The Pegasus: Performance

A single 2TB Hitachi Deskstar 7K3000 is good for sequential transfer rates of up to ~150MB/s. With six in a RAID-5 configuration, we should be able to easily hit several Gbps in bandwidth to the Pegasus R6. The problem is, there's no single drive source that can come close to delivering that sort of bandwidth.

Apple sent over a 15-inch MacBook Pro with a 256GB Apple SSD. This was the first MacBook Pro I've ever tested with Apple's own SSD, so I was excited to give it a try. The model number implies a Toshiba controller and I'll get to its performance characteristics in a separate article. But as a relatively modern 3Gbps SSD, this drive should be good for roughly 200MB/s. Copying a large video file from the SSD to the Pegasus R6 over Thunderbolt proved this to be true:

Apple's SSD maxed out at 224MB/s to the Thunderbolt array, likely the peak sequential read speed from the SSD itself. Average performance was around 209MB/s.

That's a peak of nearly 1.8Gbps and we've still got 8.2Gbps left upstream on the PCIe channel. I needed another option.

Without a second Thunderbolt source to copy to the array at closer to the interface's max speed, we had to generate data. I turned to Iometer to perform a 2MB sequential access across the first 1TB of the Pegasus R6's RAID-5 array. I ran the test for 5 minutes, the results are below:

Promise Pegasus R6 12TB (10TB RAID-5) Performance
  Sequential Read Sequential Write 4KB Random Read (QD16) 4KB Random Write (QD16)
Promise Pegasus R6 (RAID-5) 673.7 MB/s 683.9 MB/s 1.24 MB/s 0.98 MB/s

The best performance I saw was 683.9MB/s from our sequential write test, or 5471Mbps. Note that I played with higher queue depths but couldn't get beyond these numbers on the stock configuration. Obviously these are hard drives so random performance is pretty disappointing.

That's best case sequential performance, what about worst case? To find out I wrote a single 10TB file across the entire RAID-5 array then had Iometer measure read/write performance to that file in the last 1TB of the array's capacity:

Promise Pegasus R6 12TB (10TB RAID-5) Performance
  Sequential Read (Beginning) Sequential Write (Beginning) Sequential Read (End) Sequential Write (End)
Promise Pegasus R6 (RAID-5) 673.7 MB/s 683.9 MB/s 422.7 MB/s 463.0 MB/s

Minimum sequential read performance dropped to 422MB/s or 3.3Gbps. This is of course the downside to any platter based storage array. Performance on outer tracks is much better than on the inner tracks, so the more you have written to the drive the slower subsequent writes will be.

At over 5Gbps we're getting decent performance but I still wanted to see how far I could push the interface. I deleted the RAID-5 array and created a 12TB RAID-0 array. I ran the same tests as above:

Promise Pegasus R6 12TB (10TB RAID-5) Performance
  Sequential Read Sequential Write 4KB Random Read (QD16) 4KB Random Write (QD16)
Promise Pegasus R6 (RAID-5) 673.7 MB/s 683.9 MB/s 1.24 MB/s 0.98 MB/s
Promise Pegasus R6 (RAID-0) 782.2 MB/s 757.8 MB/s 1.27 MB/s 5.86 MB/s

Sequential read performance jumped up to 782MB/s or 6257Mbps. We're now operating at just over 60% of the peak theoretical performance of a single upstream Thunderbolt channel. For a HDD based drive array, this is likely the best we'll get.

To see how far we could push things I pulled out all six drives and swapped in four SF-2281 based SSDs. To really test the limits of the interface I created a 4-drive RAID-0 array sized at only 25GB. This would keep drive performance as high as possible and reduce the time required to fill and test the drives.

Unlike the hard drive based arrays, I had to take the queue depth up to 16 in order to get peak performance out of these SSDs. The chart below shows all of my performance attempts:

Promise Pegasus R6 Sequential Read Performance

With highly compressible data, I managed to get just over 1000MB/s (8021Mbps to be exact) to the 4-drive SF-2281 Pegasus R6. Note that this isn't a shipping configuration, but it does show us the limits of the platform. I'm not entirely convinced that we're limited by Thunderbolt here either - it could very well be the Pegasus' internal controller that's limiting performance. Until we get some other Thunderbolt RAID devices in house it's impossible to tell but at around 8Gbps, this is clearly an interface that has legs.

The Pegasus: Quirks It's Bootable & Power/Thermals/Noise


View All Comments

  • Exodite - Friday, July 8, 2011 - link

    Point being, they have no reason to do so.

    What possible use is Thunderbolt to consumers?

    Had LP been piggybacked upon USB instead it would have allowed for both back- and forwards-compatibility with a huge market of devices and could have slowly permeated said market until it had become a de-facto standard.

    The choice of using mini-DP may well have condemned it to being yet another Firewire.
  • Exodite - Friday, July 8, 2011 - link

    Mini-DP isn't vastly superior to any other display interface.

    The differences between DP 1.2 and HDMI 1.4a are at best a trade-off, with the latter being the better choice for consumers - even disregarding the huge difference in market penetration.

    Computers currently shipping with mini-DP - Apple Macs and those equipped with discrete AMD 5- and 6-series graphics cards.

    Unfortunately that's not the main issue, which is the distinct lack of /displays/ using the standard.

    Displays shipping with mini-DP - Apple Cinema displays. The end. A few other high-end professional displays ship with full-sized DP but you can't argue it's got any significant market penetration.

    Then there's HDMI, which is pretty much ubiquitous in the consumer electronics world, seeing rapid adoption for computer displays and is also used in modern projectors.

    Anyway, that's pretty much an aside - the real gist of it is what I already mentioned regarding USB/LP being the vastly superior choice.

    As for USB connectors only being allowed to be USB connector, that's not a physical limitation - it's a matter of licensing and what they choose to name the standard. A theoretical USB/LP standard could just as well have been named 'USB 4.0' or whatever.
  • Focher - Saturday, July 9, 2011 - link

    Not sure what your point is here. Is it just about the connector type? USB and Thunderbolt are different technologies. TB needed a display standard. DP was chosen because 1) Apple participated in its creation and 2) it isn't saddled with the restrictions and costs that HDMI licensing does. Reply
  • repoman27 - Sunday, July 10, 2011 - link

    Mini DisplayPort IS DisplayPort, just using a smaller connector. The size of the connector is irrelevant to the interoperability of the devices, you simply connect them using an appropriate cable.

    A quick search of Newegg shows displays available from Asus, BenQ, Compaq, Dell, HP, Lenovo, NEC, and Samsung all with DisplayPort connectors.

    ATI/AMD has been shipping cards with DisplayPort connections since early 2008, and DP is native to the GPUs used for every 5 and 6 series device. Intel began including DisplayPort capability with GMA 4500 in 2008 and currently ships it to you in every CPU with integrated HD graphics. NVIDIA is the only major player who seems to be lagging on DP adoption, but there are still plenty of Fermi based cards on the market rocking DisplayPort.

    HDMI was developed for consumer electronics, i.e. televisions and home theaters. DisplayPort was developed for PCs. The lineage is distinct and continues to this day as the two evolve. DP is packet based so that multiple displays can be daisy chained off of one port, which is not a common usage model for TVs. Newer HDMI specs include such home theater relevant features as support for 3D formats, 100 Mbps Ethernet and an audio return path—something that makes no goddamned sense in the PC context. DP originally supported high-resolution displays but not audio, whereas HDMI included audio from the outset but could only drive a 1920x1080 display because that’s all that HDTV required. DP 1.2 has an AUX channel that can be used to provide a USB 2.0 connection to the display over the same cable as video and audio, as well as offering more than twice the total bandwidth of HDMI 1.4.

    USB is a shared serial bus based on a tiered-star topology. It’s great for connecting lots of relatively slow devices that don’t require much bus power or tight timing. When you try to use it for devices that require lots of bandwidth, like high speed external storage, things go downhill fast. Even with a single device connected to a USB host controller, you’re lucky to get throughput equal to 60% of the oft touted “480 Mbps” or “5 Gbps” due to insanely high protocol overhead. Although it’s finally moved beyond half-duplex and added better support for bulk data transfers, the cluster that is USB 3.0 tops out at around 385 MBps in best case scenarios. To create backwards compatibility with 2.0, they merely created Siamese Frankenconnectors, doubling the dimensions of the B connectors in the process, and added more conductors to the cable. How would bastardizing this poor port any further be a good idea? The differences between Thunderbolt and USB in architecture, implementation, and intended use are vast. How would the average user make this distinction if they were somehow cobbled onto the same port?
  • taltamir - Friday, July 8, 2011 - link

    it doesn't have such potential, because it is an active cable that costs 50$ per cable.
    Daisychaining doesn't help reduce the amount of CABLES you need, you still need one cable per device. It just reduces the amount of ports you need.

    So I will stick with my 5 cables for 2$ each instead of 5 cables for 50$ each.

    Now, if they made a thunderbold cable that is passive and cheap as hell, and integrated the controller into southbridge, then it will have the potential to replace all other cables.

    Not having royalties and being an extension of PCIe are very powerful features as you said yourself.
  • HW_mee - Friday, July 8, 2011 - link

    I though the unofficial Apple slogan was "It just works", but after reading the description for using an Imac as secondary display, that slogan seems like a joke. You have to boot up the Imac, own a recent keyboard and press a slightly odd key combination, that does not fit my impression of something that "just works".
    You use the Imac as a screen and Displayport is part of the cable, why could they not just have a on/off switch for the screen and one for the complete Imac, the screen switch can only control the screen and the Imac switch starts screen and computer, if the Imac is on, the screen switch is deactivated, not exactly advanced science.

    Reading the review I also got the impression that Apple have given up on security, is there no password protection or something similar in "Target disc mode"? Can you just buy a Thunderbolt cable for a new Macbook pro and start stealing data from other Thunderbolt equipped Apple computers, just by connecting the cable and holding down t when the "victim" is started?
  • HW_mee - Friday, July 8, 2011 - link

    Replying to myself :-/

    Target disk mode is seriously a horrible feature, from a security view, it even works with FireWire and I can not find any references to any security.

    Remember kids, a login password protects you data, unless the bad guy remembers a 4$ cable.
  • xype - Friday, July 8, 2011 - link

    Uhm, as soon as someone has physical access to your computer (which Target Disk Mode implies), short of encrypting your whole disk (with, say, or any other encrypting software), you can kiss your data goodbye.

    Also, how’s Target Disk Mode any different than an USB key? If anything, the latter is easier to deal with, since you can just plug it into a running machine and off you go. Target Disk Mode might annoy you with stupid things like a Login and whatnot.
  • Penti - Friday, July 8, 2011 - link

    Physical access is physical access. Just set up a EFI/BIOS password if you want false security. If your concerned about your noisy friend with another mac and firewire/thunderbolt-cable. Or whatever. If you like to protect your data, then encryption is not really enough either, but it helps. It does protect against someone stealing your shut down computer. If you have it on, they encryption key is in memory however. When you have physical access it doesn't really matter what the firmware tries to do, passwords and lojack can all be circumvented, and of course you could just remove the drive from the computer when it's not encrypted any way. Computers don't have protection from and are never protect from physical access. You need physical security for that. Reply
  • HW_mee - Friday, July 8, 2011 - link

    It is one thing if the thief steals the HDD/computer, learns a password or hacks the computer using some piece of software, as this often takes time, requires a lucky break or leaves obvious traces.
    Another thing is if someone can just hook up a cable turn the computer on and press T to access everything.

Log in

Don't have an account? Sign up now