Conclusion: Check Performance First, Then Buy
There are basically two types of USB drives: traditional thumb drives optimized for space and larger flash drives that offer somewhat more storage capacity. Thumb drives fit easily into your pocket and can be carried anywhere. The other option is a drive the size of a small smart phone.
Kingston and OCZ offer the latter, and their products clearly dominate the performance benchmarks. This advantage, however, comes at the expense of mobility. In choosing between these two devices, specifically, we recommend the OCZ Enyo, which clearly outperformed Kingston when writing various file types, while the difference between the two was hardly noticeable in the read benchmarks.
Kingston is not the only device to narrowly miss our recommendation. We would currently also hesitate to recommend the FastKey from LaCie due to inconsistencies in the accompanying software. More specifically, the device failed to support the 64-bit version of Windows 7, which we find unacceptable for a premium product. From a purely technical standpoint, LaCie performed well across the board. Its FastKey is the only drive in a compact USB format to offer genuinely impressive performance.
Our choice among the USB 3.0 thumb drives is Super Talent's SuperCrypt, which combines encryption and attractive performance without any significant weaknesses in performance.
Limited performance was demonstrated by many of the other drives, albeit in various tests. Hardly any drive was capable of handling a large number of I/O operations per second, although we have to make clear that these products are not made to handle such tasks anyway, so their failure to do so is not critical at all.
When reviewing the performance data, we compare our expectations from a USB 3.0-capable device with the measured performance numbers. And here we find some discrepancies that should not exist, especially since we're talking about products that imply high performance. When writing small files, PQI’s 300 KB/s or Patriot’s 800 KB/s are absurd. Many of the tested devices easily wrote 7.5 to almost 20 MB/s. The task is far from impossible, and these results simply mean that manufacturers with the poorly-performing products need to go back to the drawing board.
Combined sequential reading and writing of various file types also caused some products to hit their limits rather quickly. While Kingston’s Data Traveler Ultimate 3.0 performed well in other tests, it froze at 8 MB/s of combined reading/writing. PQI and A-Data didn’t fare any better.
The overall conclusion is a bit of urgent advice from us: do not purchase USB 3.0 flash drives without reliable test data in your hands. The probability that the SuperSpeed USB logo is being used to sell a technically inadequate product at a premium price is simply too high.
- Super Talent,
- Patriot,
- OCZ,
- LaCie,
- Kingston,
- Storage,
- A-Data,
- usb-3.0 ,
- thumb-drive ,
- flash-drive
Latest External Storage News
Latest External Storage reviews
- 10/04 – In Pictures: External Data Storage Through The Ages
- 10/11 – Round-Up: 15 microSDHC Cards, Benchmarked And Reviewed
- 25/05 – Qnap TS-559 Pro+: Familiar Network Storage With A New CPU
- 24/05 – 10 SDXC/SDHC Memory Cards, Rounded Up And Benchmarked
- 18/05 – Is Data Encryption Worth Destroying Your NAS' Performance?
-
Latest Super Talent News
- Super Talent Intros Storage POD Mini USB 3.0 External SSD
- Patriot, Super Talent Launch SSDs with SF-2200
- Super Talent Announces the ShuttleCraft SAS SSD
-
Latest Super Talent reviews
- Nine USB 3.0 Flash Drives For Road Warriors
- USB 3.0 On A Stick: Super Talent's RAIDDrive 64 GB
-
Latest Patriot News
- Patriot Memory Updates SSD Line With Pyro SE Series
- Patriot Memory Launches SATA 3-Charged Pyro SSDs
- Patriot, Super Talent Launch SSDs with SF-2200
- Patriot Adds SandForce SSD With Inferno Series
- New Patriot SSDs Use Latest JMicron Controller
-
Latest Patriot reviews
- Second-Gen SandForce: Seven 120 GB SSDs Rounded Up
- Nine USB 3.0 Flash Drives For Road Warriors
This was pretty helpful.
Was considering purchasing a USB 3.0 device. thanks
I hope that before testing all these devices, you first completely filled them with data by writing to all sectors twice - otherwise none of your testing is valid as you need to test read/erase/write times. Testing on brand new (blank) sticks is useless for real world benchmarking.