Random Read/Write Speed

This test writes 4KB in a completely random pattern over an 8GB space of the drive to simulate the sort of random writes 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.

I've had to run this test two different ways thanks to the way the newer controllers handle write alignment. Without a manually aligned partition, Windows XP executes writes on sector aligned boundaries while most modern OSes write with 4K alignment. Some controllers take this into account when mapping LBAs to page addresses, which generates additional overhead but makes for relatively similar performance regardless of OS/partition alignment. Other controllers skip the management overhead and just perform worse under Windows XP without partition alignment as file system writes are not automatically aligned with the SSD's internal pages.

First up is my traditional 4KB random write test, each write here is aligned to 512-byte sectors, similar to how Windows XP might write data to a drive:

4KB Random Write - MB/s

The OCZ Vertex LE does ridiculously well here, outperforming even Intel's X25-E. The Crucial RealSSD C300 does ok, a bit better than the Indilinx drives but no where near as good as the Intel based offerings. The reason being that the C300's controller doesn't actually attempt to reorganize unaligned writes on the fly. Look at what happens to 4K random write performance if we only write to 4K-aligned addresses:

Not only does the Vertex LE jump in performance to over 160MB/s, but so does the RealSSD C300. The Intel and Indilinx drives also get a small boost, but nothing huge by comparison. Modern 4K-aware OSes write data in 4K aligned addresses, which matches up perfectly with the 4K page size on all of these SSDs. You can also guarantee higher performance in older OSes like Windows XP by re-mapping the LBAs on the fly but that requires additional overhead. These newer drives forgo performance under Windows XP, for significantly better performance under Windows Vista/7 and Mac OS X 10.5/6. It makes sense given that Intel's X25-M was designed back in 2005 - 2007 while these newer controllers were more recent endeavors.

4KB Random Read - MB/s

Random read performance is great on the C300 but a bit lower on the Vertex LE. None of these drives are posting bad figures here though. For the price the Kingston SSDNow V+ doesn't appear to deliver its worth.

Sequential Read/Write Speed Overall System Performance using PCMark Vantage
POST A COMMENT

83 Comments

View All Comments

  • mertl - Thursday, March 18, 2010 - link

    The Sandforce Controller has hardware AES encryption, which could be used with OCZ Toolbox (which is not available).
    It's a self encrypting drive. No CPU Performance wasted with software encryption.
    Why doesn't the Vertex LE support FDE (TCG OPAL Standard like a Seagate Momentus 7200 FDE.4)? The BIOS ATA PWD is passed to the drive, alternatively Pre-Boot Utilities like wave can be used.
    Reply
  • jaggerwild - Saturday, March 13, 2010 - link

    It's kinda sad that they will not at least replace the drive with an equal value SSD, I suggest to anybody considering buying any OCZ product to read through there forum before making a purchase!!! People are reporting waiting seven months for warranties on the Vertex series? Today more than ever people are not getting what they paid for and to even post a review of there product is a slap in our collective faces. But I guess Aandtech is selling out too as they have there advertisement up top of this section LOL!
    Reply
  • NandFlashGuy - Friday, March 12, 2010 - link

    Hi Anand,

    I think your characterization of all IMFT flash as equivalent creates a false idea of quailty equilvalence for readers.

    "It’s uses a standard Indilinx Barefoot controller but is paired with IMFT 34nm flash branded either as Intel or Micron. It’s the same flash that’s used in the X25-M G2s."

    Intel does not sell enterprise-grade or SSD Nand to competitors. It does do more testing on the Nand used inside the X-25M G2s, ensuring it meets a higher quality level.

    While nothing prevents SSD competitors from buying USB-key grade Nand and using it their SSDs, it's incorrect to call it the same Nand that Intel uses.
    Reply
  • ayembee - Sunday, March 07, 2010 - link

    Well, my 200GB LE lasted one week -- picked it up last Sunday, and now it's so dead the BIOS can no longer even find it. That's one seriously expensive paperweight... Reply
  • deBlanc - Friday, March 05, 2010 - link

    Has anyone tried a Raid 1 (mirror) with an SSD and an HDD? I have a customer with a Dell R410 server which has very limited expansion capabilites, only one expansion slot which has a 4 port NIC in it. They are starting to run into disk I/O issues. Short of buying a new server, I was thinking of Raiding a 500 GB OCZ Colossus with a 500 GB Seagate 7200.12, the idea being much faster reads with failover to the HDD if the SSD bricks. I can't afford to buy the SSD to test this out. What do you experts say?
    Thanks
    db
    Reply
  • SeanFowler - Friday, February 26, 2010 - link

    Overclockers UK has confirmed with Crucial that the 128GB version of the drive has a far slower write speed. They've been told that the write speed is up to 140MB/s.

    The details are on page 1 here: http://download.micron.com/pdf/datasheets/realssd/...">http://download.micron.com/pdf/datasheets/realssd/...

    Compare this to 235MB/s sustained write on the LE, and 250MB/s max write, and there's no competition.

    Yes the C300 still has amazing read speeds, but I think the LE is the better all-round performer.

    Unless we find out that the numbers we've been seeing on the net for the LE are for the 256GB version and have been erroneously attributed to the 128GB version as well.

    I had the LE on order, then cancelled it to order the C300, then found out that the 128GB drive has mediocre write speeds. Talk about a mess!
    Reply
  • Beno - Thursday, February 25, 2010 - link

    in Final Words, first paragraph, fifth line.
    Crucla
    Reply
  • bizautomation - Wednesday, February 24, 2010 - link

    I contacted the folks at SandForce for a comment on the un-ceremonious death of the OCZ test drive. Below is my question to them, and their response:

    My email to SandForce:

    Dear SandForce,

    This is more for the blog community, than it is for me personally (although I am considering Enterprise SSDs for our database servers). I don't think it warrants your phone time. If you can simply prepare a statement in reaction to the test drive going bellly up at Anantech's test center, which I can post, I think that would be more than enough.

    Here's the main concern out there:

    1. CTO's like me are seriously looking at these drives to consider putting their databases on (SaaS based applications can benefit from the performance improvements).
    2. We are deathly afraid of the reliability issues, so when initial articles about SandForce controllers came out with the emphasis on reliability, everyone was cautiosly elated.
    3. We see a follow-up story on Anantech (arguably THE source of impartial news on hardware), show that the OCZ drive just "died", and I think it's safe to say, there's significant concern.

    Yes, everyone knows it was a pre-production drive (Anantech makes that clear), but the cat is now out of the bag, and without a proper forensic investigation, I'm concerned this would happen witth a production drive. Hence my reaching out to
    your firm for an explenation beyond "Oh don't worry guys, that was pre-production hardware". As you know, our community is comprised of smart people, and that type of answer just doesn't cut it.

    What say you ?

    Their answer to me:

    Dear Carl,

    We appreciate the inquiry and definitely appreciate your concern. It was appropriate that Anandtech highlight the drive that failed was a pre-production drive. The drive used Beta firmware, which by definition, had known issues.

    Since then, SandForce has released its ‘release candidate’ firmware to our customer base of SSD vendors with the intent for them to run their qualification process on the drives and intensive tests for stability. We have received positive feedback so far. We’ll take any issues that we hear of (and any that we find ourselves) and release a ‘mass production’ version of firmware with the intent for end user customer shipments and usage. We’re confident that with the large number of leading SSD vendors in both Enterprise and Client markets that are qualifying drives based on our silicon and FW, in addition to our own robust testing capabilities, production drives based on SandForce will be of the highest reliability.
    Reply
  • anuraaga - Wednesday, February 24, 2010 - link

    Looking at a Japanese retailer's listing for the RealSSD drives at http://bit.ly/clxvoa">http://bit.ly/clxvoa, the write speeds of the 128GB and 256GB versions differ by around 50%!

    Random 4K Write:
    128GB - 140 MB/s, 30K IOPS
    256GB - 215 MB/s, 45K IOPS

    Out of curiosity, which version was tested in this article? The numbers are close to the listed 128GB specs, so that's what I'm guessing. What could result in such a big difference in write performance in this as well as other drives(DRAM size?)? Or is this just some retailer manipulating the numbers to sell a more expensive part (work in research long enough and you get to know the amazing things possible through the use of normalization)...
    Reply
  • eldiablopotato - Tuesday, February 23, 2010 - link

    Don't suppose C300 vs. Vertex LE review is around the corner? Reply

Log in

Don't have an account? Sign up now