to throw at an SSD, and it's also one of the best chances for a well engineered controller to shine. I love it when a new controller technology comes along and creams everything else in the pack, and this time the Vector has done just that. The Vector turns in the best figures in everything save the Web Server test, where it rides along to nearly match the Samsung 840 Pro. Sustaining 55,000 IOPS is not easy to do with our Workstation test, but the OCZ Vector pulls it off nicely."
     
  9. jwilliams4200

    jwilliams4200 Senior member

    Joined:
    Apr 10, 2009
    Messages:
    532
    Likes Received:
    0
    The problem is that there are literally millions of possible mixed read-write workloads. Besides percentage read/write mix, there is the average block size, and how the reads and writes are interleaved (one IO each, 10 IOs each, etc.)

    A lot of the synthetic read/write mix tests (like IOMeter) tend to interleave the read/write IOs at one-to-one (or almost that), and often with small block sizes. Most work I do on my desktop has reads and writes coming in chunks, not interleaved small blocks of reads and writes, and I think this is common for many consumer workloads. So I would not place much weight on those IOMeter tests unless you know for a fact that your workload closely matches it.

    The tomshardware tests are somewhat interesting. The first two on that page are just incompressible data sequential write speed tests (well, the first one is probably partially compressible), so nothing to get excited about there. The last one, the virus scan, is interesting. Too bad they did not tell us more about exactly what was going on during the scan. He mentions that there could be some writes along with mostly reads, but does not give us any details about that.
     
    #34 jwilliams4200, Dec 15, 2012
    Last edited: Dec 15, 2012
  10. Hellhammer

    Hellhammer AnandTech Emeritus

    Joined:
    Apr 25, 2011
    Messages:
    700
    Likes Received:
    3
    It's interesting that they haven't recorded the IO activity to know more about the blocks sizes and queue depths. Would stop guessing and explain why the Vector is faster (can be compared to synthetic tests).
     
  11. christer12

    christer12 Member

    Joined:
    Nov 25, 2012
    Messages:
    31
    Likes Received:
    0
    Some people say that there is a difference performance between a 256GB and 512GB.. So don't compare a OCZ Vector 256GB and Samsung 840 Pro 512GB..

    The performance it's more weaker the bigger SSD is.
     
    #36 christer12, Dec 15, 2012
    Last edited: Dec 15, 2012
  12. jwilliams4200

    jwilliams4200 Senior member

    Joined:
    Apr 10, 2009
    Messages:
    532
    Likes Received:
    0
    I don't think that is true for all SSDs. For late model SSDs like the 512GB Plextor M5P and indeed, the 512GB Samsung 840 Pro SSDs, the specs on the 512GB models are at least as good as the 256GB models. Now that does not prove that there could not be some specific tests where the 512GB models fare worse than the 256GB models (perhaps something involving full-span sustained heavy writes where the larger FTL of the 512GB models would take longer to scan through than the FTL of the 256GB models), but in general I think the 512GB 840 Pros and M5Ps should have similar (or slightly better) performance to the 256GB models.
     
  13. corkyg

    corkyg Elite Member<br>Super Moderator <br>Peripherals
    Super Moderator

    Joined:
    Mar 4, 2000
    Messages:
    25,942
    Likes Received:
    10
    Instead of either or, what would be the result of having both the Samsung and OCZ in the same computer? How would the different controllers get along?

    Play it either way - Samsung for OS and OCZ for data, or vice versa.
     
  14. jwilliams4200

    jwilliams4200 Senior member

    Joined:
    Apr 10, 2009
    Messages:
    532
    Likes Received:
    0
    Unless they are in a RAID, or you are copying data from one drive to another, there is no interaction between SSDs or HDDs in your system.
     
  15. corkyg

    corkyg Elite Member<br>Super Moderator <br>Peripherals
    Super Moderator

    Joined:
    Mar 4, 2000
    Messages:
    25,942
    Likes Received:
    10
    OK - assume a program is on the Samsung, and the OCZ has the data file. What then? IOW, the program on the Samsung needs to save the data on the OCZ. Is that a concern or problem?

    My laptop is currently set up with a Samsung 830 256GB as the OS and program drive. The data from any of the programs gets saved on a Momentus XT 500GB HDD. Could the Momentus XT be replaced with a OCZ Vector?

    Data would be created on the Samsung and saved on the OCZ. Is that a problem?
     
  16. Amnesia1187

    Amnesia1187 Junior Member

    Joined:
    Dec 11, 2012
    Messages:
    19
    Likes Received:
    0
    Data isn't really created on a disk drive. It's created in Ram and then written to whatever disk it's told to. Assuming whatever program you are using lets you choose to save files to drive 2, then they will never exist or have any interaction with drive 1. Even if there isn't a way to choose file locations in your program, there are still ways to go about doing it such as using mklink to connect a shortcut on 1 drive to a folder on another.

    The only scenario where data would go through drive 1 and then be saved on drive 2, is if you were using Intel RST SSD Caching, but there is no reason at all to use RST with another SSD.

    Edit: It did occur to me, there is one other case where data could be written to 1 disk than the other, and that's if the program has an explicit on disk cache, or was something like a torrent program where you specify a location for downloads in progress and another for completed ones, but then its the same basic concept as dragging a file from 1 drive to another which should never have any issues.
     
    #41 Amnesia1187, Dec 16, 2012
    Last edited: Dec 16, 2012
Loading...