POST A COMMENT

38 Comments

Back to Article

  • fancarolina - Tuesday, August 30, 2011 - link

    The first picture says it all. You can tell they never tested the card with any real world USB 3.0 Devices like eg. those flash drives. When you have to force them in place to connect them. That is bad, you can't help but believe with a bit more testing they could have found better spacing for the ports. So that devices could plug you know plug in. Reply
  • StealthX32 - Tuesday, August 30, 2011 - link

    On the other hand, people rarely use two USB flash drives simultaneously.

    If you do it a lot, then fork up the $2 for an extension cable.
    Reply
  • Googer - Thursday, September 08, 2011 - link

    With 4 ports, i think its stupid to call this anything more than a minor flaw. I doubt anyone will ever have 4 thumb drives attached at any given moment. Persoanlly, I'd rather see them cram an extra closely spaced port in than to have few more widely spaced USB ports. Just like when a wall wart blocks some of my outlets, I grab the trusty old extension cable and free them up. (in this case, a usb3 hub would suffice) Reply
  • rtallmansu - Tuesday, August 30, 2011 - link

    This is not a lack of testing at all. While I agree that it would be nice if the ports were spaced out more for these larger flash drives, you have to understand that there is a specification that HighPoint is following for the distance that two USB ports are apart. This is why all USB ports that you see in either the vertical or horizontal position are always the same distance apart from any manufacture. The problem really lies with the flash drive makers that are building devices out of USB connector specs, because people want larger capacity drives that can’t be fit into the proper form. If you want a larger capacity drive that requires a larger form factor then buy a couple dollar adaptor cable or a drive that comes with one. Reply
  • PyroHoltz - Tuesday, August 30, 2011 - link

    I think your point is a little unrealistic. The card has 4 ports across the back plate. If someone wanted to use two highspeed(read:wide) thumb drives, then they could easily just move to another port. I think the author was making a very specific point while also trying to saturate one leg(side) of the bus. Reply
  • fcx56 - Tuesday, August 30, 2011 - link

    I agree completely, however I feel I should point out that each slot has it's own 3.0 controller, so your conclusion that the author was attempting to push "one leg(side) of the bus" is flawed. Reply
  • JohnMD1022 - Tuesday, August 30, 2011 - link

    I keep USB cables in an assortment of lengths, from 6 inches on up.

    Problem solved.
    Reply
  • mino - Wednesday, August 31, 2011 - link

    How many of those are USB3 ... just noting as USB3 cables are nowhere as cheap as USB2 ... Reply
  • Spazweasel - Wednesday, August 31, 2011 - link

    Thumb drives are not the targeted devices for this product. External drives and drive arrays, video sources, and high-end scanners are. Such devices are attached with cables which will live side-by-side without issue. For people using thumb drives and who want to use the same controller at the same time as their other devices, NewEgg has USB 3.0 extension cables for $4.99. (as a point of comparison, the cheapest USB 2.0 extension cable they list is $2.99, so it's a $2 premium).

    Cost is not an issue.
    Reply
  • Lohkay - Tuesday, August 30, 2011 - link

    Would it possible to verify if you can select from the bios a device connected to that card for booting? Almost every card out there and motherboards uses the NEC chip which doesn't allowing booting off USB3. Reply
  • damianrobertjones - Tuesday, August 30, 2011 - link

    If you have the correct driver you should be able to boot... if not... Windows will ask you for one. Reply
  • ggathagan - Tuesday, August 30, 2011 - link

    Lohkay's question had nothing to do with Windows drivers.

    The question was whether or not a device connected to the Highpoint card would be recognized *in the BIOS* as bootable.

    It's probably entirely dependent on the motherboard chipset.

    Highpoint's manual makes no mention of booting capabilities, and Asmedia makes no mention of it in their description of the ASM1042 controller.
    Reply
  • LauRoman - Tuesday, August 30, 2011 - link

    I can boot from an sd card connected to my usb printer. But then again that is usb 2.0. Reply
  • peterfares - Thursday, September 01, 2011 - link

    That's because your USB controller allows you to boot from connected media. Once again, the OP is asking whether this USB controller card allows you to boot from something connected to it. Reply
  • HibyPrime1 - Tuesday, August 30, 2011 - link

    "The SATA II interface's maximum bandwidth is 3Gbps, and the latest SATA III standard maxes out at 6Gbps. While most devices are incapable of saturating these interfaces, higher-performance SSDs and SSD arrays can sometimes deliver in excess of 1Gbps."

    Almost all SSDs and most hard drives are able to exceed 1 Gbps, 1 Gbps = 0.125 GBps.
    Reply
  • JarredWalton - Tuesday, August 30, 2011 - link

    You're correct on the definitions of Gb vs. GB, but read the statement: "While most devices are incapable of saturating these interfaces, higher-performance SSDs and SSD arrays can sometimes deliver in excess of 1Gbps."

    Very few consumer SATA HDDs will saturate even a 1Gbps link (125MBps), let alone a 3Gbps link (375MBps)--though SATA uses 8/10b encoding, so really it's only capable of 300MBps maximum throughput on "SATA II". Move up to 6Gbps/600MBps and the only devices capable of hitting that right now are SSD RAID sets. But we're talking about USB 3.0, which is a 5Gbps link capable of up to 625MBps (in theory). And we're looking at sequential (i.e. best-case) throughput; for other workloads you'll be more like ~85% of the sequential performance on average.

    That is why he states that high-performance SSDs and SSD arrays can "sometimes deliver in excess of 1Gbps." Good SSDs can certainly hit that mark and exceed it (even on random accesses for some), but if we're talking real-world usage I'd say it's accurate as Zach has it right now. We could say that "high-performance SSDs and SSD arrays can sometimes deliver in excess of 3Gbps" as a more specific statement, and in the future I expect a single SSD will even saturate 6Gbps links. HDDs really need multiple drives and sequential access to break 2Gbps though, and on random workloads they'll drop to <2MBps.

    Anyway, in light of the preceding paragraph where Zach points out that a single USB 3.0 controller with a x1 PCIe link will give 1.25Gbps to each of the four ports, putting SSD performance requirements at that level makes sense. I did however change the "sometimes" to "often", as just about any modern SSD will be able to push more than 1Gbps for sequential transfers.
    Reply
  • HibyPrime1 - Tuesday, August 30, 2011 - link

    I guess I misinterpreted what he was saying then, I assumed he was saying that very high end SSDs (revodrive and the like) and SSD arrays were able to hit 1GBps (which is true), and thought the headroom from 6GBps SATA (which isn't true) was still sufficient. It would be a terrible SSD array that couldn't deliver well over 125MBps :) Reply
  • Jambe - Tuesday, August 30, 2011 - link

    What method (and specific hardware) did you use to hook up the SSDs? Reply
  • Z Throckmorton - Tuesday, August 30, 2011 - link

    Each drive was on its own port (USB 3.0 hard drive docks). I did not have a multi-drive external RAID enclosure that would've utilized only one port on hand for this card's write-up. Edited the article to clarify this useful point. Thanks! Reply
  • bji - Tuesday, August 30, 2011 - link

    Why would you write that? What logic concludes that two drives used in RAID 0 cannot transfer data as fast as the two individual drives? If both drives are loaded with I/O then I don't see why you wouldn't get more total throughput from two drives used individually than in RAID 0, given the (very slight) overhead of RAID 0 but more importantly, the fact that every operation then has the maximum latency of both drives' contributions to each operation.

    And your graphs even bears out the fallacy of your statement; total throughput with two individual drives was 589 MB/s, whereas the RAID 0 configuration only gave 578 MB/s.
    Reply
  • bji - Tuesday, August 30, 2011 - link

    Sentence two should have been written as:

    "What logic concludes that two drives used drives used individually cannot transfer data as fast as the two drives in RAID 0?"
    Reply
  • bji - Tuesday, August 30, 2011 - link

    Obviously I have a typo even in the correction. Man I am tired today. Reply
  • Z Throckmorton - Tuesday, August 30, 2011 - link

    Of course you are correct, you simply misinterpreted what I wrote. Individual implies one. That is, an individual drive isn't as fast as multiple drives in RAID 0. That's why I said individual instead of summed or total. Reply
  • bji - Tuesday, August 30, 2011 - link

    There was no reason to put the drives into RAID 0 if it was not going to make their throughput higher since the purpose of adding devices and changing their configuration was to try to increase total throughput to stress the USB card. Therefore putting the SSDs into RAID 0 was counter to that purpose since it reduced the overall throughput of the devices.

    Also the comment that "Even the two of the fastest SSDs in RAID 0 and two of the fastest USB 3.0 flash drives don't choke the RocketU" further implies that the RAID 0 array of two drives was expected to have higher throughput than the two drives working individually.

    Unless you were trying to test the throughput on one individual USB port, but that's not how the article was written.

    Anyway, I'm being pedantic. It was a fine article.
    Reply
  • ltcommanderdata - Tuesday, August 30, 2011 - link

    It might be useful to include a direct comparison against a motherboard's own USB3.0 ports to show whether bandwidth saturation of PCIe x1 is common and significant in these scenarios. Reply
  • peternelson - Wednesday, August 31, 2011 - link

    Asus suggest that it could be a restriction when using fast peripherals across a single lane backplane connection and I do find this plausible.

    That was the thinking behind their Asus U3S6 expansion card which has some similar features to this product and has been around a while now.

    As well as increasing the link speed and going full duplex, the USB3 standard also increased the current available to devices.

    I would quite like to see some supplementary power connector on the card (perhaps Molex) to avoid needing to draw all that power from the system backplane.
    Reply
  • Radnor - Tuesday, August 30, 2011 - link

    Hi Zach T.

    If you really want to test this controller, just Install a WS2008R2 ( Free Evaluation from MS Technet ) and assemble a Software Raid 0, with 4 decent Pen Drives. Might not be good for booting but some of would love to use a fast drive for alot of things. Or for just the sake of saturating that bus.

    Thanks
    Reply
  • blckgrffn - Tuesday, August 30, 2011 - link

    Seriously, this looks good to go for a nice external disk array without all the fun and expense of a Thunderbolt setup. Reply
  • repoman27 - Tuesday, August 30, 2011 - link

    Well, a single Thunderbolt port with two drive arrays connected to it would yield ~19% more throughput than this card with four individual SSDs connected to it.

    But since the cheapest Thunderbolt drive array is etailing for $949 right now and the only cable is going for $49, you're looking at $1996.00 worth of enclosures and cables. Or you could get four appropriate drive enclosures with cables and the HighPoint card for $205.91...

    81% of the performance for 10% of the cost? Bah! I might need that extra 19% of bandwidth someday... or the ability to attach a high resolution display... or the ability to fit the card into a laptop...
    Reply
  • repoman27 - Tuesday, August 30, 2011 - link

    USB 3.0 and PCIe 2.0 share the same nominal data rate of 5 Gbps and both use 8b/10b encoding thus they are really 500 MB/s interfaces. However, the protocol overhead for USB 3.0 in a best case scenario is still way higher than that of PCIe 2.0 under normal circumstances (at least 20% vs. less than 10%). Therefore one dedicated PCIe 2.0 lane will always provide more than enough bandwidth for one USB 3.0 controller.

    A single SF-2281 SSD connected to a good SATA 6 Gbit/s to USB 3.0 bridge can saturate a single USB 3.0 host controller, as you demonstrated. The numbers you posted for the Corsair Force GT are much lower than the > 400 MB/s which that drive is capable of when connected directly to a SATA 6 Gbit/s controller. They are in fact though, some of the highest numbers I've ever seen posted for USB 3.0 throughput. You've saturated USB 3.0 way before coming close to the marketing bullshit number of "5 Gbps". As controllers and drivers mature and are further tweaked we might see throughput closer to 400 MB/s over USB 3.0, but for now it looks like 326 MB/s is all she wrote.

    I'm not sure why you bothered with USB flash drives for this review, or why you tried to cram them into adjacent ports. It would have been more in keeping with this site to use four separate SF-2281 drives, each with their own SATA 6 Gbit/s to USB 3.0 bridge, connected to each of the 4 ports on the RocketU 1144A to see what this card was really capable of. Also, this could have been used to demonstrate the further impact of USB 3.0 on storage system performance beyond just sequential reads and writes.
    Reply
  • howardC - Tuesday, August 30, 2011 - link

    it appears that the driver for this card at the official HighPoint support site has issues. can someone else confirm if that is the case?

    http://www.highpoint-tech.com/USA_new/series_Rocke...

    thanks.
    Reply
  • alfredska - Wednesday, August 31, 2011 - link

    I'm with many of the other user's comments: "Just use a USB extension cable". I mean really, if you're buying a card with 4USB slots on a single expansion card bracket, you can't expect them to be widely spaced. Reply
  • howardC - Wednesday, August 31, 2011 - link

    i'm looking at the silly thing right now. it's a design flaw.

    it's not about available space on the expansion card bracket. it's really about how they are spaced. if all four ports were evenly spaced and further apart there would not be any issues. but then again HighPoint would have to rethink how they're gonna attach the board to the bracket or put in another way, find another place (or another way) to do it because the mount points and screws and whatnot also took up space.

    this is not exactly an el Cheapo Card and yea, i do expect some 'engineering.'
    Reply
  • ctgoodman - Wednesday, August 31, 2011 - link

    I have one of these cards on order. What I plan to use it for is to transfer a weekly backup offsite to multiple 2Tb hard drives. I will post back my numbers. We are using some SATA3 and SATA6 7200RPM 3.5" drives. I will have each one in a startech USB3.0 dock each connected to it's own USB port. Right now, using a single drive and acheaper 2port controller, I am getting transfer numbers ranging from 108MB/s to 125MB's in real world copy jobs using a single drive. All my files are larger than 20GB each. I agree this review wasn't the most useful. the USB flash drive test was pointless. The SSD test was nice to know maximum throughput but I need large drives and a 7200RPM disk should have been in your test. I'll post back my findings. I will be running three drive simultaneously and will post the throughput numbers middle of next week. Reply
  • buster2003 - Wednesday, August 31, 2011 - link

    Hi Zach, I really like your article. Quick question - how did you RAID those SSD USB devices in Windows 7? I have tried to setup a RAID with USB HDD's but have I have not been able to convert the USB 3.0 HDD's to dynamic in windows Disk Management in my Win7 system. Reply
  • repoman27 - Monday, September 05, 2011 - link

    I was looking at some block diagrams of the ASMedia ASM1042 used on this card, and it only utilizes a PCIe 2.0 x1 connection anyway. This seems to be the case for every USB 3.0 host controller currently on the market, regardless of whether they support 2 or 4 USB 3.0 ports. This card only uses one port per controller and hooks the four PCIe 2.0 x1 connections coming off of the four ASM1042 host controllers to a 4 lane PCIe 2.0 connector via a PLX PEX8609 PCI Express switch. The PEX8609 is actually an 8-port, 8-lane switch, so it should be more than adequate for the task at hand. Since USB 3.0 and PCIe 2.0 both offer raw bandwidth of 500 MB/s, and PCIe 2.0 has lower latency and protocol overhead, a single PCIe 2.0 lane should always be able to keep pace with a single USB 3.0 host controller. The limitation of this card is therefore not at the back-end at all, but right at the front with the performance of the USB 3.0 host controllers.

    As far as this was tested, I have several questions.

    1. What drivers were used and are they (and the ASM1042 in general) capable of UASP, or were these speeds achieved using BOT?

    2. What kind of SATA 6 Gbit/s to USB 3.0 bridges were employed to connect the SSDs, and were they themselves potentially a bottleneck?

    3. What was CPU usage like during these tests, and is there potential for this card to be CPU bound?

    4. Was the PCIe slot used pulling its lanes directly off the CPU or was it connected instead to the PCH?

    5. How does the USB 3.0 connected performance of these drives compare to a direct SATA 6 Gbit/s connection on the same system?

    6. What speeds are achievable when copying data between two SSDs, each connected to a separate port on the RocketU 1144A? How about when they are sharing a single port through a hub?

    7. What testbed was used for benchmarking? (The sequential write speeds for the SSDs were higher than the sequential reads, which I presume was because compressible data was used in conjunction with SandForce SSDs.)

    I came across some ASMedia marketing literature depicting a Crystal DiskMark 2.2 run with an SSD connected to an ASM1042 showing sequential reads of 356.2 MB/s and sequential writes of 229.5 MB/s. Renesas claims that with their new controllers and UASP driver they have achieved 370 MB/s according to ATTO Disk Benchmark 2.46 with a USB 3.0 connected SSD. I feel like it would take a couple 240GB OCZ Vertex 3 MAX IOPS drives and a couple 250GB Intel SSD 510s, potentially in RAID 0 enclosures, to really lean on this thing properly.
    Reply
  • Googer - Thursday, September 08, 2011 - link

    How about posting some benchmarks which compare the highpoint with other usb3 options, including integrated motherboard and PCi-e USB3 solutions from other vendors and other USB3 chipsets too. Reply
  • ruggb - Monday, October 21, 2013 - link

    There appears to be a number of issues with USB3 add ons.
    The boot issue involves the fact that the BIOS must load the USB3 drivers prior to trying to boot. Some do, some don't. My new ASUS MB has a switch in BIOS to bring up all USB or just some prior to booting.

    The other issues (most all of them) apprear to be a result of interference created by the USB3 signals in the 2.4 GHz band. Read this intel white paper
    http://www.intel.com/content/www/us/en/io/universa...

    Since my problems acted like noise I have done and posted the following everywhere I see the problem posted --------- As it is appearing more like the solution every hour it stays up. It woke up this AM still connected. It has not done that in the past.
    =============
    There r a lot of posts about this problem. No answers - until now - maybe.
    I read a post about a problem with a wireless mouse. Someone posted an answer ref to an Intel white paper RE USB3 noise in the 2.4GHz band. The problem was only helped using some copious shielding of connectors and components. This was done on a laptop.

    OK, my problem is on my desktop... after trying every driver update and fw update I could find, I decided my issue with my PCIe to USB acted like it was a noise thing so I pulled it out, disassembled the PCIe connector, lined it with copper foil tape and connected it to a grd pin on the PCIe connector. Then I wrapped the tape around the cable up to the front panel box and connected that foil to the foil in the connector.
    It has not been a long time 2 hrs maybe, but it never stayed up this long from a boot.
    I am hopeful.
    If u try this, other methods and other metal foil may work. I have had this copper tape for a long time. Make sure ONE end it grounded preferably close to the MB. If u ground both ends u will get a ground loop and maybe make things worse. I might cover it with scotch tape to prevent the cable from touching the chassis, though the cable is much stiffer now.
    ===============
    Reply

Log in

Don't have an account? Sign up now