Now that both the MacBook Pro Retina Display and MacBook Air ship with USB 3 I was curious to find out which was actually faster? Going with my existing Firewire 800 drives and the Thunderbolt to Firewire adapter or USB 3 drives. I did searches online and really didn’t come up with much in the way of current data or test. Most of what I found was comparisons to USB 2 or tests done before Macs had USB 3 built-in. Since I couldn’t find the data I wanted, I decided to do my own tests.
Making it fair
I wanted this to be as fair as possible. That meant that I didn’t want to use two different drives. Ideally I would want to test the same drive that has both Firewire 800 and USB 3 built-in. That configuration isn’t as easy to find as you would think on a portable drive. I found two: LaCie’s Rugged 1TB (which I’ve used in the past with no issues) and the Oyen Digital Mini Pro. Since I was looking to have a smaller enclosure, I decided to give the Oyen Digital a shot. It was also a few bucks cheaper.
For the benchmarks I went with the Free App – BlackMagic Disk Speed Test.
Get it here
I recorded my results in the video above.
Thanks, Terry. I like the real world test of writing the 16GB folder. I guess the real world read test would be about the same.
I’d love to see a comparison between eSata (via a Thunderbolt adapter) and Thunderbolt itself
I would also be curious what sort of numbers are achievable when the external drive is instead a fast ssd drive
The limitation for thunderbolt is the drive itself since it is essentially a PCIe connection fast enough for a graphics card. Even Internal SATA is only 6Gb compared to Thunderbolts two channels of 10 and SSDs can go far beyond eSatas limits.
That’s really helpful, Terry. Thank you.
Terry–might the real world test have been compromised a little by the volume you were pulling data from? That may have been a bottleneck…
It was an internal SSD.
Wow, I thought 1TB SSD drives were still $1100+ without an enclosure
LaCie’s Rugged 1TB – linked in article – $189
Oyen Digital Mini Pro. – linked in article and used in test – $169.56
960GB Mercury Electra MAX 3G SSD – MacSales – $1129 (without enclosure)
http://eshop.macsales.com/item/OWC/SSDEX3G960/
Apologies, you were talking about the drive in the Mac (I misread)
USB 3.0 is usually more than twice as fast as FireWire 800. If you were using an SSD inside the enclosure you would see a much larger difference in performance.
The USB 3.0 chipset can make all the difference too. I have seen USB 3.0 speeds as much as 5x faster than FireWire when using an SSD for testing and when the USB 3.0 design is optimized for SATA 6G. Performance very close to Thunderbolt will be available with quality USB 3.0 enclosures in the future – for a fraction of the price.
Terry, This is great for just copying data, but how about video editing. I seem to remember that firewire had an advantage over USB in that firewire sends data both ways at the same time. How does that figure into USB3?
This is my concern as well. I am also still trying to afford to make the switch from my old Mac to a new Macbook Pro – So I’m thinking about getting one of these drives with BOTH capabilities so that my current MAC will be compatible, and it will also be compatible with a new one MAC so I don’t just invest in the older technology…
This is my concern as well. I mean how about video editing
I think the thing to remember here is that Firewire is going Away.
No more card readers in FW 800. only on ebay.
USB 3.0 will be able to handle faster loads & writes with SSD drives as they are coming down in price. & FW would not be able to get the best performance from the SSd anyway
I have always loved FW, but I love spending less time on the computer . & more time with my Family. So 3.0 here I come.
Very well articulated testing. I have always had the big hunch that in ‘real world’ performance, that it wouldn’t be significantly different. Nice going and thanks.
Steve Douglas
The test comparison I would like to see is how USB 3.0 and Firewire 800 compare when you string multiple devices together. In the past this is where Firewire had the real advantage.
So this flimsy connector is a usb 3 what a step backwards… way to easy to break never mind the speed if your drive is broken because of the built quality.
@Jojo – +1. I would also like to see this test, although I reckon USB 3.0 may have the edge overall.
Thanks for doing the test and explaining what’s done is great but personally I’d rather see that written up than have to sit through an 8 minute video where we watch you mounting and unmounting drives and waiting for the speedtest to happen. I can digest information a lot faster when it’s written and time is a precious commodity. And, yes, I just spent time writing this (not much, though as I type fast) but I mention it because if all your information is delivered in youtube videos then you lose me and others like me as an audience (fwiw).
I like the video very much. the information engages more sense and stays with me longer.
thank you.
lily
I liked the video as well.
Agree. I’ll search elsewhere.
Thbaks for doing this. Exactly what I needed and well done.
Hi Terry,
Seeking an answer to this same burning question ever since I purchased both a mid-2012 MacBook Pro and a Lacie Rugged 500GB FW/USB3.0 drive, I came across your excellent “real-world test” and, while the differences weren’t very dramatic, the answer…and the method by which you demonstrated the comparison…was timely, useful, and very much appreciated. My hat’s off to you for job well done. Many thanks, Terry.
Regards,
Mike
Hi Terry,
a good initial test comparing the two technologies – thank you.
With regard to the real world, I did not see how many files and the sizes of the files in your 16GB folder. I would guess this would impact on the speed as the OS would be allocating and polling the source and destination drives more often for a lot of files than if it were one 16 GB video file for example.
So allow me to suggest more real world comparison tests:
One large video file (or BPAV folder with one large clip) 15GB – 16GB.
A 16GB folder containing a mix of files, perhaps jpg’s, psd’s, mp3’s, and mp4’s, which are perhaps around the 5 – 50 MB each. My guess is a 16GB folder containing 1,000 or 3,000 files will take a longer duration to copy than a folder with only 1-5 files.
thanks for this. One more thing I’m curious about is CPU and memory usage during these tests. Firewire supposedly uses less system resources than USB, and I’d love to see if this is the case. It’s more important in real world use, because we may get a file copy started, then perform other tasks in the foreground. The point might be moot with multicore desktops with 8gigs of ram and up, but I’m still curious.
agree, I remember USB vs Firewire favoured to firewire because the CPU had to do all the transfer work on USB devices. Firewire managed that by it’s own chips, leaving your cumputer faster while transferring data
Great comparison Terry. Just what I needed to see too! Thanks!
Thank you Terry, that was well done and impossible to find on other sites.
BTW, I’m a different Dean than the previous poster (not a lot of Dean’s around!).
http://terrywhite.com/firewire-800-vs-usb-3-which-is-faster/
Interesting results. Benchmark test showed USB 3.0 wins by a mile. Real world copying from the internal drive to the external drive showed USB 3.0 by a hair (closer than the Obama vs. Romney polls). Call it a dead heat. What is going on? FireWire is full duplex while USB is half duplex. Copying a file requires handshaking, so the back channel sees a small amount of data being transferred, too. While the USB forward channel is waiting for the USB back channel to free up the half duplex link FireWire keeps on working away at its sustained full speed. Nice to have a two lane road in the real world. What would have been very interesting would to time copying a file FROM the external drive BACK to the SAME external drive, where full duplex Firewire should be significantly faster than half duplex USB, I suspect.
Sent from my iPad
Now I test on my external HDD WD Caviar Blue.
For HDD about Firewire 800 vs USB 3.0 it same fast.
Because HDD has a limit to read – write.
if use SSD it deference sure.
Uh….what?
??????????????
Interesting results, especially the real world test. I can think of a few alternative testing scenarios where CPU loading and simultaneous read/write could produce results more favorable to Firewire. Your research shows that USB 3.0 is faster than Firewire, but not always enough to be noticeable.
It would take an extreme scenario to make USB 3.0 look like a bad choice, although I have no doubt it can happen, possibly on a Mac Mini that someone has pressed into service as a file server with multiple users and simultaneous read/write. At that point, Ethernet becomes the bottleneck so maybe it doesn’t matter.
Finding this video a little late .. But you’re using a MBP that has USB 2.0 ports. I conducted my own tests for work purposes transferring 100GB of RAW 5K 25fps RED Epic footage from an external SSD onto another external SSD via USB 3.0 using a 2012 retina MBP and the transfer was 7 minutes on the dot. I then repeated the same transfer using a 2011 MBP with Firewire 800 and the transfer was just under 30 minutes.
I assure you Jack that my MacBook Pro 15″ Retina Display does have USB 3 ports.
did the 2011 and 2012 have the same harddrive?
Very cool video and very informative. Thank you. Reading the responses above, I decided to do my own little test with Black Magic. Results confirmed your overall analysis. USB 3.0 edges Firewire 800. Barely.
Results:
USB 3.0— Write 256 / Read 501
Firewire 800 — Write 250 / Read 502
I was testing on G-Drive Mini 7200 RPM. And that’s a huge factor if you are using the drive for Video. Go with 7200 RPM over 5400 RPM.
I would say, use whatever is convenient. Personally, I use Firewire 800. Why? Because one USB 3 port is taken by my mouse wireless transmitter and the other USB 3 port is used by my video capture device. I also give a big-time agreement with Mr Schmidt above— what’s up with the 3.0 connector?
Anyhow, my two cents. Once again, Thank you for fantastic and highly informative video. I now have a new bookmark in Safari. 🙂
fantastic – answers my Q’s re concerns over no fw ports on some of the news macs
Terry, that was an excellent comparison and food for thought. I myself was conducting these tests but not with USB 3. I was more in line with FireWire 800 and different drives in different enclosures which included an SSD 120gig drive. I started thinking about the comparison between FireWire 800 and USB 3 and did a search online and your website with the info came up. Great job and thanks.
Just discovered your column and I will be following your blog for more tech insights. Excellent test and commentary on these two interfaces. Especially nice to have the benchmark vs. real work sample. I have firewire 800 drives and will likely purchase the adapter regardless but this really helped me make a decision on what type of drives to purchase going forward.
I’m putting together a system now, and this exact question occurred to me. Thank you so much for providing a straight forward video test that clearly covers all the bases…Now I just need to research what this thunderbolt drive you mentioned is! lol
Thank you for this test!
Overall you did a good job testing the transfer this way. It is interesting to see real world figures instead of just ‘800 mb/s’ , ‘5 gb/s’ etc. I see one remaining issue in the method you used however. Different partitions on a disk might result in different speeds, because the transfer rate the hard disk is able to reach depends also from the physical position of the heads on the platters. On the outside of the platters the disk reaches a higher transfer speed. The circle is bigger, the path is longer, so you can put more data on it.
Firewire is a dying technology, there is no development, ports are vanishing and the price is not competitive. The question is, not even whether USB 3 outperforms Firewire in every respect, but whether it is a usable alternative.
My new favorite storage strategy is rather simple:
– one thunderbolt drive for projects I am currently working on
– cheap storage for the rest. USB 3 seems to do the job.
A very interesting test would be esata over thunderbolt. This should be cheap, have high performance and put a low burden on cpu.
This test is pretty much nonsense. You should do this with something that can actually handle the extra bandwidth. This is a “Blackmagic disk speed” test. Not a bandwidth test. Come on Terry. I have seen thunderbolt move a 7GB file to a RAID in less than 30 seconds. This test of yours is pointless. Not trying to be rude but it kinda annoyed me because it’s just straight up misinformation and I have watched your videos on editing many times.
Not a fair comparison! You failed to state whether you were using a SSD, 5400rpm, 7200rpm, 10,000rpm 2.5 HD. There are many variables left out and what seems to me to be a plug for the Oyen and the LaCie! I was really hoping to learn some factual data.
firewire 800 eats usb 3.0 alive don’t be fooled. just because a computer has support for the buggy usb 3.0 doesnt mean the drive will run at 3.0 speeds. in many case the drive will run at 2.0 but you can’t really tell which is being used. I had WD external drive was supposedly USB 3.0 was horrible.
it would take about 30 minutes to transfer 10GB of data. I took it out of the pile of crap enclosure and put it in a lacey enclosure that offered both usb 3.0 and firewire 800 both and I tried both. the firewire blew the USB 3.0 away. the same 10GB of data took only 3 minutes to move verse close to 30 minutes.
Well done Terry, another good video as always. You get right to the point and don’t talk about a lot of fuss that no one cares about. Thanks.
Hi Terry- do you think a new 7200 Seagate HYBRID Hard Drive connected via FW800 is faster than the stock 2009 hard drive in my white imac (internal)? It seems to be really slow!
Definitely! any newer 7200 rpm drive is going to be faster than the one they were putting in back in 2009.
even connected via Firewire vs the old old drive in SATA?
Yes if the old drive was 5400 rpm which I suspect it was.
This was a great video! Very informative. Well spoken. Easy to understand! Thank you so much for taking the time to make this! Very cool! Great job!!
Thanks Simon for saving me the time of sitting through an 8 minute video to find out what I could read in 5 seconds. I got my information from reading the comments then the web post. I’ll get the rest of my information elsewhere, thanks.
3.0 with video editing is MUCH faster for me than Firewire 800. Just thought I’d throw my 2 cents in.
This isn’t correct, FireWire 800 does up stream and downstream all at once. USB 3.0 can only downstream or upstream at any one given time. This is not an accurate test.
That’s not a fair test, Apple developed Firewire 800 in the 1990’s and it’s 15 year old technology, when only USB 2.0 was offered on PC’s, you should be comparing Thunderbolt to USB 3.0.
http://www.apple.com/thunderbolt/
So which one is faster? Sorry, don’t have time to watch an 8 minute video just to find out if USB 3.0 is faster, or if FireWire is.
Interesting and very helpful, Terry! Thank you for sharing the test and the results! I’m just up a large external drive for storage and Time Machine backup and a 2nd external drive to back up the first drive. Good to know using USB3 will save me a little backup time!
Thank you for taking your valuable time to do this test for us!.
I have that exact same cabinet and it is great for copying files in/out of a 2010 generation Macbook Pro where FW800 is loads faster than USB2 which is the next lower speed alternative, followed by Ethernet…
So with that era MacBooks this cabinet is wonderful for disk cloning, backups and such. You can hook it up to USB3 to copy some large stuff over from a modern computer then hook up to the 2010 Macbook and copy stuff over at FW speeds.
Regarding that test; Actual implementation of Firewire and USB3 protocols in the cabinet product is not representative to all products out there. Some manufacturer could have succeeded a lot better (or worse) with the performance attainable with their protocol implementation of USB3 and Firewire than have some other vendor. It´s about software algorithms and developer skills. So your test is only a hint at that USB3 can be faster than FW800 om some cabinet products.