Sign in with
Sign up | Sign in

Super Talent USB 3.0 RAIDDrive (USB 3.0, 64 GB)

USB 3.0 On A Stick: Super Talent's RAIDDrive 64 GB
By

This is one of the first USB 3.0 thumb drives we received in our test lab. Super Talent offers the RAIDDrive, the self-encrypted SuperCrypt, and the Express Drive. All utilize the USB 3.0 interface. The RAIDDrive is available at 32, 64, or 128 GB, but be prepared to fork out at least $250 just for the 32 GB model. The 128 GB model more than doubles this.

However, Super Talent gives a lot in return. As the name indicates, the drive is built on a RAID stripe set based on two flash memory channels designed to increase throughput on USB 3.0. Super Talent published a performance expectation document in which the vendor talks about capabilities and performance requirements. It claims to reach more than 300 MB/s, but there’s a catch.

First of all, you might need the 128 GB model to ever see this throughput. The 32 GB and 64 GB models are more limited. You’ll more likely see results like those found in the benchmark section of this article: up to 178 MB/s on an X58 platform with an NEC USB 3.0 controller on a PCI Express slot.

Super Talent adds a necessary driver for unleashing the full potential of the RAIDDrive.Super Talent adds a necessary driver for unleashing the full potential of the RAIDDrive.

Super Talent went a step further by offering a modified driver that replaces Microsoft’s mass storage driver for the RAIDDrive. It is the first and most important step to increase performance, but it’s also part of the catch. The driver effectively makes throughput CPU-dependent. On our Core i7-920 reference system, we can't reach those rated 300 MB/s speeds, as the CPU simply is too slow. Super Talent even refers to overclocking as a possible way of increasing RAIDDrive performance.

The Super Talent driver refused to work on our storage test system under Windows 7 Ultimate. The PC wouldn’t even boot anymore after installation. We had no choice but to recover to a system restore point to start over. A separate test run on a Core i7-870 system returned 220 MB/s maximum read throughput, but this didn’t solve the issues on our reference system. Plug-and-play device detection would reset the driver to the Microsoft default and the system didn’t survive a restart.

Be that as it may, the RAIDDrive still provides performance that’s so far ahead of conventional USB 2.0 and eSATA thumb drives that the driver shouldn’t matter too much. There’s little competition, and we doubt that your system infrastructure can provide sustained 170 MB/s write rates to accept data from the RAIDDrive. Even when looking at write throughput, we’re seeing between 2x and 4x increases compared to USB 2.0 hardware. The RAIDDrive runs hot, reaching 55°C after intensive operation, so be sure not to cover it.

Display all 31 comments.
This thread is closed for comments
Top Comments
  • 10 Hide
    cdillon , June 4, 2010 4:20 PM
    JPForumsIn truth, USB3 itself currently has no practical advantage over eSATA.


    USB3 has integrated power and physical port compatibility with the millions of USB1/2 devices out there. Seems like a whoppingly huge "practical advantage" to me.
Other Comments
  • 0 Hide
    spidey180 , June 4, 2010 6:29 AM
    freakin awesome
  • 0 Hide
    wintermint , June 4, 2010 6:46 AM
    Rev your porn engines!
  • 3 Hide
    pinkfloydminnesota , June 4, 2010 6:48 AM
    No readyboost tests?
  • -6 Hide
    zoridon , June 4, 2010 7:30 AM
    Whats the price?
  • -1 Hide
    Gin Fushicho , June 4, 2010 7:42 AM
    So does the Super Talent drive work fine without the drivers? No need to deal with the issues it brings?
  • -1 Hide
    Anonymous , June 4, 2010 7:52 AM
    Ethernet has more latency, wich is the delay when doing stuff. Sending things over high bandwitdth network with latency can be like A:Hi! B:Hm.......................................................Yeah, Hi! Where the bad performance hit comes not from the overall bandwidth but from the delay when doing things in the protocol stack. That's why disk fibre channel with equal speed is much faster than ethernet, and this thinking should go into the comparison of USB too.
  • 4 Hide
    JPForums , June 4, 2010 11:32 AM
    These are fast drives no doubt, but I take issue with the idea that eSATA drives are inherently inferior. The current crop of eSATA flash drives (represented by the OCZ drive) aren't fast enough to push the standard and thus slower than the new USB3 drive (which also isn't fast enough to saturate eSATA). I suspect this has more to do with the flash and flash controller the makers chose to use. After all, there isn't much incentive to produce super fast flash (non-SSD) architectures when the benefits can't be used in the vast majority (USB2) of the market. I'd also like to point out that 64GB of flash allows for more parallel channels than 8GB.

    In truth, USB3 itself currently has no practical advantage over eSATA. Case and point, I am currently running an OCZ Vertex 2 on a Thermaltake eSATA dock at work. The drive runs, for all intents and purposes, at the same speed as it does internally. Enthusiasts covering the topic know this Sandforce based drive runs much faster than the drives in this comparison. I know it's not fair to compare an SSD to a flash disk (thumb drive), but the comparison made here is equally invalid when answering the question presented at the beginning of the article:

    Quote:
    Wondering how fast USB 3.0 runs compared to USB 2.0 and eSATA?


    The article presents a relatively good answer to the question: "Which format currently holds the title of fastest flash drive?" However, to answer the question presented, you should use the fastest devices available for each format. In this case, you've limited the performance of two of the formats in question as there are faster devices available.
  • -2 Hide
    ta152h , June 4, 2010 12:43 PM
    Intel non-graphics PCIe 1x slots are not version 1.1, they are 2.0, but they run at PCIe 1.1 speeds. This isn't a huge distinction, but one nonetheless.

    Because Intel cut corners with their southbridge, you could see organic USB 3.0 implementations faster than those with discrete cards (since they can use two lanes).
  • 3 Hide
    ta152h , June 4, 2010 12:50 PM
    JPForumsThese are fast drives no doubt, but I take issue with the idea that eSATA drives are inherently inferior. The current crop of eSATA flash drives (represented by the OCZ drive) aren't fast enough to push the standard and thus slower than the new USB3 drive (which also isn't fast enough to saturate eSATA). I suspect this has more to do with the flash and flash controller the makers chose to use. After all, there isn't much incentive to produce super fast flash (non-SSD) architectures when the benefits can't be used in the vast majority (USB2) of the market. I'd also like to point out that 64GB of flash allows for more parallel channels than 8GB. In truth, USB3 itself currently has no practical advantage over eSATA. Case and point, I am currently running an OCZ Vertex 2 on a Thermaltake eSATA dock at work. The drive runs, for all intents and purposes, at the same speed as it does internally. Enthusiasts covering the topic know this Sandforce based drive runs much faster than the drives in this comparison. I know it's not fair to compare an SSD to a flash disk (thumb drive), but the comparison made here is equally invalid when answering the question presented at the beginning of the article:The article presents a relatively good answer to the question: "Which format currently holds the title of fastest flash drive?" However, to answer the question presented, you should use the fastest devices available for each format. In this case, you've limited the performance of two of the formats in question as there are faster devices available.


    I agree, it's unclear to me what this article is about. If seems to indicate that it's about comparing technologies behind drives, but if that's the case, it fails horribly. It's obvious the drives themselves have a lot to do with the performance, not the interface, and since they didn't make a real effort to compare drives as identically as possible on eSATA and USB 3.0, it's impossible to conclude anything useful about the interfaces.

    Having said that, we do get a better picture of USB 3.0 vis-a-vis USB 2.0. I know this isn't that useful, but it's better than nothing.
  • 1 Hide
    annymmo , June 4, 2010 1:26 PM
    The ethernet interface is made to operate at larger distances.
    Thus must be prepared to latency and hence doesn't mind encountering.
    Using fiber over distances longer than two meters actually makes sense.
    Hopefully intel Lightpeak will allow us to be able to do very long distances. I can't wait for it.

  • 0 Hide
    Pei-chen , June 4, 2010 2:17 PM
    eSATA is long dead. An external power brick just doesn't make much sense and now with USB3, there is no reason for its existance.
  • 10 Hide
    cdillon , June 4, 2010 4:20 PM
    JPForumsIn truth, USB3 itself currently has no practical advantage over eSATA.


    USB3 has integrated power and physical port compatibility with the millions of USB1/2 devices out there. Seems like a whoppingly huge "practical advantage" to me.
  • 2 Hide
    segio526 , June 4, 2010 5:32 PM
    I chose to do a USB 3.0 setup at home for my backups vs eSata for several reasons. One was the backwards compatibility, so I get an extra USB port on my system (although I do have 2 wasted eSata ports), and Two, I've had way too many issues with hot disconnecting eSata drives (I've lost quite a few drives on different eSata controllers and different eSata enclosures. Yes I was telling Windows to disconnect before actually unplugging). In the end, I'm very happy with 60MB/s with an external 7200 RPM 2TB drive.
  • 3 Hide
    bildo123 , June 4, 2010 5:37 PM
    "The USB 3.0 standard is also referred to as SuperSpeed USB. This was necessary because USB 2.0 was promoted as High-Speed USB several years ago."

    USB 4.0...Ludicrous Speed USB!!!
  • -1 Hide
    WR2 , June 4, 2010 6:47 PM
    I had been wondering how fast USB 3 was compared to USB 2 / eSAT\.
    Thanks for giving us a peek at what's out there.
  • 1 Hide
    tharkis842 , June 4, 2010 8:11 PM
    Wow, can't wait to get my hands on a usb 3.0 mobo.
  • 2 Hide
    jimmysmitty , June 4, 2010 8:15 PM
    bildo123"The USB 3.0 standard is also referred to as SuperSpeed USB. This was necessary because USB 2.0 was promoted as High-Speed USB several years ago."USB 4.0...Ludicrous Speed USB!!!


    "They've gone plaid!!!!"

    hehe....

    For the article, on the second page it had the nominal thourough put in Mgeabytes as Mb/s and it should be MB/s. Should fix it so it doesn't confuse people.

    But USB 3.0 looks promising. Its pretty fast and has just started really. I am sure new enhancements will come to make it faster and better and hopefully get close to the 500MB/s nominal speed.

    Maybe when Lightpeak hits it will be able to do that since light has been known to be able to maximize nominal speeds much easier than copper wire.
  • 0 Hide
    NucDsgr , June 4, 2010 11:00 PM
    An interesting article. Yes, Native USB 3.0 and SATA 6G is worth it once devices eventually become available that can utilize the bandwidth.

    Still for a Thumb Drive to utilize USB3.0 requires special drivers and RAID architecture on a "Thumb Drive". For USB3.0 to become mainstream, the flash architecture within the "Thumb Drive" will have to be significantly revamped. This things look much larger than the typical thumb drive.

    I myself would be better off with the OCZ USB2.0 product for the files I move.
  • -4 Hide
    JOSHSKORN , June 5, 2010 12:31 AM
    Wow. I can see it now. Byebye HDD and SSD, Hello USB3!
  • 1 Hide
    hakesterman , June 5, 2010 5:30 AM
    SSD is the future for Hard Drives, USB3 is for thumb drives and external add-on's. They are both Awsome!
Display more comments