Samsung’s Portable SSD T5 reviewed

The density wars are in full swing between the major flash players, and the name of the game is 64-layer 3D NAND. We recently experienced Toshiba’s 64-layer BiCS TLC firsthand in the company’s excellent XG5 SSD. IMFT’s 64-layer stuff is already out on the market in the Intel 545s, but we’ve yet to spend any hands-on time with it. Samsung’s 64-layer V-NAND has been on our radar for just about a year now, but the company has really been talking up a storm about the technology in the last couple of months.

Well, brace yourselves: we’ve made it to the eye of the storm. Today, Samsung officially unveils the first client drive with 64-layer V-NAND inside: the Portable SSD T5. The company was kind enough to send us 500GB and 1TB versions of its latest to run through the wringer.

The T5 closely follows in the footsteps of its forebear, the Portable SSD T3. It retains the same compact size and feathery weight, ringing in at 3 x 2.3″ x 0.4″ and 51 grams. The housing construction has changed slightly in this version, though. The T5 sheds the T3’s two-piece metal and plastic body for an all-aluminum unibody. The 250GB and 500GB versions are finished in, ahem, “Alluring Blue,” while the 1TB and 2TB models are clad in a simpler “Deep Black.”

The big news, of course, is the move from 48-layer TLC V-NAND to the 64-layer stuff. Samsung reckons that the newfound speeds engendered by that transition require outfitting the T5 with a USB 3.1 Gen 2 interface instead of Gen 1. The company claims the T5 can hit peak speeds of 540 MB/s, which should theoretically fit within the 5 Gb/s maximum that USB 3.1 Gen 1 (or USB 3.0, if you prefer) affords. In reality, encoding overhead can be a buzzkill for the older standard. More on that later.

The 500GB drive drew the short straw, so it had to endure being stripped down to its skivvies. The 64-layer TLC V-NAND is distributed across only two packages on the same side of the PCB as the controller. That controller appears to be a variant of the same MGX chip found inside the T3. In a rare but sensible deviation from complete vertical integration, Samsung seems to be passing USB controller duties off to an ASMedia mSATA-to-USB bridge controller.

The T5 supports AES 256-bit hardware encryption, accessible via Samsung’s Portable SSD software. As it did with the T3 before, the software provides easy interfaces for encrypting, decrypting, and updating the drive’s firmware. The software is available for PC, Mac, and Android. Unlike the T3, the T5 includes a USB Type-C to Type-C cable in addition to the vanilla Type-C to Type-A cable, so consuming your encrypted files on-the-go should be dead simple. At the time of this writing, however, the Android app for T5 was not available for experimentation.

Now let’s get down to testing and and bring RoboBench to the stage.

TR RoboBench — Real-world transfers
First, a quick update on our test setup. Our Intel Z97 storage rigs are equipped with mere USB 3.1 Gen 1 ports. Therefore, we secured a Rosewill add-in card to let the T5 run wild at Gen 2 speeds. The card uses the same ASMedia ASM1142 controller that is very commonly found on motherboards with 3.1 Gen 2 capabilities.

You may notice some data in our results from a drive we haven’t yet reviewed: Adata’s SD700. The full review is in the works, but since we already had the raw numbers we decided to include them in our graphs this time around. Stay tuned for our full look at the SD700 soon.

RoboBench comprises real-world transfers with a range of file types. Developed by our in-house coder, Bruno “morphine” Ferreira, this benchmark relies on the multi-threaded robocopy command build into Windows. We copy files to and from a wicked-fast RAM disk to measure read and write performance. We also cut the RAM disk out of the loop for a copy test that transfers the files to a different location on the SSD.

Robocopy uses eight threads by default, and we’ve also run it with a single thread. Our results are split between two file sets, whose vital statistics are detailed below. The compressibility percentage is based on the size of the file set after it’s been crunched by 7-Zip.

The media set is made up of large movie files, high-bitrate MP3s, and 18-megapixel RAW and JPG images. There are only a few hundred files in total, and the data set isn’t amenable to compression. The work set comprises loads of TR files, including documents, spreadsheets, and web-optimized images. It also includes a stack of programming-related files associated with our old Mozilla compiling test and the Visual Studio test on the next page. The average file size is measured in kilobytes rather than megabytes, and the files are mostly compressible.

RoboBench’s write and copy tests run after the drives have been put into a simulated used state with 30 minutes of 4KB random writes. The pre-conditioning process is scripted, as is the rest of the test, ensuring that drives have the same amount of time to recover.

Let’s take a look at the media set first. The buttons switch between read, write, and copy results.

The T5 drives immediately set themselves apart from the USB 3.1 Gen 1 drives. Read speeds are a solid 10% ahead of the T1 and SD700, but write speeds are an incredible 50% higher in the single-threaded test. Across read, write, and copy, the T5 is far closer to the SATA 850 EVO 2TB than to any of our older USB drives. 10 Gbps USB is some pretty good stuff.

Next up, the work set.

The story is much the same with the work set, at least with a single thread. The T5 drives seem to dislike reading the work set with eight threads, putting up numbers about 20% worse than the T3 did. But again, write numbers are vastly improved.

The T5 performed better than our USB 3.1 Gen 1 drives across all our tests except one, and it generally won by large margins. Additionally, the 500GB drive’s results were just as good as the 1TB’s. “Alluring Blue” is no misnomer.

Ordinarily we’d be done here, but since we had some extra time, we decided to rerun a few tests with the T5 plugged into one of our motherboard’s USB 3.1 Gen 1 ports to see how much of a difference Gen 2 makes. Let’s take a look.

When mated to a slower Gen 1 port, the T5 500GB’s speeds are almost identical to that of the T3. Samsung’s 64-layer NAND might be bringing some new mojo with it, but it appears that much of the speed increase might be credited to the USB 3.1 Gen 2 upgrade. But without a 48-layer V-NAND device running on Gen 2 to compare against, we can’t definitively attribute the gains to one factor or another.

That’s it for testing. Hit the next page for our test methods, or skip ahead to the conclusion.

Test notes and methods
Here’s are the essential details for all the drives we tested:

USB 3.1 Gen 1 drives were connected to the motherboard via USB 3.0 port. USB 3.1 Gen 2 drives were connected via a 3.1 Gen 2 add-in card. The 2.5″ drives were connected via a USB 3.0 drive dock for SATA drives.

We used the following system for testing:

Thanks to Asus for providing the systems’ motherboards, to Intel for the CPUs, to Adata for the memory, to Fractal Design for the cases, and to Corsair for the system drives and PSUs.

We used the following versions of our test applications:

• IOMeter 1.1.0 x64
• TR RoboBench 0.2a

Some further notes on our test methods:

• To ensure consistent and repeatable results, the SSDs were secure-erased before every component of our test suite. For the IOMeter database, RoboBench write, and RoboBench copy tests, the drives were put in a simulated used state that better exposes long-term performance characteristics. Those tests are all scripted, ensuring an even playing field that gives the drives the same amount of time to recover from the initial used state.

• We run virtually all our tests three times and report the median of the results. Our sustained IOMeter test is run a second time to verify the results of the first test and additional times only if necessary. The sustained test runs for 30 minutes continuously, so it already samples performance over a long period.

• Steps have been taken to ensure the CPU’s power-saving features don’t taint any of our results. All of the CPU’s low-power states have been disabled, effectively pegging the frequency at 3.5GHz. Transitioning between power states can affect the performance of storage benchmarks, especially when dealing with short burst transfers.

The test systems’ Windows desktop was set at 1920×1080 at 60Hz. Most of the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.