Random Read/Write Speed

The four corners of SSD performance are as follows: random read, random write, sequential read and sequential write speed. Random accesses are generally small in size, while sequential accesses tend to be larger and thus we have the four Iometer tests we use in all of our reviews.

Our first test writes 4KB in a completely random pattern over an 8GB space of the drive to simulate the sort of random access that you'd see on an OS drive (even this is more stressful than a normal desktop user would see). I perform three concurrent IOs and run the test for 3 minutes. The results reported are in average MB/s over the entire time. We use both standard pseudo randomly generated data for each write as well as fully random data to show you both the maximum and minimum performance offered by SandForce based drives in these tests. The average performance of SF drives will likely be somewhere in between the two values for each drive you see in the graphs. For an understanding of why this matters, read our original SandForce article.

Desktop Iometer - 4KB Random Write (4K Aligned) - 8GB LBA Space

Low queue depth operation isn't going to show any advantage on the RevoDrive 3 X2. This is to be expected. I include these results to point out that for the majority of desktop users, you won't see any benefit from a 4-drive RAID-0 or the RevoDrive 3 X2. I already talked about how most modern SSDs deliver similar real world performance in our last SSD article. The RevoDrive 3 is no exception.

Desktop Iometer - 4KB Random Write (8GB LBA Space QD=32)

What happens during periods of intense IO activity however? The RevoDrive 3 X2 excels. With incompressible data the RevoDrive 3 X2 is over 2x faster than the 240GB Vertex 3 and with compressible data we're almost at 800MB/s for a single PCIe card.

Note that OCZ's specs for the RevoDrive 3 X2 promise up to 200,000 IOPS, however we're only seeing around 180K IOPS in our QD32 test. What gives? In order to hit those sorts of numbers you actually need to run in a multithreaded/ultra high queue depth configuration (two threads + QD64 in each case). If you actually run in this configuration but hit 100% of the LBA space, a reasonable workload for a high traffic server you'll get numbers similar to ours above (766MB/s vs. 756MB/s). If you limit the workload to an 8GB LBA space however you'll hit the 200K that OCZ advertises:

OCZ RevoDrive 3 X2 (480GB) 4KB Random Write Performance (IOPS/MBps)
  QD=3 QD=32 QD=64
IOPS 52131 IOPS 184649 IOPS 202661 IOPS
MB/s 213.5 MB/s 756.3 MB/s 830.1 MB/s

Desktop Iometer - 4KB Random Read (4K Aligned)

Low queue depth random read performance is nothing to be impressed by.

Sequential Read/Write Speed

To measure sequential performance I ran a 1 minute long 128KB sequential test over the entire span of the drive at a queue depth of 1. The results reported are in average MB/s over the entire test length. These results are going to be the best indicator of large file copy performance.

Desktop Iometer - 128KB Sequential Read (4K Aligned)

Desktop Iometer - 128KB Sequential Write (4K Aligned)

We're still going through the unimpressive tests here - we see some benefit for larger transfer sizes, but the real advantages come when you start loading up the RevoDrive 3...

Introduction AnandTech Storage Bench 2011
Comments Locked


View All Comments

  • SonicIce - Tuesday, June 28, 2011 - link

    lol last sentence >evironment
  • mmaestro - Tuesday, June 28, 2011 - link

    You can do secure erase, is there any sort of garbage cleanup tool they also supply for maintenance? I know Intel supplies these with their SSDs for if TRIM is unavailable, and that seems like something which you ought to have to make up for this shortcoming.

    It's a shame OCZ started you all off with such an expensive offering. I'd be far more interested in the performance of the Revodrive 3 (no x2) 120gb or 240gb.
  • don_k - Tuesday, June 28, 2011 - link

    I started reading this review, saw the specs and that they removed the pci-x bridge and said to myself "This thing is gonna hit 2GB/s sequential." Close.

    Man is this thing fast. It is very interesting especially on a enterprise environment and that is due to the data workload, nothing else.

    I personally have been running the original 120GB Revodrive (2 Sandforce 1200 controllers) and I've had absolutely zero reliability issues in 24x7 running on a linux box that is my personal workstation. Bought September 2010 so almost a year old now. I'd call that 'good', so far. Tentatively.

    So what is the driver used for this under Linux? Is there one? If Ocz have made changes to firmware then they need to release a driver for linux no?
  • skrewler2 - Tuesday, June 28, 2011 - link

    This thing probably wont be used by consumers. Why not at least include some more relevant DB benchmarks? tpcc would be a good place to start.
  • FunBunny2 - Tuesday, June 28, 2011 - link

    me too.
  • hurricanepkt - Tuesday, June 28, 2011 - link

    OCZ uses a pretty confusing model number schema. Are either of these the revodrive 3?
    It seems very difficult to tell
  • Conficio - Wednesday, June 29, 2011 - link

    Dear Anand,
    first off all for doing what you do her. I have learned so much from your SSD (and other technology coverage).

    As you really have this powerful voice in the industry and have always used it in favor of the users and consumers I'd like you to shed more light on supported platforms of SSDs as part of your reviews and tests.

    I have at work a Mac Pro with a couple of 120GB OCZ Vertex SSDs which got corrupted by a power outage Monday morning. Not a big deal as the data on them was of temporary nature.

    However, my attempt to use that opportunity and secure erase the drives and may be update the firmware failed miserably:
    * OCZ does not make it easy to find the tools
    * OCZ makes you download a package for all platforms which then you have to select (w/o a guide) the "right" model and burn the tools on a CD
    * The Insturctions require to set BIOS modes for ATAPI, so are PC only
    * The required BIOS mode (IDE) is again PC only
    * Finally I had to use PC Laptop to to boot the CD and it painfully failed to boot because of some CD driver for the CD drive missing.

    So to actually restore the drive to factory conditions one needs to have some very specific hardware (if it works at all then, never got to that point).

    In my book this is clearly poor customer support and platform support. I'd like to see more details about the actually supported platforms for this new technology including the secure erase tooling (or other tooling that OCZ does provide, like some form of garbage collection tool).
  • neotiger - Wednesday, June 29, 2011 - link

    1) All the benchmarks are for desktop apps. I really doubt people would be buying PCIe SSD for "Starcraft II & WoW" (!!) Far more likely people would be using it for MySQL, Solr, Hadoop, VM. Can we get a benchmark based on those apps?

    2) You only tested random read at low queue depth and simply concluded that "Low queue depth random read performance is nothing to be impressed by." What about high queue depth? Random read IO at high concurrency and high queue depth is just about the *MOST* important metrics for server SSD uses -- yes even more important than random write IO. Yet you completely skip over that. Any chance you can update the benchmark to include results at high concurrency and high queue depth?
  • alphadog4 - Thursday, June 30, 2011 - link

    One way this review would be nice is if we built 4x80GB striped RAID with a SAS controller setup and compared. I'm just not sure what I am gaining here.
  • chadwilson - Thursday, June 30, 2011 - link

    I really really really would like to see some server testing done with server class devices. Specifically I'd like to see the testing done on a modern *nix kernel, as this type of device will likely see usage in this environment.

Log in

Don't have an account? Sign up now