Mac Thunderbolt speed tests obliterate USB 2.0, FireWire 800

“Newly published results show Apple’s newly adopted Thunderbolt technology blows FireWire 800 out of the water with data transfer speeds to an external RAID system at 700MB/s,” Blake Steven reports for AppleInsider.

“After the release of Apple’s Thunderbolt cable on Tuesday, early impressions have begun to surface on the Web. AnandTech got their hands on both the $49 cable and the $1,999 Promise Pegasus R6 system and have subsequently stated that they are able to write to the 12TB RAID array at nearly 700MB/s while on a notebook,” Steven reports. “The speed obliterated that of the commonplace USB 2.0, as well as FireWire 800.”

Advertisement: Students, parents and Faculty save up to $200 on a new Mac.

Steven reports, “Macworld also received the new cable with the same RAID system, and their detailed results show Thunderbolt is between 4 and 21 times faster than FireWire & USB 2.0. When compared to both on a 2.2GHz Core i5 Macbook Pro, Thunderbolt could write a 2GB file at 210.5 MBps.”

Read more in the full article here.

35 Comments

    1. The cable is more sophisticated than you might think:

      As noted by Arstechnica and iFixIt, Apple’s new $50 Thunderbolt cable is actually a “smart” or “active” cable that contains circuitry and firmware. Arstechnica explains:
      A source within the telecom industry explained to Ars that active cables are commonly used at data rates above 5Gbps. These cables contain tiny chips at either end that are calibrated to the attenuation and dispersion properties of the wire between them. Compensating for these properties “greatly improves the signal-to-noise ratio” for high-bandwidth data transmission.
      iFixIt tore down the new ThunderBolt cable and found two Gennum GN2033 chips in the connector, one on each side. Additional support chips and resistors were also found for total of 12 chips and “tons” of smaller electronic components.

      1. Your description regarding active cables certainly applies for a new compute element at work that uses a XAUI high speed connection (10 Gbps, I believe) between a Xylinx V5 FPGA to a multicore Tilera Tile64. The XAUI cable includes electronics embedded in the connectors at each end of the cable, and it is also heavily shielded for EMI/RFI. Assuming that the same thing applies to the Thunderbolt cable, It isn’t just a simple crimp or solder job.

      1. Actually it’s very valid. USB 2.0 is in common use. USB 3.0 is not common and there is almost no support for it. Thunderbolt would be replacing USB 2.0 not 3.0. So you’d test it against the thing it’s replacing, not the thing that no one uses.

          1. I’ll grant that it would be nice to see how it stacks up against USB 3.0. It still does make sense to test the new (hopefully) standard against the old one since thats the one it would be replacing.

      2. NOT silly, but lacking. Testing against the interfaces we know well does make sense, but only for positioning Thunderbolt.

        Another, more fundamental reason to test now, is to determine which of the emerging standards offers the best prospects to make it as something you should find on every system that respects itself.

        BTW what went wrong with Firewire in the first place? It started to appear on PCs, but then gradually waned away.
        -Just politics, to break the competition’s plans?
        -Or excessive pricing of FireWire devices? (E.g., another $50 extra if you wanted FireWire+USB).
        -Or was it because it turned out that Firewire connections were too sensitive to cable defects? (I’ve had my share of such problems, in particular when FW devices were needed most, or for backup!!)

        1. What killed FireWire was the ill-conceived change in connector when FW800 was released. They should have found a way to reuse the FW400 connector for its successors, like they managed to do for USB 1,2,3. 
          You’d need a new cable to get the full speed but you could still use the old cable if that’s all you had on hand. The FW400 connector design is very solid and robust but the one for FW800 is fiddly and unreliable, and is highly prone to disconnecting if bumped slightly out of alignment. 

          All the things that Thunderbolt/Lightpeak promise to bring were also in the spec for future FW implementations – ultra high speed over copper followed by a huge jump in speed when optical fibre was integrated in the cable design, ability to carry multiple bi-directional HD video, audio and data streams. 

          If they just hadn’t changed the damn connector, things today would almost certainly be different and FW would rule the world.  

    1. It’s probably because FireWire can transfer data in both directions at the same time. USB cannot. In a normal file transfer (not a Photoshop operation), FireWire 400 is only marginally faster than USB 2.0.

      1. Even in casual use, Firewire 400 is noticeably faster than USB 2.0. People who say otherwise are generally depending on reading the specs rather than actual experience.

        1. Well, I timed it before on my Mac, with a 1GB file transfer (folder with about 1GB of files), and FireWire 400 was only about 20% faster, which to me is “only marginally faster.” I was using the same external drive that has both FireWire 400 and USB 2.0.

          If I was just “depending on reading the specs,” I would have said USB 2.0 was faster.

        2. It varies somewhat depending on the application. USB 2.0 can do OK in burst, but tends to lag in sustained throughput in my experience. It seems to make a difference whether that 1GB of data is a single video file or 1000 1MB files.

          I consider 20% to be fairly substantial, even if ken1w doesn’t. I have also found FW400 and FW800 to be far quicker and more reliable for Time Machine use.

  1. This has me extremely worried, and you should be too. Here’s why: the transmission throughput could be so great that you may actually receive data BEFORE you ever created it. In other words, the Thunderbolt cable is so damn fast that you might find yourself time traveling into the future. You read right. The future. And if that happens, we’re all screwed, because this could well rupture the time-space continuum, bringing about the end of all life as we know it.

    Don’t believe me? Then read these reviews of the Denon AKDL1 Dedicated Link Cable by customers on amazon.com. So great is the speed of this cable that people found that they encountered time travel problems and had to wrap their children in aluminum foil to protect them. Be sure to read these reviews – they’re priceless. Here’s the link:

    http://www.amazon.com/Denon-AKDL1-Dedicated-Link-Cable/product-reviews/B000I1X6PM

    Don’t say I didn’t warn you. See you in the future. Or the past. You choose.

    1. Thanks Bernard, that made my day, after watching the
      Tigers lose to the Mets, I connected to Amazon and
      fell out. The product reviews were hilarious. 🙂 🙂 🙂

    2. You may jest but I do remember experiments some years ago where on a sub atomic level information was ‘supposedly’ received before it was sent proving in the experimenters eyes that time travel was indeed possible. Like cold fusion all quiet on that front since.

  2. Macworld never was something for tech-literate people.

    But a whopping 210.5 MBps == 210.5 MegaBit per second over a 10 GigaBit per second link beats any pessimistic opinion about the skills of Macworld’s writers.

    1. Most conventional drives are still stuck using SATA 2.0 (around 300 Mbytes/sec transfer), although WD is shipping Caviar Black drives with SATA 3.0 (600 Mbytes/sec peak transfer).

      However, their sustained read/write is much lower at only 140 Mbytes/sec.

      Getting 700MB/sec across the RAID pack is pretty good, but I suspect where you’ll see the big performance news is when they start teaming full-blown SATA 3.0 SSDs with 500MB/sec read/write. Two of those in a RAID 0 pack will use 80% of a Thunderbolt bus’ 10gb link.

  3. If Apple didn’t ask for high royalties with FireWire to begin with (they changed it later but it was too late), FireWire should be at 3200 vs 800 today, being that it would have been adopted by more OEMs.

    But ThunderBolt makes up for all that loss time that we could have had faster file transfers…

    Less time to multitask i guess…

    1. The “High royalties” were $1 per port. Although some people come up with mental gymnastics to assert that this would have somehow added $50 to the price of the drive.

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.