Mac USB3 performance: worth the cost?

Mac USB3 performance: worth the cost?

Summary: Now that some new Macs support USB 3.0, is the flexibility worth it? I test a USB3 to eSATA interface to find out.

SHARE:
TOPICS: Storage
9

The newest Ivy Bridge MacBook Airs and MacBook Pro include USB 3.0 that operate at a nominal 5Gb/sec along with 1 or 2 Thunderbolt ports with a 40Gb/sec cross-sectional bandwidth. That's incredible total bandwidth even before adding in the 6Gb/sec for the internal SATA SSD.

I own 5 eSATA and no USB 3.0 drives, so I checked to see if there were any USB 3.0 to eSATA adapters. The Bytecc adapters had higher user ratings and a lower price than the Belkins, so I bought several to test.

Test philosophy
My test approach differs from the dedicated hardware sites like Tom's and Anandtech. I look for obvious improvement - improvement you can feel - and assuming that, then look at the value: is the improvement worth the cost?

I don't worry about stopwatch timings or specs. Either I see a difference or I don't.

Test setup
A new MacBook Air with 2GHz i7, 8GB RAM and 500GB SSD connected to a Uspeed USB 3.0 7 port hub with a single Bytecc SuperSpeed USB 3.0 to eSATA 3Gbs adapter plugged in, connected to a half-full PopDrive with 2 mirrored 2.5" drives. Then I ran the Blackmagic Disk Speed Test app (BDST) - which is designed to tell video editors how many video streams a device can support - as the benchmark program.

To set the bar, I first ran BDST on the MacBook Air's SSD.

Screen Shot 2012-07-03 at 5.15.24 AM

Impressive.

Then I ran BDST on the PopDrive connected using its USB 2.0 port.

Screen Shot 2012-07-01 at 11.23.10 AM

BDST measured USB 2.0 performance was lower than expected. Previous experience with this drive - and in a year and a half of use, it has been flawless - has been in the 20MB/s range. But this is what BDST found, which may reflect the special requirements of dropped-frame-sensitive video.

Finally, here's the BDST results for the Bytecc eSATA adapter.

Screen Shot 2012-07-01 at 11.19.20 AM

The Storage Bits take
Is USB 3.0 on the Mac worth it? I have 3 answers:

  1. No. The 10Gb/s of USB 3.0 pales beside the 40Gb/s cross-sectional bandwidth of Thunderbolt.
  2. Maybe. That 7 port USB 3.0 hub cost $50. Each eSATA adapter is ≈$18 or $126 to fill the hub. And the eSATA cables are another ≈$40. Spend over $200 to get eSATA connectivity? Only worth it if you, like me, have several eSATA storage devices.
  3. Yes. The convenience of USB 3.0 is undeniable. The performance is significantly better than USB 2.0. And we can expect many more options for 3.0 than Thunderbolt is ever likely to have.

"Maybe" and "yes" win the day for me. But for many Mac users I suspect the answer will be "no" because they don't need anything more than Wi-Fi and USB 2.0 peripherals, let alone Thunderbolt.

If you do decide to start using USB 3.0 with your new Mac, take a moment to read this Apple tech note on USB 3.0. There are some USB 3.0 quirks you'll want to know.

Comments welcome, of course. I bought the MacBook Air, 3.0 hub and Bytecc adapters myself. I won the PopDrive at the 2011 CES.

Topic: Storage

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

9 comments
Log in or register to join the discussion
  • Last 2 Pictures

    Aren't the last 2 pictures, actually the same picture twice?
    lehnerus2000
    • 2 pics?

      Yes. New CMS taking some getting used to. Fixed.
      Robin Harris
  • Eclectic

    "My test approach differs from the dedicated hardware sites like Tom's and Anandtech."
    My research approach is to read all sites ... and only then decide what to test. For example before deciding on the design of my graphics PC (PC note) I looked up ...
    http://macperformanceguide.com/
    ... after which it was plain Adrian Kingsley's Hughes' Build Your Own Ultimate Photoshop PC was positively feeble.

    Coincidentally I have just finished reading ...
    http://www.tomshardware.co.uk/usb-3-uas-turbo,review-32467.html
    ... which discusses the merits of the UAS protocol for speeding up USB 3.0 ports. The bottom line is that only a careful choice of controller and software stack will offer a performance improvement from the typical 125MB/s to 350MB/s.

    Even more coincidentally only yesterday macperformanceguide tested a USB 3.0 external disk on a new macbook pro ...
    http://macperformanceguide.com/Reviews-OWC-MercuryEliteProMini-SSD-USB3.html
    ... also seeing performance approaching the 300MB/s figure.

    My preliminary conclusions then:
    - the ESATA and USB hub Robin used are inferior parts - MAC USB 3.0 is not the bottleneck
    - careful selection of parts will also achieve 300+MB/s from a PC
    Note too that 300MB/s is in the same region as the Pegasus Thunderbolt disk array with 4 drives in RAID 5.

    "I look for obvious improvement - improvement you can feel"
    Why then include 4 benchmark charts? (The last does not look a switch up from USB 2.0 to USB 3.0 but a duplicate copied in error.)
    jacksonjohn
    • Re eclectic

      Should not have assumed readers would know that half-full RAID 1 would be slow. Suspect the disk, not the link, is the limiting performance factor.
      Robin Harris
  • Is PC USB3 performance worth the cost?

    Since your blog is about storage in general and your blog isn't supposed to be an Apple advertisement, can you please answer the question as to whether or not PC USB3 performance is worth the cost?
    toddbottom3
  • PC usb3?

    Disn't test that but expect that most PC users will go with external USB3 drives, not eSATA, and will be pleased.
    Robin Harris
  • This is an IDIOTIC and disreputable article.

    The title has nothing to do with the article, and it is misleading. It's NOT testing performance of USB 3 I/O. And the conclusions have nothing to do with USB 3 performance.

    What's been tested is the performance of a USB 3 port after throwing in a chain of devices to allow I/O over USB 3 for a storage device NOT INTENDED for USB 3. Namely: 1 USB hub, 1 USB to eSATA adaptor.

    The author feels free to draw lazy, specious conclusions without addressing the possiblity of there being affects from the USB hub or, more likely, the USB to eSATA adaptor. I doubt they would let this kind of disreputable testing methodology anywhere near Tom's Hardware.

    The article should have been titled "Is Mac USB 3 performance worth it for my particular idiosyncratic setup?" But it isn't, and that makes this article a misrepresentation, leading a reader to conclude that there is some deficiency in Apple's USB 3 implementation. The author's conclusions have nothing whatsoever to do with USB 3 performance on a Mac.

    BTW, I am not an Apple fan boy - I don't own any Apple products. But I'm outraged by the misrepresentations in this article. Truly shameful.
    cosmolee501
    • Wow. I guess you missed the "MAC" in the title...and the Summary.

      Dude. Really? The author shared a specific setup showing how a PARTICULAR application which might give MAC users some insights regarding USB3. Apply some of your verbosity back at ya, as your reaction is waaayyy over the top for someone that doesn't even own a Mac. THIS WAS FOR MAC USERS, nimrod!
      gahendrix
  • Really?

    Like most Mac Pro users, I bought an eSATA interface card to get something faster than FireWire 800. So I have a number of eSATA drives that I haven't been able to use except with FireWire or USB 2. Now that I have USB 3.0, I'd like to use them as active storage, not just as archives.

    That's what I was testing. True, many PC users won't have that problem, but that's why I specified Mac, not PC. If that real world problem is "idiosyncratic" so be it.
    Robin Harris