Just this year, I’ve added two new Lenovo laptops to my stable of PCs. Both include 8th generation Kaby Lake Intel CPUs along with a single USB-C (3.1) port, and an attendant UCSI driver that supports ACPI transport. I’ve long had access to Sabrent USB 3.0 SSD enclosures for m.2 SSDs (e.g. a Samsung 860 EVO 1 TB devices, plus numerous smaller 850 EVO devices). Last week, I decided to pony up for a Sabrent NVMe USB-C SSD enclosure (EC-NVME US$39) and a Samsung 970 EVO NVMe SSD (MZ-V7E1T0BW US $170) to put in it. Total cost: just under US$210.
At 4.1×1.75×0.6″/10.4×4.5×1.6cm) this compact unit is made of milled aluminum, so it’s sturdy and keeps the SSD surprisingly cool (28 °C/82 °F).
A Bigger Difference Than I’d Imagined: USB 3.0/M.2 vs. USBC(3.1)/NVMe
To begin with, I ran the well-known CrystalDiskMark (v6.0.2) with the same settings against both devices attached to my Lenovo Yoga X380 ThinkPad PC. That is, the NVMe device was plugged into the unit’s USB-C port, while the M.2 device was plugged into a USB 3.0 port. All other factors were kept equal, so this should be a pretty valid comparison, starting at the USB port for each device.
The difference is pretty staggering with ratios ranging from 1.77:1 on the low end to 22.75:1 on the high end! Overall, this was a much bigger speed difference than I’d expected, even knowing the theoretical throughput differences between the two interfaces to be 2:1, with real-world performance (I thought) likely to be somewhat better. Here’s a comparison, with M.2 on the left and NVMe on the right:
This is just for side-by-side comparison, but the ratio tops out at 22.75:1 (M.2:NVMe). Wow!
[Click image for full-sized view.]
Benchmarks Never Tell the Whole (or a Completely True) Story
I used to write hardware reviews for Tom’s Hardware back in the 2000s. One thing I learned in benchmarking PCs and other computing gear was that benchmarks only tell part of the performance story for any device. So I put these two drive enclosures head-to-head in an activity I engage in every day, and that matters a LOT to me: Macrium Reflect image backup. To me, this represents a much better — and more significant — measure of speed differences between the two devices. Here’s the concluding report from Reflect for each image backup (again, M.2 left and NVMe right):
Do the math in seconds. 13:25=805; 3:49=229. That’s a 3.5:1 ratio In other terms, the NVMe finishes 9:36 before the M.2 device.
[Click image for full-sized view.]
Don’t get me wrong. US$210 is a hefty price to pay for a 1TB backup device/external laptop storage drive. But personally, I think the cost is justified by the device’s speed. I’m going to acquire a 2.5″ drive enclosure for the various 2TB 2.5″ drives I also keep around for this purpose. Then I can compare USB-C for a spinner against USB-C for NVMe, which may be more meaningful for most readers buying new gear. For myself, and my sizable collection of old as well as new storage media, this comparison is a real eye-opener. Perhaps for other readers, too?
Author: Ed Tittel
Ed Tittel is a 30-plus-year computer industry veteran. He’s a Princeton and multiple University of Texas graduate who’s worked in IT since 1981 when he started his first programming job. Over the past three decades he’s also worked as a manager, technical evangelist, consultant, trainer, and an expert witness. See his professional bio for all the details.