Review & Benchmarks: eSATA-II + MacBook Pro
I have to admit I have always been jealous of Mac Pro users with the ability to mix and match components and add hard drive space as needed. About a week ago I was tasked with the job of editing a large number of very large video files, much more than the crowded 160GB drive in my MacBook Pro would allow for. At the time I thought the solution was to get an external USB 2.0 or Firewire 400/800 drive to tackle my space issues. There isn’t anything wrong with this arrangement but I always found myself wanting more speed. The solution? Adding some eSATA-II ports to my Macbook Pro of course.
eSATA-II is rated at a theoretical speed of up to 3.0GB/s but you can expect to achieve real world write speeds of up around 80mb/s. Firewire 800 is rated at 80MB/s but most benchmarks I see have it around 30-40MB/s. USB 2.0 hobbles in around 15-20MB/s. So in the real world eSATA-II is about five times faster than USB 2.0 and nearly three times faster than Firewire800. After testing all of these connections independently I can tell you that jumping to eSATA-II over USB/Firewire really is worth it.
You can buy a two port eSATA-II card for your Mac’s ExpressCard slot (pictured here) for around $50 or more. ExpressCard slots can be found on both 15″ and 17″ MacBook Pros equipped with a PCI-Express bus. I am not partial to any brand, and as near as I can tell, they are using the si3132 chipset, and use the same driver. You may want to check out Apiotek as they have a card that enables RAID, as well as passes on your drive’s SMART status to the OS. Once you get yourself a card simply plug it in, load the driver, plug in your external hard drive, and you should be up and running.
For an external hard drive enclosure I went with a unit from Vantec because of the great price point and simple design, but there are really plenty of better enclosures to choose from. Be sure you buy one that supports eSATA-II and not simply eSATA if you want maximum throughput. You may also want a fan in yours but since I work in a cool environment this didn’t seem necessary. I stuffed in a Seagate 750GB and have been really pleased ever since.
Benchmarks
How pleased? Take a look at the numbers below which I tested with my MacBook Pro. I pitted 4 drives against eachother:
- The stock internal eSATA Hitachi 160GB in the MacBook Pro
- External Maxtor 250GB 7200RPM 8MB Firewire 400
- External Western Digital Passport 160GB 5400RPM USB 2.0
- External Seagate 750GB 7200RPM 16MB eSATA-II (SATA-300)
Sorry I don’t have benchmarks for Firewire800 available but most XBench tests run around 32MB/s for sustained transfers. Also keep in mind that benchmarking in XBench is by no means a perfect test, but all drives were erased and all applications closed for all passes.
Internal Hitachi 160GB 5400RPM 8MB SATA-I (SATA-150)
External Maxtor 250GB 7200RPM 8MB Firewire 400
External Western Digital Passport 160GB 5400RPM USB 2.0
External Seagate 750GB 7200RPM 16MB eSATA-II (SATA-300)
Conclusion
Now it is important pointing out that you will need a newer MacBook Pro with a PCI-Express bus to use eSATA-II. For those of you with 15″ MacBook Pros that came minus a Firewire 800 port this should be really appealing. You can see from the benchmarks above that you should actually achieve better transfer rates with an external SATA-II setup than with your Mac’s own stock internal drive. Another good tip is to be sure and set your scratch drives to this new drive for some of your memory hungry apps like Photoshop and Final Cut Pro to see a real nice speed boost.
For another good comparison of these transfer mediums with benchmarks over other connections, see this thread.
But what about the Firewire 400 setup that scored the same as the eSATA?
Bot
Can you make some more reliable test then Xbench? Something like dd in the terminal? Here are the commands:
time bash -c “dd if=/dev/zero of=/Volumes/Clone/bigfile bs=8k count=500000 && sync”
for writing and
time dd if=/Volumes/Clone/bigfile of=/dev/null bs=8k
Use the appropriate Volumes to get the result from the drive. I get around 65MB/sek from a WD MyBook connected with Firewire 800 to a 17″ MBP. Good enough for me.
These tests don’t make any sense:
First, it’s a given that the internal drive is slower, its 5400 rpm with a smaller 8 MB cache than the drive you used with the eSATA-II;
Second, the it’s also a given that the external USB drive is slower, its 5400 rpm with a smaller 2MB cache (I looked it up) than the drive you used with the eSATA-II;
Three, the drive used with the Firewire 400 is also probably slower overall, since it has a 8MB cache vs. the 16MB cache of the eSATA-II’s drive. But if you notice the Firewire drive has the same over all score as the eSATA-II because it kicks the crap out of it on the Random writes, probably because the eSATA-II’s drive has more area to cover with 750GB of storage space.
Your testing methodology is flawed. With the exception of the MacBook Pro’s internal drive to use as a base comparison, you should have used the same drive to test all three interfaces, in order to eliminate the drive’s performance as a variable.
Also, there are other tests using Firewire 800 and eSATA each using an identical drive (identical to each other), which shows Firewire 800 reaching 50-75 Megabytes per second. This is still well below the 95 Megabytes per second of the eSTAT interface in that test, but much higher than the 32 Megabytes per second that you grabbed from the Xbench’s database (I assume).
Bottom of page…
http://www.barefeats.com/hard91.html
Its also worth mentioning that the theoretical speed of eSATA-II is 3Gbps, not 3GB/s: Gbps is “gigabits per second” and GB/s is “gigabytes per second”, which are two entirely different speeds. Firewire 800 is 800mbps and Firewire 400 400mbps and dead last is USB 2.0 with 480mbps: Mbps is “megabits per second” and MB/s is “megabytes per second”. If you understand how Firewire and USB 2.0 works yes USB 2.0 is still slower despite is extra mbps.
On the other hand the theoretical speed of Firewire is actually 3.2Gbps. Speed that is set to be implemented in Firewire 1600 and 3200, which I believe use the same interface (plug) as Firewire 800.
Its also a shame that eSATA is becoming more popular than Firewire 800, since Firewire does not labor under the limitations that eSATA does, when it come to daisy-chaining devices; Firewire can have 63 vs. one per port with eSATA.
Firewire also supplies power to devices and does not use CPU cycles like USB 2.0 to perform data transfers, which becomes crucial when more than one device is attached. Firewire 800 can also be used with CAT-5 networking cable and carry an IP address. Try all that with eSATA!
In truth, eSATA was design for “consumer” not professional use but like USB 2.0, an inferior standard, it will become popular because the Firewire consortium messed up licensing by charging a dollar per port fee, way back when. But considering that the average PC today probably has 4 USB ports, because it cannot be daisy chained, a dollar would have been a bargain, not to mention the space savings on laptops. Did I mention that Firewire 800 and Firewire 400 can be daisy chained together with a simple $5 cable adapter (not a converter)?
Too bad eSATA is dragging us back down that same damn hole as USB. Firewire could have replaced USB and eSATA, and complemented CAT-5 networking.
@doc logic:
as everyone else has already pointed out, leave the benchmarking to the big boys @ bare feats etc: u obviously are totally unqualified to create testing methodologies.
also: any fool that doesnt know the difference between BITS and BYTES – that represents one order of magnitude difference, for those of you keeping score at home – is someone who should not be allowed to post his _opinion_ on anything let alone something purporting to be _analysis_
@ james: yes & no …
a) yes, it is a shame that apple screwed up every chance for firewire to succeed
the interesting question isL where the hell is the 1600/3200 chipsets?! —
after steve job’s axed the R&D dept at apple upon his return, he had to meekly go out & buy the major independent firewire merchant silicon house in order (which included one of the original firewire architects who left apple); the high-speed chips (not to mention the RF/wifi; and the optical; and the copper MAN version) – none have never seen the light of day!
while it is no surprise the firewire would wither given apple’s withdrawal of support, it is PERVERSELY sick for apple not to provide a substitute – namely e-sata!
b) no, the stumbling block for the royalty price was not $1 … it was $5.
and this was apple’s doing, not the firewire consortium’s fault (or even microsoft’s fault for not shipping firewire drivers or intel in chipping integrated firewire chipsets).
it is the consequent long delay in changing the royalty regime that derailed the oem & general marketplace momentum for firewire.
nearly 100% apple’s fault.
NOTICE THE PATTERN:
* steve (the “genius”) destroys a core piece of apple technology & competitive advantage (just like shutting down the ‘useless’ newton);
* then steve (the “genius”) does a “my bad” and tries to buy his way out of his dufus error (just like the offer to acquire Palm, the ex-apple founded pda);
* then steve (”the genius”) backtracks with a last-minute brilliant defensive maneuver (ipod/iphone)
SAME PATTERN WITH FIREWIRE and lots of other apple tech that has been “steved.”
steve jobs is second only to bill gates as a serial destroyer of intellectual property.
all this “only steve could have saved apple” genuis-talk vastly over-states his talent.
steve jobs 2.0 “success” is more likely a reflection on how god-awful the previous management was rather than a tribute to the singular brillance that is uniquely steve jobs.
hell, even michael dell could have probably rescued apple in the 90’s – given the stunning DNA at apple which any newcomer would have had at his disposal (though even Jobs’ hatchetry did not manage to completly purge).
so: firewire is orphaned but we are left in limbo for years without e-sata ….this is business-as-usual with Jobs 2.0: all we get is half measures (just like the perennially broken Finder TEN (10) YEARS after NeXTstep arrives at apple!) … so many flashes of brilliance almost fatally tainted with compromise.
sigh.
@ zahadum
I think your wrong about the $5 per port cost, it was $1. No wait you are wrong.
Apple To Charge “Per-Port” Licensing On FireWire
January, 15th, 1999
http://www.macobserver.com/news/99/january/990115/firewire.html
As far as the chip sets, with Firewire adoption slowing they’ve been slow to develop FW 1600/3200 but it’s still underway from what I’ve read recently, but I don’t remember where I read it.
As for Steve Jobs I think you have it all wrong. It was the work he did at Next that saved Apple. OS X came from Next and thats been central to Apple’s resurgence. OS X runs not only the Mac but the Apple TV, the iPod, and the iPhone.
Sure Jobs makes mistakes, but he brings true innovation to the computing industry, unlike Bill Gates. Gates’ only accomplish was to make PCs as common as house hold phones via an illegal monopoly.
hey. can someone take a look at my esata-II results and let me know if i am way under?
everything should be sata-2 capable, the pci express card, enclosure and 750gb disk. here are my results:
I guess this site doesnt like html tags..
my results:
Results 31.21
System Info
Xbench Version 1.3
System Version 10.4.10 (8R4061a)
Physical RAM 4096 MB
Model MacBookPro3,1
Drive Type Hitachi HDS721075KLA330
Disk Test 31.21
Sequential 48.64
Uncached Write 32.70 20.08 MB/sec [4K blocks]
Uncached Write 38.02 21.51 MB/sec [256K blocks]
Uncached Read 54.08 15.83 MB/sec [4K blocks]
Uncached Read 145.70 73.23 MB/sec [256K blocks]
Random 22.97
Uncached Write 7.45 0.79 MB/sec [4K blocks]
Uncached Write 44.43 14.22 MB/sec [256K blocks]
Uncached Read 96.14 0.68 MB/sec [4K blocks]
Uncached Read 143.85 26.69 MB/sec [256K blocks]
@ Alexey
Can’t you see they’re in the middle of a pissing contest. How dare you bring up something on-topic.
bf6c525a002c…
bf6c525a002c890b54f3…
I’m a long term Apple user, though am still fairly ignorant when it comes to ports/buses/plugs and so on…
I have a 500gig+ lossless music collection I want to run from an external drive… I was going to get a new macbook pro but the guys at the local PC store brought up the lack of eSata support…. any help here? Should I be holding off for an eSata supported mac? or will firewire do the trick?
answers/pointers more than welcome, though keep it lamens terms please….
Cheers..
I love Apple “Support” for its own creations taking such stupid decisions:
iBook…Firewire YES (10 years of support)… and then suddenly….MacBook… Firewire NO!!!…. mmmm… wait…new MacBook Pro 13…Firewire 800 YES….. PowerBook 15…. Firewire 800 NO….. PowerBook 17 Firewire 800 YES….. wait….MacBook Pro 15 …. Firewire 800 YES….
What the hell are these guys thinking about? How can a company make such stupid marketing decisions!. Can you imagine the same in the PC world? …. USB yes… USB no…. USB yes… USB no…. well… make your own judgment.
I have a brand new Mac Book Pro and it doesn’t have the same card slot as the MacBook Pro I purchased about a year ago, which means that I cannot use the eSata adapter which I previously purchased.
Does anyone know if there is an eSata capability on the motherboard of the most recent Mac Book Pros and how to get at it in terms of ports or adapters?
Also, as a photographer, I transfer loads of digital images from a Compact Flash currently using a USB2 card reader. I currently use ScnDisk Extreme IV Compact Flash with a stated speed of 45 MB/s. Does anyone know of a Firewire 800 or eSata Compact Flash reader, and what the transfer speed would be likely to achieve?
Some people have to learn the hard way. ,
Nato members realise that but governments have spending constraints. ,
Re: Compact Flash USB2 users.
Sanddisk makes a FireWire 800 Compact Flash card reader.
Think I paid about 80.00 from B&H about a year ago for this.
They are a little pricey but worth the difference in uploading speeds vs USB2 Readers for Photographers